تحليل تكاملي يكشف عن ارتباطات بين اختلال ميكروبيوتا الفم والتشوهات الجينية والوراثية المضاعفة في سرطان الخلايا الحرشفية في تجويف الفم Integrative analysis reveals associations between oral microbiota dysbiosis and host genetic and epigenetic aberrations in oral cavity squamous cell carcinoma

المجلة: npj Biofilms and Microbiomes، المجلد: 10، العدد: 1
DOI: https://doi.org/10.1038/s41522-024-00511-x
PMID: https://pubmed.ncbi.nlm.nih.gov/38589501
تاريخ النشر: 2024-04-08

تحليل تكاملي يكشف عن ارتباطات بين اختلال ميكروبيوتا الفم والتشوهات الجينية والوراثية المضاعفة في سرطان الخلايا الحرشفية في تجويف الفم

ليويانغ كاي هينغيان زو تشين تشين مو بو يي وونغ لينلين لان تشيري و. ك. نج بو لي مان كيت تشيونغ © دايجوانرو وانغ إيدي و. ي. وونغ إريك إتش. إل. لاو زينون و. س. يونغ رونالد لاي كاتي ميدان شيروود فنج كوان تشي أ. تشان فيفيان و. ي. لوي ألفريد س. ل. تشينغ جون يو بول ك. س. تشان جيسون واي كيه تشان وزيغي تشين

(أ) التحقق من التحديثات

الملخص

تم الإبلاغ عن أن اختلال التوازن في الميكروبيوم الفموي البشري مرتبط بسرطان الخلايا الحرشفية في تجويف الفم (OSCC)، بينما تظل التفاعلات بين المضيف والميكروبيوم فيما يتعلق بالتأثير المحتمل للبكتيريا المسببة للأمراض على الشذوذات الجينومية والإيبيجينية للمضيف غير مدروسة بشكل كافٍ. في هذه الدراسة، تم تحليل المجتمع البكتيري المخاطي، والتعبير الجيني على مستوى الجينوم للمضيف، وميثيلين DNA CpG في وقت واحد في الأورام وأنسجتها الطبيعية المجاورة لمرضى OSCC. لوحظت زيادة ملحوظة في الوفرة النسبية لسبع أنواع من البكتيريا (Fusobacterium nucleatum، Treponema medium، Peptostreptococcus stomatis، Gemella morbillorum، Catonella morbi، Peptoanaerobacter yurli وPeptococcus simiae) في بيئة الورم لـ OSCC. شكلت هذه البكتيريا الغنية بالورم 254 ارتباطًا إيجابيًا مع 206 جينات مضيف مرتفعة التعبير، تتعلق بشكل رئيسي بمسارات الإشارات المرتبطة بالالتصاق الخلوي، والهجرة، والتكاثر. حدد التحليل التكاملي لارتباطات البكتيريا-التعبير الجيني وارتباطات البكتيريا-الميثيلين ما لا يقل عن 20 جينًا مضيفًا غير منظم مع ميثيلين CpG مقلوب في مناطق المحفز الخاصة بها مرتبطة بزيادة البكتيريا المسببة للأمراض، مما يشير إلى إمكانية البكتيريا المسببة للأمراض في تنظيم التعبير الجيني، جزئيًا، من خلال التغيرات الإيبيجينية. أكدت نموذج في المختبر أن Fusobacterium nucleatum قد تساهم في غزو الخلايا من خلال التفاعل مع E-cadherin. -إشارات بيتا-كاتينين، TNF إعادة تشكيل المسار والمصفوفة خارج الخلوية من خلال زيادة تنظيم جين SNAI2، وهو عامل نسخي رئيسي في الانتقال الظهاري-الم mesenchymal (EMT). عملنا باستخدام نهج متعددة الأوميات استكشف التفاعلات المعقدة بين المضيف والميكروبات وقدم رؤى مهمة حول الأساس الجيني والوظيفي في نشوء أورام سرطان الفم والبلعوم (OSCC)، والتي قد تكون بمثابة مقدمة لدراسة مدفوعة بالفرضيات لفهم العلاقة السببية للبكتيريا المسببة للأمراض في هذا السرطان القاتل.

سرطان الخلايا الحرشفية في تجويف الفم (OSCC) هو أكثر الأورام الخبيثة شيوعًا في الفم على مستوى العالم ويرتبط بمعدلات وفيات ومراضة كبيرة. . حول من سرطانات الفم يتم تشخيصها في مرحلة متقدمة، مما يؤدي إلى معدلات بقاء لمدة 5 سنوات تقل عن تشمل العوامل المسببة المعروفة لتكون أورام سرطان الفم والبلعوم استهلاك التبغ وإساءة استخدام الكحول.
ومضغ جوز البتلة؛ ومع ذلك، فإن انتشار سرطان الفم والبلعوم بدون عوامل الخطر التقليدية قد زاد في السنوات الأخيرة لذا هناك حاجة ملحة لتحديد أسباب أخرى أساسية ذات صلة بالتشخيص المبكر وتحسين العلاج لدى مرضى سرطان الفم والبلعوم. نظرًا للزيادة في الأدلة التي تشير إلى الإمكانات المسرطنة للإنسان
الميكروبيوم في السرطانات إن توضيح الميكروبيوم البشري في سرطان الفم والبلعوم قد يفسر، جزئيًا، حقيقة أن مجموعة من المرضى الذين لم يتعرضوا لعوامل الخطر التقليدية ينتهي بهم المطاف إلى تطوير السرطان. في الواقع، تم الإبلاغ عن فقدان كبير في التنوع الميكروبي لدى مرضى سرطان الفم والبلعوم. ; تم تعزيز العديد من مسببات الأمراض اللثوية، على سبيل المثال، الفوسوبكتيريوم، البيبتوستربتوكوكوس والتريبونيما، بشكل ملحوظ في بيئة الورم لسرطان الخلايا الحرشفية الفموية تم الإبلاغ عن أن هذه البكتيريا المسببة للأمراض تعزز عدوانية سرطان الفم من خلال تفعيل مسار إشارات التكامل/FAK المحفز بواسطة TLR/MyD88. مما يجعل فهم البكتيريا المسببة للأمراض في السرطان ذا أهمية كبيرة.
هناك وعي متزايد بأن الأنماط العالمية للتغيرات الجينية والإبيجينية تلعب دورًا حاسمًا في الخصائص الجزيئية لسرطان الفم. على سبيل المثال، تعتبر مسارات EGFR وPIK3CA وNOTCH وجين TP53 من بين الأكثر تغييرًا في سرطان الخلايا الحرشفية في الرأس والعنق (HNSCC). لقد حددت عدة دراسات لسرطان الرأس والعنق ميثلة المحفز لجين CDKN2A (p16) وجين DAP كيناز (DAPK) وجينات إصلاح الحمض النووي. و يمكن أن تؤدي البكتيريا إلى تحفيز التعديل الوراثي البشري، مما يؤدي إلى كتم تعبير جينات مثبطات الأورام. لذا، فإن تحليل الأحداث الجينية والوراثية التي قد تتأثر بخلل التوازن في الميكروبات الفموية قد يوفر لنا فهماً أساسياً لدورها في تنظيم الجينات ويمكننا من إنشاء علامات محددة للتشخيص المبكر واستراتيجيات العلاج.
في هذه الدراسة، تم تحليل أورام سرطان الخلايا الحرشفية الفموية (OSCC) وأنسجتها الطبيعية المجاورة (AN) من مجموعة من المرضى السلبيين لفيروس الورم الحليمي البشري (HPV) بشكل متزامن من حيث الميكروبيوم المخاطي الفموي، والتعبير الجيني على مستوى الجينوم المضيف، وميثيلين الحمض النووي CpG لاستكشاف الأساس الجيني لتفاعلات المضيف والميكروبيوم. كشفت التحليلات التكاملية باستخدام أساليب متعددة الأوميات عن شبكات معقدة بين اختلال توازن الميكروبيوم الفموي والاختلالات الجينية والوراثية للمضيف. توفر نتائجنا رؤى مهمة حول الأساس الجيني والوظيفي لفهم أفضل لدور البكتيريا الفموية في مسببات سرطان الخلايا الحرشفية الفموية.

النتائج
مواضيع الدراسة

تمت دراسة 98 مريضًا بسرطان الخلايا الحرشفية الفموي (OSCC) قدموا أنسجة ورم وأنسجة طبيعية متطابقة. من بينهم، تم استبعاد ثمانية مرضى مصابين بأنواع فيروس الورم الحليمي البشري عالية الخطورة (7 مصابين بـ HPV16 وواحد بـ HPV18). تألفت هذه المجموعة المتبقية من 57 ذكرًا و33 أنثى، بمتوسط عمر 65 عامًا (الانحراف المعياري: 12 عامًا). تتوفر معلومات ديموغرافية وسريرية مفصلة في الجداول التكميلية 1 و2.

اختلال ميكروبيوم الفم في سرطان الفم الحرشفي

كانت قراءات البكتيريا عالية الجودة من النوع 16 S من الأنسجة الورمية والأنسجة الطبيعية المتزاوجة متاحة من 81 مريضًا بسرطان الفم (OSCC) (متوسط القراءات من ) (الجدول التكميلي 3 والبيانات التكميلية 1-3)، مع Firmicutes (متوسط الوفرة النسبية لـ الانحراف المعياري لـ ) كأكثر فصيلة بكتيرية سائدة، تليها الفوسوبكتيريا ( بروتيوبكتيريا ( بكتيرويديتس ) وسبعة أخرى من الفصائل (الشكل التوضيحي 1a). على مستوى متغير تسلسل الأمبليكون (ASV)، لوحظ انخفاض كبير في تنوع ألفا لمجتمع الميكروبات في الغشاء المخاطي الفموي في أنسجة الورم كما تم قياسه بواسطة مؤشرات الثراء، وشانون، وسيمبسون مقارنةً بمجموعة التحكم AN (مان-ويتني اختبار، ) (الشكل 1أ)؛ وبالمثل، أظهر تحليل المكونات الرئيسية (PCoA) المستند إلى مصفوفة GUniFrac الموزونة فصلًا واضحًا بين مجموعتي الورم وAN ( تؤكد هذه النتيجة على المجتمع الميكروبي المتميز المرتبط بمواقع أورام سرطان الفم والبلعوم مقارنةً بالأنسجة الطبيعية المجاورة.
استخدمنا تحليل التباين المتعدد المتغيرات القائم على التباديل (PERMANOVA) لت quantifying مساهمة العوامل (حالة المرض، الجنس، الشيخوخة، التدخين، استهلاك الكحول، ومراحل سرطان T و N) في اختلافات التركيب الميكروبي لدى المرضى الذين يعانون من سرطان الفم والبلعوم (OSCC). حدد هذا التحليل حالة المرض كعامل مستقل كان له أكبر تأثير على الهيكل العام للميكروبيوم الفموي (GUniFrac الموزون). ; GUniFrac غير الموزون ; براي-كورتيس ) (الشكل التوضيحي التكميلي 1ب). بالإضافة إلى ذلك،
عوامل أخرى، بما في ذلك مرحلة T (T1 و T2 مقابل T3 و T4) (وزن GUniFrac ; GUniFrac غير الموزون ; براي-كورتيس ) والتدخين (نعم مقابل لا) (GUniFrac الموزون ; براي-كورتيس كان لها أيضًا تأثيرات كبيرة، وإن كانت أقل. يبرز تأثير حالة المرض على تركيبة الميكروبات الفموية دورها المحتمل في مسببات سرطان الفم.
استخدام اختبار حجم التأثير لتحليل التمييز الخطي (LEfSe) (LDA > 3، )، والذي تم التحقق منه أيضًا من قبل اثنين على الأقل من الأدوات الثلاثة المعنية بالتكوين، ANCOM-BC2 و ALDEx2 واختبارات ZicoSeq المعدلة لمتغيرات مرحلة T والتدخين ( قمنا بتطبيق خوارزمية تحديد الموضع النشوي، pplacer، لرسم قراءات ASV مقابل قاعدة بيانات مرجعية لـ 16S rRNA من NCBI باستخدام أقصى احتمال، ولاحظنا 6 و 13 جنسًا بكتيريًا تم إثراؤه بشكل كبير وانخفاضه في الوفرة النسبية في بيئة الورم OSCC، على التوالي (الشكل التكميلي 1c، الجدول 1 والجدول التكميلي 4a). علاوة على ذلك، وُجد أن سبعة أنواع بكتيرية مخاطية (Fusobacterium nucleatum، Treponema medium، Peptostreptococcus stomatis، Gemella morbillorum، Catonella morbi، Peptoanaerobacter yurli و Peptococcus simiae) كانت مستعمرة بشكل رئيسي في تجويف الفم لمرضى السرطان (الشكل 1c، الجدول 1 والجدول التكميلي 4b)، مما يشير إلى إمكانية وجود بكتيريا غنية بالورم في مسببات مرض OSCC.
تمكن التجميع الهرمي المستنتج من 6 أنواع بكتيرية تمييزية (LDA > 3، AUC > 0.65) من تمييز العينات المدروسة إلى مجموعتين، تتكون من غالبية الأورام (62/92) والشواهد (51/70)، مع نسبة احتمالية تبلغ 5.48 (95% CI 2.66-11.68، ) (الشكل 1د). حققت هذه الأنواع البكتيرية منطقة مجمعة تحت منحنى التشغيل الاستقبالي (ROC) (AUC) تبلغ 0.858 (95% CI 0.803-0.914). قد تكون هذه التمييز أساسًا لتطوير علامات تشخيصية لسرطان الفم الحرشفي.

فوسوبكتيريوم نوكليتوم المرتبط بمرضى سرطان الفم والبلعوم الحرشفي دون عوامل خطر تقليدية

نظرًا لأن F. nucleatum كان النوع البكتيري الأكثر انتشارًا في بيئة الورم في سرطان الفم والبلعوم (OSCC) مقابل ) (الشكل التوضيحي التكميلي 2a والجدول 1)، قمنا بتحليل إضافي لاستعمارها المرتبط بخصائص المرضى. إن إثراء . كان النوكليتوم مرتبطًا بشكل كبير بغير المدخنين (متوسط تغيير اللوغاريتم الثنائي بمقدار 1.48 مقابل. ) (الشكل التوضيحي 2ب) وغير المدخنين (1.35 مقابل ) (الشكل التوضيحي 2c). نظرًا لأن جميع المرضى المجندين كانوا سالبين لفيروس الورم الحليمي البشري عالي الخطورة، فإن الاكتشاف يوحي بإمكانية الدور المسبب للأمراض لـ . النوكتوم في مرضى OSCC السالبين لعدوى HPV، والتدخين، واستهلاك الكحول (ثلاثي السلبية). كما ارتبط F. النوكتوم بتحسن البقاء على قيد الحياة المحدد بالسرطان (1.42 مقابل. ) (الشكل التوضيحي الإضافي 2d)، والذي كان من المحتمل أن يدعمه بقاء محدد للمرض لمدة 3 سنوات أفضل ( مقابل عند 36 شهرًا، ) (الشكل التوضيحي 2e). تشير هذه النتائج إلى أن يمكن أن يكون Nucleatum علامة مميزة في حالات سرطان الفم والبلعوم الحرشفي (OSCC) السلبية لفيروس الورم الحليمي البشري (HPV)، لا سيما بين غير المدخنين وغير الشاربين، وقد يكون أيضًا دليلاً على تشخيص أكثر إيجابية في هؤلاء المرضى.

تعبير الجينات المضيفة المختلفة في ترانسكريبتوم سرطان الفم والبلعوم

تمت دراسة مجموعة فرعية من 40 ورمًا و22 نسيجًا من الأنسجة الطبيعية (AN) لتوصيف التعبير الجيني على مستوى الجينوم المضيف (الملحق الجداول 2 و5)، مما كشف عن إعادة تنظيم عالمية لخلل الجينات المضيفة في بيئة الورم لسرطان الفم والبلعوم (OSCC) (الملحق الشكل 3a والملحق الجدول 6). من بين 17,225 نسخة جينية بمتوسط نسخ لكل كيلوباز مليون أكبر من 1 (TPM > 1)، حددنا 1407 جينًا مرتفع التعبير (مرتبطة بالورم) و1525 جينًا منخفض التعبير (مرتبطة بالأنسجة الطبيعية) في مجموعة OSCC المدروسة (log2FC > 1 أو ) (الأشكال التكميلية 3a و3b والجدول التكميلية 6)، متوافقة بشكل كبير مع ملف النسخ الجيني لبيانات TCGA OSCC (معامل سبيرمان ) (الشكل التوضيحي 3c). على سبيل المثال، كانت تعبير LAMC2 (المعني بهجرة الخلايا الظهارية) مرتفعًا بشكل ملحوظ في كلا ، ) (الشكل 2ب) ومجموعات بيانات TCGA-OSCC (الشكل التكميلي 3د). وبالمثل، تم تصنيف MMP1 (المعني بغزو خلايا الورم) كواحد من
الشكل 1 | اختلال ميكروبيوم الفم المرتبط بسرطان الخلايا الحرشفية في تجويف الفم (OSCC). أ مقارنة تنوع ميكروبيوم الفم بين ورم OSCC (الورم) والأنسجة الطبيعية المجاورة (AN) على مستوى متغير تسلسل الأمبليكون (ASV). تشير الخط المركزي في الرسم البياني إلى القيمة الوسيطة، وتمثل حدود الصندوق الربع الأول والثالث، وتمتد الشعيرات إلى أصغر وأكبر القيم في البيانات، على التوالي. ب رسم تحليل المكونات الرئيسية بناءً على مصفوفة مسافة GUniFrac الموزونة المستنتجة من ASVs. ج الأنواع البكتيرية التمييزية كما تم اكتشافها بواسطة تحليل التمييز الخطي (LDA) حجم التأثير (LEfSe)
تحليل (الدرجة )، والذي تم التحقق منه بشكل إضافي من قبل اثنين على الأقل من الأدوات الثلاثة المعنية بالتكوين، ANCOM-BC2 و ALDEx2 واختبارات ZicoSeq المعدلة لعوامل التداخل لمرحلة T والتدخين ( يمثل طول الشريط درجة LDA على مقياس اللوغاريتم العشري. تم اختبار الفروق في الوفرة النسبية بشكل إضافي بواسطة اختبار مان-ويتني الزوجي. اختبار و Tukey HSD بعد الاختبار كما هو موضح في الألواح اليمنى. ; ; تحليل التجميع الهرمي باستخدام مصفوفة المسافة لستة أنواع بكتيرية تمييزية (LDA الجامعة الأمريكية في القاهرة صنفت عينات الأنسجة التي تم مسحها إلى فصيلتين.
الجينات الأكثر تنظيمًا للأعلى ” ). تشمل أمثلة الجينات المنخفضة التنظيم مثبطات الأورام CRISP3 ( ) و EMP1 ( ).
تحليل إثرائي يصنف 1292 جينًا مختلفًا معبرًا عنه مرتبطًا بالأورام (DEGs) ) في قاعدة بيانات موسوعة كيوتو للجينات والجنوم (KEGG) وجدت أن معظم المسارات الكانونية المتغيرة كانت مرتبطة بالإشارات المسرطنة وتنظيم المناعة ( ) (الشكل 2 ج والجدول التكميلي 7). على سبيل المثال، تفاعل مصفوفة خارج الخلية – مستقبل (ECM-receptor) ( ) التي تتعلق بتنظيم حركة الخلايا، التكاثر والتمايز كانت المسار الأكثر غنى، حيث تم التعبير عن مجموعة متنوعة من الجينات المعبر عنها بشكل مختلف التي تشفر الكولاجين، الإنتجرين واللامينين بشكل مرتفع ( ) (الشكل 2د). تم ملاحظة اضطرابات في مجموعة متنوعة من مسارات الأيض عندما تم تحديد الجينات المعبر عنها بشكل منخفض.
( تم تلخيصها للتحليل الوظيفي (الشكل التكميلي 3e والجدول التكميلي 7). على سبيل المثال، تم الإشارة إلى أن تقليل استقلاب التيروزين مرتبط بتقدم الورم في عدة أنواع من السرطانات مثل سرطان المريء. سرطان الفم وسرطان الخلايا الكبدية .
تحليل البقاء حدد أيضًا 62 جينًا معبرًا عن الورم و48 جينًا معبرًا عن AN مع قيم تنبؤية لـ OSCC. ) (الجدول التكميلي 8). وشملت هذه MMP1 (المعنية بالتكاثر والتمايز) ( ), سيربين1 (المعني في هجرة الورم وإعادة تشكيل الأنسجة) ( ) ، COL5A2 (المعني في تكاثر الخلايا والغزو) ( )، و FAP (المعنية بالبوليبوز العائلي الأدينوماتي) ( ) التي كانت مرتبطة بشكل كبير بالبقاء على قيد الحياة المحدد بالمرض (الشكل 2e). إن تحديد هذه الجينات المعبر عنها بشكل مختلف يوفر مشهدًا تنبؤيًا يمكن أن يكون
الجدول 1 | الأنواع والأجناس البكتيرية المميزة التي لوحظت في أنسجة أورام سرطان الفم والبلعوم مقارنةً بالأنسجة الطبيعية المجاورة المأخوذة من نفس المريض
الضرائب وفرة متوسطة ( sd) (%) اختبار LEfSe اختبار توكي HSD اختبار MWU اختبار ANCOM-BC2 اختبار ALDEx2 اختبار زيكوسيك الممثل ASV
أن ورم مجموعة LDA درجة LDA تحليل التخصيص اللاتيني قيمة الفرق (%) قيمة قيمة p قيمة قيمة قيمة p قيمة قيمة قيمة p قيمة قيمة قيمة
جنس
فوسوبكتيريوم ورم ٤.٧٨ 0.0000 12.45 (8.71, 16.20) 0.0000 0.0000 0.0000 1.12 ٤.٢٧ 0.0000 0.0002 ٥.٧٢ 0.0000 0.0001 0.0600 0.0979
تريبونيما ورم ٤.٢٩ 0.0180 4.07 (1.85, 6.29) 0.0004 0.0001 0.0028 0.81 2.57 0.0114 0.0235 ٢.٥٧ 0.0156 0.9819 0.0100 0.0017
بيبتوانيروباكتير ورم 3.82 0.0017 1.19 (0.47, 1.92) 0.0014 0.0000 0.0001 1.46 6.16 0.0000 0.0000 3.58 0.0012 0.3104 0.0100 0.0070
بيبتوستربتوكوكوس ورم 3.80 0.0019 1.20 (0.26, 2.14) 0.0125 0.0004 0.0066 0.71 ٢.٧٤ 0.0070 0.0151 3.77 0.0004 0.1432 0.0300 0.0468
كاتونيلا ورم 3.54 0.0001 0.62 (0.29، 0.95) 0.0003 0.0000 0.0003 0.72 3.23 0.0017 0.0044 3.94 0.0003 0.1222 0.0100 0.0063
بيبتوكوكوس ورم 3.06 0.0002 0.23 (0.10, 0.36) 0.0008 0.0001 0.0018 0.87 ٤.٢٣ 0.0001 0.0003 3.78 0.0008 0.2074 0.0100 0.0017
ستربتوكوكوس أن ٤.٥٥ 0.0000 -7.39 (-10.66, -4.13) 0.0000 0.0000 0.0000 -0.91 -3.64 0.0004 0.0013 -3.04 0.0046 0.7240 0.0100 0.0313
فيلونيلة أن ٤.٣٠ 0.0000 -3.97 (-5.22, -2.71) 0.0000 0.0000 0.0000 -1.11 -3.95 0.0001 0.0005 -3.98 0.0002 0.0660 0.0100 0.0067
روثيا أن 3.98 0.0000 -1.80 (-3.62, 0.01) 0.0518 0.0000 0.0000 -1.27 -4.40 0.0000 0.0001 -4.38 0.0001 0.0220 0.0100 0.0208
جرانوليكاتيلا أن 3.69 0.0000 -0.96 (-1.58، -0.34) 0.0027 0.0000 0.0000 -0.85 -3.21 0.0017 0.0044 -4.02 0.0002 0.0768 0.0100 0.0428
شاليا أن ٣.٦٤ 0.0000 -0.85 (-1.26، -0.45) 0.0000 0.0000 0.0000 -0.98 -4.03 0.0001 0.0004 -٤.٠٠ 0.0003 0.0867 0.0100 0.0013
أكتينوميس أن ٣.٥٧ 0.0000 -0.68 (-0.93، -0.43) 0.0000 0.0000 0.0000 -1.08 -5.32 0.0000 0.0000 -4.92 0.0000 0.0073 0.0100 0.0081
لاوتروبيا أن 3.36 0.0002 -0.42 (-0.68، -0.15) 0.0023 0.0000 0.0005 -0.87 -4.08 0.0002 0.0007 -3.18 0.0059 0.5965 0.0100 0.0013
كورينباكتيريوم أن ٣.٣٠ 0.0000 -0.36 (-0.59، -0.14) 0.0018 0.0000 0.0002 -0.75 -3.53 0.0008 0.0026 -3.96 0.0006 0.1612 0.0100 0.0013
يوبيكتيريوم أن 3.28 0.0393 -0.38 (-0.67، -0.09) 0.0117 0.0007 0.0113 -0.56 -2.42 0.0174 0.0337 -0.63 0.5459 1.0000 0.0100 0.0208
هالوموناس أن 3.22 0.0024 -0.28 (-0.65, 0.09) 0.1305 0.0045 0.0545 -1.15 -6.82 0.0000 0.0001 -1.99 0.1134 0.9503 0.0100 0.0470
ميغاسفيرا أن 3.16 0.0028 -0.27 (-0.47، -0.08) 0.0063 0.0003 0.0060 -0.63 -2.98 0.0042 0.0096 -2.42 0.0295 0.9625 0.0100 0.0072
لانسفيلديلا أن 3.11 0.0115 -0.24 (-0.55, 0.08) 0.1447 0.0001 0.0018 -1.06 -4.66 0.0000 0.0001 -1.20 0.2663 1.0000 0.0100 0.0070
فوكايكولا أن ٣.٠٤ 0.0040 -0.22 (-0.41، -0.03) 0.0212 0.0006 0.0109 -0.74 -3.57 0.0008 0.0027 -2.49 0.0387 0.8908 0.0100 0.0070
نوع
فوسوبكتيريوم نوكليتوم ورم ٤.٦٨ 0.0000 9.75 (5.91, 13.59) 0.0000 0.0000 0.0000 1.15 ٤.٥٤ 0.0000 0.0001 3.92 0.0002 0.1623 0.0100 0.0229 4c305539242bc00f72118f6a70d5d103
تريبونيما ميديوم ورم ٤.٠٢ 0.0030 2.06 (0.97, 3.16) 0.0003 0.0000 0.0013 1.01 ٤.٠١ 0.0001 0.0005 ٣.١٥ 0.0036 0.8050 0.0100 0.0043 88d5e4fcfd68b0cef9f267bd04d6b934
بيبتوستربتوكوكوس ستوماتيس ورم ٣.٦٧ 0.0021 0.95 (0.08, 1.82) 0.0323 0.0005 0.0094 0.67 2.78 0.0062 0.0132 3.79 0.0004 0.2569 0.0100 0.0183 feb281c58c97b847d8e32aa9dae2174b
جميلة موربيلوروم ورم 3.52 0.0000 0.67 (0.23، 1.11) 0.0034 0.0000 0.0011 0.39 1.93 0.0569 0.0879 ٤.٧٣ 0.0000 0.0140 0.0100 0.0488 66358cc06be2067caaeab6047c327466
كاتونيلا موربي ورم 3.51 0.0001 0.62 (0.29، 0.95) 0.0003 0.0000 0.0005 0.71 ٣.٤٥ 0.0008 0.0024 ٤.٠٠ 0.0003 0.1628 0.0100 0.0092 de1e79ff05735a3eb21b59e27188410c
بيبتوانيروباكتير يورلي ورم 3.38 0.0145 0.47 (0.00، 0.95) 0.0514 0.0008 0.0133 1.08 ٥.٦٤ 0.0000 0.0000 2.61 0.0246 0.9239 0.0100 0.0213 74a1666eb102cd17c22028f87467e2e4
بيبتوكوكوس سيمي ورم 3.06 0.0002 0.23 (0.10، 0.36) 0.0008 0.0001 0.0019 0.88 ٤.٦٨ 0.0000 0.0001 3.76 0.0011 0.3176 0.0100 0.0025 e5944afd1dc39fe0c43f907049fd6ac4
بورفيروموناس جينجيفاليس أن 3.96 0.0016 -1.60 (-2.77, -0.42) 0.0080 0.0000 0.0003 -0.78 -3.33 0.0013 0.0037 -2.62 0.0179 0.9642 0.0100 0.0051 204ae45c366a21148f2675b55295831c
شاليليا أودونتوليتكا أن ٣.٥٧ 0.0000 -0.72 (-1.11, -0.32) 0.0004 0.0000 0.0000 -0.78 -3.63 0.0004 0.0015 -3.86 0.0004 0.2285 0.0200 0.0043 845954a9086d53d7993b092f560b4fcc
لاوتروبيّا ميرا بيلس أن 3.32 0.0004 -0.42 (-0.66، -0.19) 0.0005 0.0000 0.0007 -0.93 -5.00 0.0000 0.0001 -2.86 0.0166 0.8482 0.0100 0.0025 14e61b2c1523a6a67df3d576d3c01fdc
أكتينوميسيس أوريز أن 3.31 0.0000 -0.41 (-0.57، -0.25) 0.0000 0.0000 0.0000 -0.86 -5.31 0.0000 0.0000 -4.51 0.0001 0.0855 0.0100 0.0025 37e783541d8270c9329ce720d8331512
كامبيلوباكتر غراسيليس أن ٣.٠٠ 0.0009 -0.18 (-0.37, 0.01) 0.0679 0.0000 0.0013 -0.72 -4.73 0.0000 0.0001 -2.50 0.0324 0.9517 0.0100 0.0124 42cb1a145d5871ce13065a619775cab6
المسارات الجينومية (KEGG) الغنية بشكل ملحوظ بواسطة الجينات المعبر عنها بشكل مرتفع في سرطان الفم والبلعوم . أحجام الدوائر تشير إلى عدد الجينات المعبر عنها بشكل مختلف في المسار. د شبكة جينات KEGG تصور الجينات المعبر عنها بشكل مرتفع في OSCC ( ).
تقدير كابلان-ماير لبقاء المرضى لمدة 3 سنوات بناءً على مستويات وفرة أربعة جينات معبرة بشكل مختلف (MMP1، SERPINE1، COL5A2، وFAP). تم استخدام اختبار لوغ-رانك لتحديد الأهمية.
الشكل 2 | تحليل النسخ الجيني للمضيف في سرطان الخلايا الحرشفية الفموي بواسطة تسلسل lncRNA. أ رسم بركاني يظهر الجينات المعبر عنها بشكل مختلف (DEGs؛ abs( ) بين ورم OSCC و الأنسجة. تشير النقاط الحمراء والزرقاء إلى الجينات المعبر عنها بشكل مرتفع ومنخفض في الأورام عند مقارنتها بالأنسجة الطبيعية، على التوالي. ب تعبير عن أربعة جينات تمثيلية (LAMC2، MMP1، EMP1، وCRISP2) المشاركة في تكوين أورام OSCC. ج أفضل 10 جينات من موسوعة كيوتو للجينات و
أداة لتصنيف المخاطر وتخطيط العلاج الشخصي في سرطان الفم والبلعوم.

التفاعلات بين الميكروبات الفموية وترانسكريبتوم المضيف في سرطان الفم الحرشفي

في دراستنا، نهدف إلى استكشاف الروابط المحتملة بين البكتيريا المسببة للأمراض والترنسكريبتوم في سرطان الفم والبلعوم (OSCC) من خلال التحقيق في الارتباطات بين الجينات المعبر عنها بشكل مختلف (DEGs) والوفرة النسبية للأنواع البكتيرية الغنية بالأورام في بيئة الورم الخاصة بـ OSCC. باستخدام تحليل الارتباطات باستخدام معامل سبيرمان، لاحظنا وجود ارتباطات بين 814 جين مضيف (370 مرتبط بالورم و444 مرتبطًا بـ AN) و7 أنواع بكتيرية غنية بالأورام التي وُجدت بشكل متكرر في 34 نسيج ورمي مع بيانات تسلسل 16 S البكتيري وlncRNA المضيف المتاحة، مما أسفر عن 990 ارتباطًا إحصائيًا ذا دلالة بين البكتيريا والترنسكريبتوم. ) (الشكل 3أ والجدول التكميلي 9أ). من الجدير بالذكر أن من بين أهم هذه الارتباطات ( ) ، عرضت أنماطًا متماسكة؛ أي، ارتباط إيجابي بين البكتيريا الغنية بالأورام والجينات المعبر عنها بشكل مرتفع (مرتفع وإيجابي) أو ارتباط سلبي مع الجينات المعبر عنها بشكل منخفض (منخفض وسلبي) (الشكل 3ب، ج). على سبيل المثال، لاحظنا ارتباطًا إيجابيًا كبيرًا بين وفرة C. morbi وتعبير HSPH1 (المعني في زيادة تنظيم مسار إشارة Wnt المتعلق بالتكاثر الخلوي والهجرة) (رو) ) (الشكل 3د). وبالمثل، وُجدت علاقة إيجابية بين وفرة السالمونيلا وتعبير GALNT6 (المعني في تقدم الورم والنقائل) (رو) ). على العكس من ذلك، فإن جينات مثبطات الأورام مثل PRKN (رو) ) و ” (رو) ) أظهر ارتباطًا سلبيًا مع وفرة C. morbi. بالإضافة إلى ذلك، أظهر COLGALT2، وهو جين تم تقليله في سرطان الثدي، ارتباطًا سلبيًا مع عدة أنواع بكتيرية بما في ذلك C. morbi (rho ) و T. متوسط (رو . ومن الجدير بالذكر أيضًا العلاقة الإيجابية الملحوظة بين C. morbi و T. medium في أورام OSCC التي تم مسحها (ارتباط SparCC ) (الشكل 3ج والجدول التكميلي 9ب). نود أن نوضح أن تحديد هذه الارتباطات قد يساهم في فهم أفضل للتفاعلات المعقدة بين الميكروبيوتا واستجابة المضيف، ولكنها تتطلب دراسات وظيفية لاحقة لتوضيح أي علاقات سببية محتملة. نظرًا للقيود الجوهرية في الدقة التصنيفية على مستوى الأنواع مع منطقة V3-V4 من جين 16S rRNA، قمنا أيضًا بإجراء تحليلات ارتباط على مستوى الجنس (الجدول التكميلي 10أ، ب). تعتبر هذه التحليلات أساسًا لفرضيات أكثر استهدافًا يمكن أن تُفيد العمل التجريبي المستقبلي المصمم لاستكشاف تعقيدات تفاعلات المضيف والميكروبيوتا في سرطان الفم والبلعوم.
تم تنظيم الجينات المعبر عنها في المضيف بشكل إيجابي المرتبطة بأنواع البكتيريا الغنية بالأورام (مرتفع وإيجابي، N = 206، تم تلخيص ( ) لعمليات البيولوجيا في علم الأحياء الجيني (GO) لتوصيف إمكانيات البكتيريا المسببة للأمراض في مسببات سرطان الفم والبلعوم (OSCC) (الجدول التكميلي 11a). كما هو موضح في الشكل 3e، كانت المصطلحات الأكثر غنىً في GO تشمل دورة الخلية، الحركة، الالتصاق، التكاثر والهجرة. وكان هذا متسقًا مع نتائج إثراء علم الأحياء الجيني (GO) التي تم الحصول عليها على مستوى الجنس (الجدول التكميلي 11b). على سبيل المثال، كان CDK6 مرتبطًا بشكل إيجابي مع T. medium؛ كان هذا الجين قادرًا على فسفرة الورم الشبكي (Rb) في مرحلة G1 من دورة الخلية عن طريق إزالة كبت E2F لتعزيز التكاثر الخلوي. وبالمثل، تم التعبير عن عدة جينات ترمز للبروتينات المشاركة في تساقط الأورام، الالتصاق والهجرة، مثل الكولاجينات (COL4A5، COL4A6)، الإنتغرينات (ITGB4) واللامينات (LAMA3)، بشكل مفرط وكانت مرتبطة إيجابيًا مع البكتيريا المسببة للأمراض في بيئة ورم OSCC.

اختلال التعبير الجيني بسبب الميثيلation الحمض النووي

تمت دراسة أنسجة الورم وأنسجة AN المأخوذة من 38 مريضًا عشوائيًا مصابًا بسرطان الفم (OSCC) من حيث ميثيلation DNA CpG (الجدولان التكميلان 2 و 12). بلغ إجمالي مواقع CpG 4,584,317 مع تمت caracterization لقراءات كل عينة، التي كانت تقع ضمن مناطق المحفز (الشكل التكميلي 4a). قمنا بت quantifying إشارات الميثيل على كل كروموسوم باستخدام نافذة منزلقة بحجم 1 كيلوبايت لتنعيم التوزيع، مما ميز بوضوح الأورام التي تم مسحها عن أنسجة AN من خلال تحليل المكونات الرئيسية.
(تحليل المكونات الرئيسية، ) (الشكل التوضيحي التكميلي 4ب). ركزنا على المناطق الميثيلية المختلفة (DMRs) المرتبطة بمناطق تعزيز الجينات (حتى 3 كيلو بايت) بسبب ارتباطها بتعبير الجينات. بشكل عام، تم تحديد 17,056 منطقة ميثيلية مرتفعة (زيادة في ميثيل CpG، meth.diff > 1.0) و26,474 منطقة ميثيلية منخفضة (انخفاض في ميثيل CpG، meth.diff تم تحديد مناطق المحفزات التي تنظم 9052 و 13,087 جينًا، على التوالي، عند مقارنة أورام OSCC مع أنسجة AN. ) (الشكل التكميلي 4c والجدول التكميلي 13). تضمنت الجينات المفرطة الميثلة أهدافًا تم الإبلاغ عنها سابقًا من تكرار الميثلة المفرطة في سرطان الفم والبلعوم (OSCC)، مثل DDAH2 وCCNA1 وDCC، بالإضافة إلى بعض الجينات المرتبطة بالسرطان بما في ذلك HRAS (الشكل التكميلي 4d). وبالمثل، تضمنت أمثلة على الجينات ناقصة الميثلة PI3 وAIM2 وPTHLH وIFNG وCEACAM1. قمنا بتحديد 477 جينًا تم قمعه و636 جينًا تم التعبير عنه بشكل مفرط (DEGs) مع ميثلة مفرطة (Hyper-Down) وناقص الميثلة (Hypo-Up) في مناطق المحفز الخاصة بهم، نظرًا لإمكانية اختلال التعبير الجيني المضيف بسبب التعديلات الوراثية (الشكل التكميلي 4e وf والجدول التكميلي 14).

الانحرافات الجينية المرتبطة بالبكتيريا في تنظيم الجينات المضيفة في سرطان الفم والبلعوم

قد يؤدي إثراء البكتيريا المسببة للأمراض في بيئة الورم في سرطان الفم والبلعوم إلى عدم تنظيم التعديلات الجينية. لاختبار هذه الفرضية، قمنا بإنشاء ارتباط بين مناطق التعديل الجيني المختلفة (DMRs) وأنواع البكتيريا الغنية في الورم باستخدام ارتباط سبيرمان (الجدول التكميلي 15a). بشكل عام، شكلت 13,172 منطقة تعديل جيني (مسؤولة عن 8690 جينًا) و7 أنواع بكتيريا غنية في الورم 16,630 زوجًا مرتبطًا بالبكتيريا والميثيل. )، حيث تساهم P. simiae و F. nucleatum في ( ) من الارتباطات الإيجابية مع فرط الميثلة (Hyper & Pos) و 50.4% (1071/2125) من الارتباطات السلبية مع نقص الميثلة (Hypo & Neg)، على التوالي (الشكل 4a، b). تحليل تكاملي لارتباطات البكتيريا-الميثلة وارتباطات ترنسكريبتوم البكتيريا حدد أيضًا 15 جينًا معبرًا عنه بشكل مكبوت قد يتم كتمه بواسطة فرط الميثلة المرتبط بالبكتيريا في مناطق المحفز الخاصة بها (الشكل 4c والجدول التكميلي 15b). على سبيل المثال، كان تركيز C. morbi مرتبطًا في الوقت نفسه مع فرط الميثلة (Hyper & Pos) وتثبيط التعبير الجيني (Down & Neg) لجين AOX1 و PITX1 (الشكل 4d). تشمل أمثلة أخرى من مثبطات الأورام DHRS3 و CES1. في الوقت نفسه، وُجدت ارتباطات عكسية بين NRG1 و ITGB4 بين نقص الميثلة في DNA CpG وزيادة التعبير الجيني، والتي كانت مرتبطة بوفرة T. medium و C. morbi، على التوالي (الشكل 4e)، مما يوحي بإمكانية البكتيريا الغنية بالأورام لتغيير ميثلة DNA التي قد تترجم إلى تنشيط التعبير الجيني. يرتبط NRG1 كجين مسرطن بأعضاء عائلة مستقبلات التيروسين كيناز ErbB، مما يؤدي إلى تفعيل مسارات الإشارة السفلية، مثل مسارات PI3K/AKT و MAPK/ERK، التي تشارك في نمو الخلايا وبقائها وتكاثرها. يتم التعبير عن ITGB4 بشكل غير طبيعي في عدة أنواع من السرطان بما في ذلك سرطان الثدي وسرطان القولون وسرطان الرئة. تم تأسيس ارتباطات البكتيريا-الميثلة على مستوى الجنس وعرضها في الجدول التكميلي. .
لفهم أفضل لإمكانات F. nucleatum في مسببات سرطان الفم والبلعوم الحرشفي (OSCC) من خلال تغيير تنظيم جينات المضيف، قمنا أولاً بتطبيق 384 جينًا مختلفًا تم تنظيمه بشكل إيجابي مرتبطًا بـ 7 أنواع بكتيرية غنية بالأورام (مرتفع وإيجابي، ” ) لتحليل إثراء مجموعة الجينات (GSEA) (الجدول التكميلي 17a). تم ملاحظة 368 اتصالًا (الشكل التكميلي 5 والجدول التكميلي 17b)، حيث تفاعل F. nucleatum مع 20 جينًا مختلفًا تم تنظيمه بشكل إيجابي (DEGs) والتي شملت 18 مسارًا رئيسيًا (الشكل 5a). من بينها، على سبيل المثال، SNAI2 ( )، التي قد يتم تحفيزها بواسطة F. nucleatum و C. morbi، هي واحدة من العوامل الرئيسية في تنظيم عملية الانتقال الظهاري-الم mesenchymal (EMT). تظهر الخلايا التي تمر بعملية EMT فقدان علامات الظهارة مثل E-cadherin وclaudins، مما يؤدي إلى انفصالها عن الورم الأساسي، وغزو الأنسجة المحيطة، وفي النهاية الانتقال إلى مواقع بعيدة. حيث أن F. nucleatum و F. periodonticum، النوعان اللذان لهما علاقة تطورية وثيقة ووظائف بيولوجية مشابهة، يمثلان و من
قراءات 16S rRNA لـ Fusobacterium، قمنا أيضًا بإجراء تحليل GSEA مستند إلى اتصالات النسخ الجيني للبكتيريا على مستوى الجنس ولاحظنا 40 جينًا معبرًا عنه بشكل مفرط (DEGs) تتضمن 22 مسارًا مميزًا مسرطنًا تم تعزيزها والتي كانت مرتبطة بشكل إيجابي مع Fusobacterium (الجدول التكميلي 17c، d)، بما في ذلك ستة جينات مميزة (LAMA3، INHBA، SNAI2، NT5E، MYLK وTPM1) ضمن عملية EMT (الشكل التكميلي 6).
ثم قمنا بتطبيق نموذج خلوى قائم على الخلايا في المختبر من خلال زراعة F. nucleatum مع خلايا الظهارة الفموية البشرية السرطانية (SAS) وغير السرطانية (HGK12) على التوالي، لتوصيف ترانسكريبتومات الجينات الخلوية باستخدام تسلسل lncRNA (الشكل 5ب والجدول التكميلي 18). أظهرت هاتان الخلايتان اختلافات.
الشكل 3 | التفاعلات بين الأنواع البكتيرية الغنية بالأورام وDEGs المضيفة المرتبطة بـ OSCC. أ رسم بياني عمودي يوضح ارتباط سبيرمان بين البكتيريا والترنسكريبتوم، يظهر وفرة DEGs المضيفة في أنسجة أورام OSCC المرتبطة بالأنواع البكتيرية الغنية بالأورام. ). تشير Up و Down إلى الجينات المعبر عنها بشكل مرتفع ومنخفض، على التوالي؛ تشير Pos و Neg إلى الارتباطات الإيجابية والسلبية، على التوالي. C. morbi Catonella morbi، G. morbillorum Gemella morbillorum، P. yurli Peptoanaerobacter yurli، P. simiae Peptococcus simiae، P. stomatis Peptostreptococcus stomatis، F. nucleatum Fusobacterium nucleatum، T. medium Treponema medium. b خريطة حرارية توضح الارتباطات العالية الأهمية بين البكتيريا والترانسكريبتوم ( ). تشير لوحة الألوان إلى معاملات ارتباط سبيرمان. تشير النجوم إلى مستوى الدلالة (** ; ). يتم ترتيب الأعمدة حسب متوسط معاملات الارتباط للجينات المعبر عنها بشكل مختلف مع البكتيريا الغنية بالأورام.
شبكة تصور الارتباطات المهمة بين البكتيريا والترانسكريبتوم (خط صلب، سبيرمان ) والبكتيريا-البكتيريا (الخط المتقطع، ارتباط SparCC ” ). د. مخططات التشتت تظهر أربعة ارتباطات تمثيلية بين البكتيريا و DEG في أنسجة أورام OSCC، حيث قوة الارتباط (معامل سبيرمان) والأهمية ( القيمة) موضحة في كل رسم بياني. تُظهر الرسوم البيانية الهامشية التعبير الجيني العام (يمين) ووفرة جين 16S rRNA البكتيري (أعلى)، مع اختبار مان-ويتني اختبار بين الأنسجة الورمية وأنسجة AN. المسارات في علم الأحياء الجزيئي (GO) التي تم إثراؤها بشكل كبير بواسطة الجينات المعبر عنها بشكل مرتفع المرتبطة بالبكتيريا الغنية بالورم. يتم تجميع مصطلحات GO ذات التشابه الوظيفي في الفضاء الدلالي بواسطة REVIGO. تم تصحيح B & H. تُشير القيم إلى لوحة الألوان. حجم الدوائر يدل على عدد الجينات المعبر عنها بشكل مختلف المرتبطة بالممرات.
تعبيرات النسخ المعبر عنها، مع 820 جينًا معبرًا عنه بشكل مرتفع و478 جينًا معبرًا عنه بشكل منخفض تم اكتشافها بشكل شائع (الشكل التوضيحي 7a والجدول التوضيحي 19). بينما كانت الجينات النشطة في HGK12 مرتبطة باستجابة المناعة التي تتوسطها الخلايا المتعادلة، ومسار الإشارات الذي تتوسطه السيتوكينات، وتنظيم الالتهام الذاتي، كانت تعبيرات الجينات في SAS المصابة بـ F. nucleatum غنية بشكل رئيسي في تنظيم نقل الإشارات، وموت الخلايا المبرمج، وعملية الاستماتة (الشكل التوضيحي 7b والجدول التوضيحي 20 و21)، وهو ما قد يكون جزئيًا بسبب أن هذين الخلايا لهما خصائص مختلفة تمامًا. ومن المثير للاهتمام، أنه من بين 148 جينًا مختلفًا تم ربطها بشكل متماسك بقراءات 16S لـ F. nucleatum في أنسجة الأورام OSCC التي تم مسحها (Up & Pos و Down & Neg، على الأقل تم تأكيد أن 15 و 10 جينات كانت مفرطة التعبير أو مثبطة في خطي الخلايا المزروعة معًا، على التوالي. ، ) (الشكل 5ب، ج والجدول التكميلي 19). تشمل هذه التعبيرات المفرطة للجينات المميزة SNAI2 وLIMA1 وPLEK2 وRAB31 المشاركة في مسارات GSEA المتعددة (الشكل 5ج، د). في المقابل، كان SLC16A7، وهو جين يشفر ناقلًا للحلائل عبر الغشاء (عضو من العائلة 16، العضو 7) وعادة ما يكون منخفضًا في أنواع مختلفة من السرطان، مرتبطًا سلبًا بـ . نوكليتوم (رو ) وتم قمعها بشكل كبير في الموقع ( ) وفي المختبر ( HGK 12 : ; ساس: ). باستخدام تقنية PCR في الوقت الحقيقي، تحققنا من فرط التعبير عن 5 جينات بارزة تمثل علامات مرتبطة بمسار EMT في المختبر، بما في ذلك SNAI2 (مرتبط بـ F. nucleatum)، INHBA و LAMA3 (مرتبطان بـ Fusobacterium)، و LAMC2 (الشكل 5e والجدول التكميلي 23)، على الرغم من أن مستويات بروتيناتها، كما تم قياسها بواسطة التحليل الغربي، كانت مثيرة للجدل (الشكل التكميلي 7c والجدول التكميلي 24). تشير نتائجنا في المختبر إلى زيادة كبيرة في مستويات التعبير عن mRNA لثلاثة جينات بارزة في EMT (LAMA3، INHBA، SNAI2) عند الإصابة بـ F. nucleatum. بالإضافة إلى ذلك، قمنا بتضمين LAMC2، الجين المشفر للبروتين الفرعي غاما من Laminin332، والذي هو جزء من نفس المجمع البروتيني مثل LAMA3، كدليل تكميلي. كما بحثنا في بروتين laminin332 ووجدنا أن تعبيرات البروتين لاثنين من الوحدات الفرعية قد زادت بشكل محتمل بواسطة عدوى النوكليتوم في كل من خطوط خلايا HGK12 و SAS. تتماشى هذه الملاحظة مع النتائج التي تم الحصول عليها من تحليل RNAseq لخطوط الخلايا المترابطة، ولكن الآليات الأساسية لمسارات التسرطن التي تشارك فيها هذه الجينات في OSCC تتطلب مزيدًا من التحقيق.

نقاش

تم الإبلاغ عن عدم التكيف بين المضيف والميكروبيوتا في أنواع مختلفة من السرطان كشفت دراستنا من خلال التحليل التكاملي للترانسكريبتوم البكتيريا والارتباطات بين الميثيل البكتيري في مرضى سرطان الفم الحرشفي غير الإيجابي لفيروس الورم الحليمي البشري عن شبكات معقدة بين اختلال ميكروبيوتا الفم والعيوب الجينية والوراثية لدى المضيف، والتي لها دور وظيفي في مسارات مرتبطة بالسرطان. تشير بياناتنا أيضًا إلى أن عدم تنظيم ترانسكريبتوم الجينات لدى المضيف قد يتأثر بالبكتيريا الغنية بالأورام أو بالتعديلات الوراثية المرتبطة بالبكتيريا. كما أكد نموذج في المختبر أن F. nucleatum، وهو نوع من البكتيريا الفموية المرتبطة ارتباطًا وثيقًا بتكون أورام سرطان الفم الحرشفي، يمكن أن ينشط الجينات الرئيسية المعنية في مسارات متعددة مرتبطة بالسرطان. قد تخدم هذه النتائج
كنقطة انطلاق لدراسة مدفوعة بالفرضيات لفهم أفضل للآليات الجزيئية للبكتيريا المسببة للأمراض التي تكمن وراء نشوء سرطان الفم.
قد تنشأ السرطنة من التغيرات في التفاعلات المزمنة بين المضيف والميكروبات عند استعمارها بواسطة مسببات الأمراض الرئيسية. أظهرت تحليل بياناتنا متعددة الأوميات العديد من المسارات المرتبطة بالسرطان التي قد يتم تحفيزها بشكل مباشر أو غير مباشر بواسطة البكتيريا الفموية الغنية بالأورام في بيئة الورم في سرطان الفم الحرشفي. على سبيل المثال، أشار تحليلنا التكاملي في الموقع وفي المختبر إلى الدور التنظيمي لبكتيريا F. nucleatum على SNAI2 في سرطان الفم الحرشفي، بما يتماشى مع دور عامل الضراوة FadA في النوكليتوم لتعديل الإيكاديرين/ إشارات -كاتينين من خلال تنشيط TNF مسار الالتهاب، إعادة تشكيل المصفوفة خارج الخلوية، وإشارات Wnt (الشكل 6) اللامينين هو المكون الرئيسي للمصفوفة خارج الخلوية (ECM) ويلعب دورًا رئيسيًا في التصاق الخلايا، وهجرة الخلايا، ونقل الإشارات. ، من بينها Laminin-332 هو عضو رئيسي في عائلة اللامينين ويحتوي على ثلاث سلاسل مشفرة بواسطة LAMA3 و LAMB3 و LAMC2. في الوقت نفسه، يمكن أن يتم نقل -catenin إلى النواة لتشكيل معقد مع بروتين عامل T-cell (TCF) وتنشيط الجينات المرتبطة بعملية التحول الظهاري (EMT)، مثل INHBA وLAMA3 وLAMC2 وNT5E وMMP1، مما يساهم في الاستجابات الالتهابية والسرطانية. تظهر الخلايا التي تمر بعملية التحول الظهاري فقدان العلامات الظهارية مثل الأوكليودينات والكلودينات وE-cadherin، وتكتسب في الوقت نفسه تعبيرًا متزامنًا عن Vimentin وN-cadherin وFibronectin. حيث يلعب بروتين SLUG (SNAI2) دورًا أساسيًا من خلال قمع عدة جينات لالتصاق الخلايا بما في ذلك تعبير E-cadherin بالإضافة إلى ذلك، السيتوكين المؤيد للالتهابات يعمل كوسيط التهابي لتحفيز التحول الظهاري للخلية السرطانية وتعزيز الانبثاث. . يمكن أن يحفز أيضًا استقرار عائلة SNAI من خلال تنشيط NF- مسار .
بالإضافة إلى F. nucleatum، يمكن أن تعزز عدة بكتيريا فموية أخرى، مثل Treponema denticola، عدوانية السرطان من خلال التفاعل بين مسارات الإشارات integrin/FAK و TLR/MyD88. تم الإبلاغ عن أن إشارات PI3K/Akt المرتبطة بالإنترجرين يتم تنشيطها بواسطة Peptostreptococcus anaerobius، الذي ينظم تقدم دورة الخلية. تم العثور على كلا النوعين البكتيريين في أنسجة أورام OSCC التي تم مسحها مع دلالة إحصائية محدودة. ومن المثير للاهتمام، أن T. medium و P. stomatis، وهما نوعان بكتيريّان مرتبطان من الناحية النشوء والتطور بـ . دينتيكولا و . الأنيروبيوس، على التوالي، زاد بشكل ملحوظ في الوفرة النسبية في بيئة الورم في سرطان الفم. تم ربط P. ستوماتيس بمختلف الأمراض الفموية بما في ذلك السرطان. . على سبيل المثال، مرتبط بـ . يمكن أن تنشط السالمونيلا المسار الكلاسيكي Wnt من خلال -إشارات بيتا-كاتينين/ MMP9 التي تنظم تكاثر الخلايا، التمايز و EMT تمت الإشارة أيضًا إلى ثلاثة جينات بارزة مرتبطة بعملية التحول الظهاري (EMT)، بما في ذلك GJA1 وMMP1 وSNAI2، والتي لديها القدرة على تعزيز هجرة الخلايا السرطانية وغزوها، وكانت مرتبطة بشكل إيجابي مع . التهاب الفم في أنسجة أورام OSCC التي تم مسحها من المهم أن نلاحظ أن عدم تنظيم تعبير الجينات المضيفة قد يتأثر بمساهمات من عدة بكتيريا ممرضة. تم تحديد LIMA1 (الذي ينتمي إلى الوصلات القمية) في البداية كجين معبر عنه بشكل مختلف في سرطان خلايا الظهارة الفموية. على سبيل المثال، كانت مرتبطة بشكل إيجابي مع C. morbi و F. nucleatum و . التهاب الفم، مما يبرز تعقيد تفاعلات المضيف والميكروبيوتا في بيئة الورم لسرطان الفم والبلعوم. تأثيرات

تعبير النسخ الجيني (DEG) ممثل بدوائر مملوءة، مع أحجام تتناسب مع القيم. يوضح الرسم البياني العمودي معاملات ارتباط سبيرمان لمثيلة CpG وتعبير النسخ مع وفرة البكتيريا الغنية بالأورام. د الرسوم البيانية المبعثرة التي تصور فرط مثيلة AOX1 و PITX1 (اللوحة العلوية) وتعبير منخفض (اللوحة السفلية) فيما يتعلق بـ C. morbi، حيث قوة الارتباط (سبيرمان رو) والأهمية ( القيم) موضحة في كل رسم بياني. تُظهر الرسوم البيانية الهامشية التعبير الجيني/الميثيل (يمين) ووفرة جين 16S rRNA البكتيري (أعلى)، مع اختبار مان-ويتني اختبار بين الأنسجة الورمية وأنسجة AN. تمثيلات بيانية توضح نقص الميثلة في NRG1 وITGB4 (اللوحة العلوية) وزيادة التعبير (اللوحة السفلية) فيما يتعلق بـ T. medium وC. morbi، على التوالي.
الشكل 5 | الإمكانات المسرطنة المرتبطة بسرطان الفم والبلعوم (OSCC) لبكتيريا Fusobacterium nucleatum في الموقع وفي المختبر. أ تحليل إثراء مجموعة الجينات (GSEA) مستنتج من 384 جينًا معبرًا عنها بشكل مرتفع مرتبطًا إيجابيًا بـ 7 أنواع بكتيرية غنية بالأورام (مرتفع وإيجابي، ). تعرض الشبكة 20 جينًا معبرًا مختلفًا تشمل 18 مسارًا رئيسيًا مرتبطًا بـ F. nucleatum. ب رسم توضيحي تخطيطي لنموذج خلية في المختبر من خلال زراعة F. nucleatum مع خلايا الظهارة الفموية البشرية HGK12 و SAS لتوصيف النسخ الجيني الخلوي باستخدام تسلسل lncRNA. عدد الجينات المعبر عنها بشكل مختلف في الموقع (abs( ) مع ارتباط متماسك مع . قراءات 16S لـ nucleatum (إيجابي للأعلى وسلبي للأسفل، ); عدد الجينات المعبر عنها بشكل مختلف في الموقع مع تنظيم غير متماسك في خط خلوى واحد على الأقل (abs( ) وتم تأكيده بواسطة خط خلوى آخر (abs(log2FC ). ج . تم ملاحظة الجينات المعبر عنها بشكل مختلف المرتبطة بالنيوكليوتوم في نفس الوقت في الموقع وفي المختبر. تشير الدائرة المملوءة إلى حالة تعبير النسخ (DEG) في المختبر وفي الموقع، على التوالي، مع الأحجام
المقابل لـ القيم. يوضح الرسم البياني العمودي معاملات ارتباط سبيرمان بين تعبير النسخ الجيني للـ DEG و وفرة النوكليتوم في أنسجة أورام OSCC. خمسة تُبرز الجينات الرئيسية المرتبطة بـ F. nucleatum بالخط العريض. تُظهر الرسوم البيانية المبعثرة زيادة تنظيم SNAI2 و LIMA1 فيما يتعلق بـ F. nucleatum في أنسجة أورام OSCC، حيث تُظهر قوة الارتباط (معامل سبيرمان) والأهمية ( القيمة) موضحة في كل رسم بياني. تُظهر الرسوم البيانية الهامشية التعبير الجيني العام (يمين) ووفرة جين 16S rRNA البكتيري (أعلى)، مع اختبار مان-ويتني اختبار بين الأنسجة الورمية وأنسجة AN. تحليل RT-PCR للتعبير المفرط المحدد لأربعة جينات رئيسية في GSEA (SNAI2، INHBA، LAMA3 و LAMC2) في خلايا HGK12 و SAS التي تم زراعتها مع F. nucleatum أو F. mortiferum. يتم تقديم البيانات كمتوسط. SD والأهمية ( تحدد بواسطة اختبار الطالب غير المتزاوج ذو الطرفين اختبار قائم على .
الشكل 6 | نموذج تخطيطي يفترض الإمكانية المسرطنة للبكتيريا الممرضة في نشوء سرطان الفم والبلعوم. تم الإبلاغ عن أن عامل الفوعة FadA لبكتيريا F. nucleatum ينشط E-الكاديرين. إشارات الخلايا المرتبطة بـ -كاتينين، والتي تنشط بدورها TNF مسار الالتهاب، إعادة تشكيل المصفوفة خارج الخلوية، وإشارات Wnt. بعد ذلك، يمكن نقل -catenin إلى النواة لتشكيل معقد مع بروتين عامل خلايا T (TCF) وتنشيط الجينات المرتبطة بعملية التحول الظهاري، مثل SNAI2 و LAMA3 و INHBA و MMP1. علاوة على ذلك، فإن F. nucleatum و T. denticola قد
تم الإبلاغ عن تعزيزها لعدوانية السرطان من خلال التفاعل بين مسارات إشارات الإنتجرين/FAK وTLR/MyD88. يمكن تنشيط إشارات PI3K/Akt المرتبطة بالإنتجرين بواسطة . الأنيروبيوس، الذي ينظم تقدم دورة الخلية. غال-غالناكس على خلايا الورم هو مستقبل فاب2 لجذب ف. نوكليتوم إلى موقع الورم. بالإضافة إلى ذلك، قد يؤدي نقص الميثلة في NRG1 إلى تنشيط مسار PI3K/AKT بينما قد تعزز زيادة الميثلة في AOX1 الغزو الخلوي والنقائل.
يمكن أن يمتد LIMA1 من هجرة الخلايا وديناميات الهيكل الخلوي إلى دورة الخلية، وتنظيم الجينات، وتكوين الأوعية الدموية، وعمليات الأيض الدهني، مما يوفر أفكارًا جديدة لاستكشاف استراتيجيات علاج السرطان المستهدفة لهذا الجين في المستقبل. .
تشير الدراسات الحديثة إلى أن التغيرات الوراثية فوق الجينية قد تلعب أدوارًا مهمة في بدء وانتشار السرطان. تم التعرف أيضًا على الآليات الوراثية اللاجينية كلاعب حاسم عند التفاعل بين الميكروبيوم البشري وخلايا الظهارة المعوية. على سبيل المثال، أدى التعرض للميكروبات المعوية المفيدة إلى تغييرات موضعية في ميثلة الحمض النووي، وهو أمر ضروري للحفاظ على توازن الأمعاء السليم. على النقيض من ذلك، فإن F. nucleatum و H. hathewayi، وهما من مسببات الأمراض البكتيرية المرتبطة بسرطان القولون، قادران على دفع فرط ميثلة المحفزات الجينية لمثبطات الأورام في خلايا الظهارة القولونية للمضيف. تشير بياناتنا إلى أن عدم تنظيم النسخ الجيني للمضيف قد يتأثر بالخلل الجيني المرتبط بالبكتيريا. على سبيل المثال، تم ربط فرط الميثلة وانخفاض تنظيم PITX1 و AOX1 بكل من C. morbi. تم تحديد PITX1 كجين محتمل مثبط للورم مرتبط بموت الخلايا وتم تقليله بواسطة فرط الميثلة في الحمض النووي في سرطان المعدة. وسرطان الخلايا الحرشفية في المريء (ESCC) . يقوم PITX1 بتثبيط القدرة على تكوين الأورام من خلال تقليل نشاط مسار RAS عبر RASAL1، وهو بروتين منشط لـ RAS-GTPase . بالمثل، فقدان ربما يرتبط بمروج مفرط الميثلة، ويساهم في الانتقال من الدرجة المنخفضة إلى الدرجة العالية خلال تقدم سرطان المثانة، مما يعزز الغزو والنقائل. . في الوقت نفسه، تشير انخفاض الميثلة في عدة جينات سرطانية مثل NRG1 وITGB4 وGALNT6 المرتبطة بالبكتيريا الغنية بالأورام إلى إمكانية الميكروبيوتا في تعزيز تعبير الجينات المضيفة من خلال التنشيط الوراثي. على سبيل المثال، لقد أظهر GALNT6 أنه يعزز تكوين الأورام والانتقال من خلال تحفيز استقرار MUCl والفيبرونكتين (FN) بواسطة جليكوزيل-أو من نوع الميوسين في خلايا سرطان الثدي. بالإضافة إلى ذلك، أفادت دراسة حديثة أن . يمكن أن يؤدي النوكلاتوم إلى انخفاض كبير في تعديلات في CRC تؤدي إلى تحسين شدة CRC لقد تم إثبات أن انتشار سرطان القولون والمستقيم يتم الترويج له بواسطة . نيكولاتوم من خلال محور miR-1322/CCL20 واستقطاب M2، الذي
مؤشرًا على أهمية تفاعلات المضيف والميكروبيوم ومع ذلك، لا يزال يتعين توضيح الآليات التي تؤدي من خلالها ميثلة محفزات الجينات المبرمجة بواسطة الميكروبات الفموية إلى التعبير غير المنظم في سرطان الفم الحرشفي.
لدراستنا عدة قيود. أولاً، لم تشمل خزعات تجويف الفم من المشاركين الأصحاء. على الرغم من أن أنسجة AN من مرضى OSCC تعتبر بمثابة ضوابط جيدة، فإن التأثيرات المشتركة للتعرضات البيئية والمتغيرات السريرية بين مرضى السرطان والأفراد الأصحاء من الصعب حسابها بسبب التداخل على المدى الطويل. أيضًا، لم تكن بعض نتائج الفحوصات المتعلقة بالحالة الفموية متاحة، مما قد يغفل التأثير التآزري لبعض العوامل المربكة على اختلال الميكروبات الفموية ويحتاج إلى معالجة في الدراسات اللاحقة. ثانيًا، يعتمد التقييم التنبؤي لمرض OSCC على تصنيف TNM السريري، لكن هذا النظام التصنيفي غير كافٍ للتنبؤ الأمثل وقد يحتاج إلى دعم بأساليب أخرى مثل التصنيف النسيجي. ثالثًا، تفتقر هذه الدراسة إلى رابط مباشر بين ترانسكريبتوم الجينات المضيفة وميثيلين DNA CpG حيث كانت هناك عشرة أورام متداخلة فقط متاحة، مما يتطلب نظرة أكثر شمولاً على البرمجة الجينية المستندة إلى مجموعة أكبر. ومع ذلك، من المثير للاهتمام ملاحظة الارتباطات المتسقة من خلال التحليل التكاملي لارتباطات البكتيريا-ترانسكريبتوم وارتباطات البكتيريا-ميثيلين. خاصة، يمكن التحقق من اختلال تنظيم العديد من الجينات الرئيسية المرتبطة بـ F. nucleatum التي تشمل مسارات EMT بشكل جيد في المختبر باستخدام lncRNAseq وRT-PCR وwestern blot. رابعًا، تكشف دراستنا عن ارتباط ولكن ليس سببية الميكروبات الفموية في مسببات OSCC؛ ستكون الدراسات الإضافية باستخدام نماذج in vivo وin vitro مفيدة لتحديد الروابط المحددة بين الميكروبات والمضيف التي تكمن وراء الآلية. وأخيرًا وليس آخرًا، قد تكون دقة تعريف الأنواع البكتيرية المستنتجة من قراءات منطقة 16S rRNA V3-V4 محدودة ولا يمكن أخذ سلالات البكتيريا في الاعتبار في الدراسة الحالية. ومع ذلك، نقدم أكثر تسلسلات ASV تمثيلاً ونشجع على مزيد من التحليل بناءً على مستويات مختلفة من التصنيف الضريبي بالإضافة إلى الثقافة البكتيرية التقليدية.
باختصار، قمنا بتطبيق أساليب متعددة الأوميات لكشف الشبكات المعقدة بين ميكروبات الغشاء المخاطي الفموي واضطراب جينات المضيف في مسببات سرطان الفم الحرشفي. وجدنا بكتيريا فموية ذات وفرة متفاوتة.
تعبير الجينات المضيفة غير المنظم، وميثيل الحمض النووي CpG الشاذ في بيئة الورم، مع وظائف غنية تتعلق بمسارات السرطان المختلفة. يكشف التحليل التكاملي بين ترانسكريبتوم البكتيريا وارتباطات ميثيل البكتيريا أن عدم تنظيم ترانسكريبتوم الجينات المضيفة قد يتأثر بالبكتيريا الغنية في الورم أو بواسطة الشذوذات الجينية المرتبطة بالبكتيريا. تمتد نتائجنا الفهم الحالي لتفاعلات المضيف والميكروبات في مسببات سرطان الفم والبلعوم، مما يوفر رؤى مهمة في الأساس الجيني والوظيفي كعلامات تشخيصية محتملة وأهداف علاجية لوضع استراتيجيات أكثر فعالية للوقاية والتدخل لإدارة هذه الكيان المرضي.

طرق

تجنيد المرضى

تم تجنيد ما مجموعه 98 مريضًا أوليًا بسرطان الفم (OSCC) من عرق الهان الصيني في مستشفيين في هونغ كونغ (مستشفى الأمير ويلز ومستشفى المسيحيين الموحدين) بين أكتوبر 2015 ويناير 2020. تم جمع المعلومات الديموغرافية والسريرية عند التسجيل، بما في ذلك العمر، الجنس، التدخين، استهلاك الكحول، موقع الورم، مرحلة T ومرحلة N، في الجدول التكميلي 2. لم يتم تضمين أي حالات من الانتقال في هذه الدراسة.

موافقة الأخلاقيات

تمت الموافقة على هذه الدراسة من قبل لجنة الأخلاقيات للبحوث السريرية لمجموعة نيو تيريتوريز الشرقية في الجامعة الصينية في هونغ كونغ (أرقام مرجع CREC 2015.396 و 2017.143). وافق جميع المرضى على المشاركة بموجب موافقة مستنيرة مكتوبة وتم مراجعتهم من قبل أخصائي علم الأمراض.

جمع عينات الأنسجة واستخراج الحمض النووي/الحمض النووي الريبي

أنسجة الورم و AN تم جمع العينات (بعيدًا عن هامش الورم) في وقت الجراحة وفقًا لمنهجيتنا السابقة . تم ري هذه العينات لفترة وجيزة بمحلول ملحي معقم لإزالة التلوث السطحي ثم تم تخزينها في حتى إشعار آخر. حول تم تجميد الأنسجة الطازجة يدويًا إلى قطع صغيرة ومعالجتها بـ بروتياز K عند بين عشية وضحاها. تم استخراج الحمض النووي DNA والحمض النووي الريبي RNA من العينات المفككة في نفس الوقت باستخدام مجموعة AllPrep DNA/RNA Mini Kit (Qiagen، فالنسيا، كاليفورنيا، الولايات المتحدة الأمريكية) وتم تصفيتها إلى تم استخدام محلول الإيلاشن بشكل منفصل وفقًا لبروتوكول الشركة المصنعة. تم اختبار الحمض النووي من أنسجة الورم لعدوى فيروس الورم الحليمي البشري باستخدام اختبار تسلسل الأمبليكون القائم على تفاعل البوليميراز المتسلسل كما تم وصفه سابقًا. .

تسلسل الأمبليكون لجين 16S rRNA للميكروبيوتا V3-V4

تم استخدام الحمض النووي المستخرج لتوصيف الميكروبات الفموية عن طريق تسلسل منطقة الجين 16S rRNA البكتيري المتغيرة بشكل كبير V3-V4 (341F: -CCT ACG GGN GGC WGC AG-3’، 806R: -جي جي إيه سي تي إيه سي إن في جي جي جي تي دبليو تي سي تي إيه تم تعيين زوج من رموز الشريط الثنائي الفريد 12 bp لكل مجموعة من الأمبليكون من خلال البرايمرات الأمامية والخلفية؛ وتم تجميع الأمبليكون الناجحة بالتساوي وتسلسلها على جهاز Illumina MiSeq باستخدام قراءات مزدوجة النهاية بطول 300 bp. لأغراض مراقبة الجودة، شمل كل دفعة تسلسل مجتمع DNA وهمي، وضوابط سلبية، وتكرارات تقنية.

تحليل البيانات الحيوية والإحصائية لتسلسل 16S للميكروبيوتا

تم استيراد قراءات 16S القصيرة المفككة التي اجتازت تصفية الجودة إلى QIIME2 (الإصدار 2021.4) لإنشاء جدول متغير تسلسل الأمبليكون (ASV) كما تم وصفه سابقًا. تم تعيين القراءات التمثيلية بشكل إضافي على مستوى الأنواع باستخدام pplacer من خلال إدخالها في شجرة تطورية مستنتجة من تسلسلات كاملة لجين 16S rRNA البكتيري لتعظيم احتماليات التطور. ASVs مع عدد إجمالي بعد إزالة القراءات المخصصة للعتائق، تم الاحتفاظ بالميتوكوندريا أو البلاستيدات الخضراء، مع عرض جداول عدد وحدات التصنيف التشغيلية (OTU) لقراءات البكتيريا لكل عينة على مستويات تصنيفية مختلفة. تم الاحتفاظ بالأنسجة الورمية وأنسجة AN التي تحتوي على أكثر من 2000 قراءة لمزيد من التحليل. تم إنشاء شجرة تطورية من خلال إدخال قراءات ASV التمثيلية في قاعدة بيانات SLIVA 128 المرجعية باستخدام تقنية وضع التطور المدعومة بـ SATe (SEPP).
طريقة تم حساب تنوع ألفا لقراءات ASV البكتيرية الملاحظة بناءً على الغنى، ومؤشرات شانون وسيمبسون باستخدام التنوع في حزمة Vegan R. تم حساب مقاييس المسافة GUniFrac وBray-Curtis المستنتجة من ملف ASV لتمييز تركيبات المجتمع (تنوع بيتا) بين مجموعات السرطان والمجموعة الضابطة باستخدام تحليل التباين المتعدد المتغيرات القائم على التباديل (PERMANOVA) مع 9999 تبديلاً باستخدام adonis2 في حزمة Vegan R. في تحليل حجم التأثير، تم التحكم في حالة المرض (ورم مقابل AN) في العلاقة بين المتغيرات السريرية، بما في ذلك الجنس، العمر، التدخين، استهلاك الكحول، مراحل N و T. تم تقدير الأنواع البكتيرية التمييزية بين أنسجة الورم و AN استنادًا إلى جداول عدد OTU باستخدام تحليل حجم التأثير للتمييز الخطي (LDA) (تحليل LEfSe). ، مع قطع LDA والذي تم التحقق منه بشكل إضافي بواسطة أدوات مختلفة مدركة للتكوين مع تصحيح الانحياز، بما في ذلك ANCOM-BC2 وزيكوسيquence ، مع تعديلها لعوامل T stage والتدخين ( ). بالإضافة إلى ذلك، تم إجراء اختبار تحليل الانحدار الخطي، واختبار مان-ويتني ويلكوكسون غير المعلمي (MWU)، واختبار الفرق المعنوي الصادق لتوكاي (Tukey HSD) بعد الاختبار، ونموذج لوغاريتمي عام ثنائي الحدين في EdgeR تم تطبيقها. تم إنشاء خرائط حرارية لأكثر الأجناس البكتيرية تمييزًا باستخدام heatmap.2 في حزمة gplots R، مع إعداد التجميع الهرمي “hclust(method=”ward.D”) و dist(method= “euclidean”)”. تم استخدام الانحدار اللوجستي ومنحنى خصائص التشغيل المستقبلية (ROC) مع حساب المساحة تحت منحنى ROC (AUC) لتقييم العلامات البيولوجية المحتملة المحددة لفحص حالات السرطان. تم استخدام تحليل كابلان-ماير لتحليل البقاء أحادي المتغير مع اختبار لوغ-رانك لتحديد الفروق الإحصائية في نتائج البقاء. تم حساب نقطة قطع مثالية مستنتجة من الوفرة النسبية للأنواع البكتيرية الفردية في العينات الممسوحة لتقسيم أنسجة الورم إلى مجموعات “عالية” و”منخفضة” باستخدام نص cutpointr R. تم استخدام طريقة انحدار كوكز النسبي للخطر بطريقة خطوة بخطوة مع التحكم في الجنس والعمر وعلاج ما بعد الجراحة وعوامل أخرى تم الإبلاغ عنها بشكل كبير في التحليل أحادي المتغير لتحليل البقاء متعدد المتغيرات. تم استخدام قيمة و/أو معدل الاكتشاف الخاطئ (FDR) المعدل قيمة أو تم استخدامه كعتبة للدلالة على الأهمية.

كشف فيروس الورم الحليمي البشري وتصنيفه الجيني

تم إجراء تصنيف نوع فيروس الورم الحليمي البشري (HPV) باستخدام اختبار تسلسل الأمبليكون القائم على تفاعل البوليميراز المتسلسل (PCR) الذي يستهدف إطار القراءة المفتوح L1 المحفوظ لفيروس HPV كما تم وصفه سابقًا. باختصار، تم فهرسة زوج من رموز الشريط المزدوج بطول 12 قاعدة إلى منتج تفاعل البوليميراز المتسلسل باستخدام بادئات أمامية وخلفية للتفكيك. تم مقارنة القراءات القصيرة التي تم إنشاؤها بواسطة Illumina MiSeq PE150 مع قاعدة بيانات مرجعية شاملة لفيروس PV باستخدام UPARSE. تم إنشاء جدول عد وحدات التصنيف التشغيلي (OTU) باستخدام عتبة الهوية، تعيين كل وحدة تصنيف تشغيلية (OTU) بنوع PV استنادًا إلى تصنيف الوكالة الدولية لبحوث السرطان (IARC)، تم تصنيف 12 نوعًا من فيروس الورم الحليمي البشري (HPV16، 18، 31، 33، 35، 39، 45، 51، 52، 56، 58، 59) على أنها عالية المخاطر (HR) وتعتبر مسرطنة. .

تسلسل RNA غير المشفر الطويل للجين المضيف (IncRNA-seq) وتحليل التعبير الجيني

تمت دراسة الأورام وأنسجة AN من مجموعة فرعية مكونة من 40 مريضًا عشوائيًا مصابًا بسرطان الفم والبلعوم (OSCC) لتوصيف النسخ الجينومية الشاملة باستخدام تسلسل lncRNA. باختصار، تم إزالة RNA الكلي من rRNA باستخدام مجموعة QIAseq FastSelect لإزالة rRNA وتم تحويله إلى مكتبة RNA اتجاهية باستخدام مجموعة NEBNext Ultra II لتحضير مكتبة RNA الاتجاهية لجهاز Illumina NovaSeq باستخدام قراءات مزدوجة النهاية بطول 150 نقطة أساسية. تم تعيين قراءات التسلسل مزدوجة النهاية إلى الجينوم المرجعي البشري hg38 باستخدام STAR. تم حساب عدد القراءات الموجهة إلى مناطق إكسون الجين باستخدام featureCounts تم الاحتفاظ بالجينات التي لديها متوسط Transcripts Per Kilobase Million (TPM) أكبر من 1. تم تحديد الجينات المعبر عنها بشكل مختلف (DEGs) بين أنسجة الورم وأنسجة AN باستخدام نموذج لوغاريتمي عام ثنائي الحدين في باستخدام حد قطع و (تغيير الطي) (الجينات المرتفعة التعبير) أو (الجينات المنخفضة التنظيم). تم تضمين العمر، والتدخين، واستهلاك الكحول، والجنس كعوامل مشوشة في النموذج. تم تلخيص الإثراء الوظيفي للجينات المعبر عنها بشكل مختلف المرتبطة بسرطان الفم والبلعوم باستخدام مجموعة ToppGene. أفضل 10 مصطلحات من موسوعة كيوتو للجينات والجنوم (KEGG) مع
الدلالة الإحصائية تم تصور ( ) بالترتيب. لتحديد الجينات المعبر عنها بشكل مختلف المرتبطة بالبقاء الخاص بالمرض، تم تصنيف تعبير كل جين بقيمة ثنائية “مرتفع” أو “منخفض” باستخدام نقاط القطع “المثلى” التي حددتها دالة cutpointr في R. تم تعيين معلمة “المقياس” في cutpointr كـ “sum_sens_spec” لتعظيم مجموع الحساسية والخصوصية. تم استخدام طريقة كابلان-ماير مع تم استخدام فاصل الثقة لنسبة مخاطر كوكس المتناسبة لتحليل البقاء أحادي المتغير باستخدام دالة coxph في R.

تحليل بيانات تسلسل RNA لسرطان الفم (OSCC) من أطلس جينوم السرطان (TCGA)

تم تنزيل مصفوفة عدد قراءات RNA-seq الخام لـ TCGA OSCC مع معلومات البيانات الوصفية (العمر، التدخين، استهلاك الكحول، الجنس، وموقع الاستئصال أو الخزعة) باستخدام حزمة R TCGAbiolinks. . تحتوي مجموعة البيانات هذه على عينات نسيجية من 502 مريضًا بسرطان الفم والبلعوم (OSCC) و44 ضابطًا غير مصاب بـ OSCC. تم تصفية العينات بشكل إضافي باستخدام المعايير التالية: (1) تمت إزالة عينة واحدة بسبب عدم وجود معلومات عن “العمر”، و(2) تم التخلص من 206 عينات لأن “موقع الاستئصال أو الخزعة” لم يكن ينتمي إلى تجويف الفم، مما أسفر عن 309 حالة و30 ضابطًا للمقارنة مع مجموعة مرضى OSCC في هونغ كونغ (HK).

تحليل الشبكة بين وفرة البكتيريا وتعبير الجينات المضيفة

تم استكشاف الأنواع البكتيرية المرتبطة بـ DEGs و OSCC لتفاعلات البكتيريا مع النسخ الجيني، حيث تم حساب معاملات الارتباط سبيرمان باستخدام دالة cor.test في R. تم تطبيع جداول عدد البكتيريا كنسبة مئوية لتحليل الارتباط. قيمة تم استخدامه كحد فاصل للدلالة. الارتباطات النادرة للبيانات التركيبية (SparCC) تم استخدامه لاستكشاف ارتباطات البكتيريا-البكتيريا، مع معامل ارتباط للدلالة. تم رسم خرائط الحرارة والشبكات باستخدام دالة corrplots في R وCytoscape v3.7.1 على التوالي. لتوصيف مسارات استهداف البكتيريا في مسببات سرطان الخلايا الحرشفية الفموية، تم تلخيص الجينات المعبر عنها بشكل مرتفع في المضيف والتي لها ارتباط إيجابي مع الأنواع البكتيرية الغنية بالورم لعمليات بيولوجية في علم الأحياء الجيني باستخدام مجموعة ToppGene. . تم الحصول على تم تصحيح القيم بعد ذلك لاختبار متعدد باستخدام طريقة بنجاميني وهوشبرغ (B & H). مصطلحات GO الغنية ( ” ) و ” المقابل تم استخدام القيم كمدخل لـ لتقليل تكرار مصطلحات GO وتصوير التجميع الدلالي لمصطلحات GO المحددة. في الوقت نفسه، تم دمج الجينات المعبر عنها بشكل مختلف المرتبطة بالبكتيريا الغنية بالأورام ومعاملات الارتباط الخاصة بها لتحليل إثراء مجموعة الجينات المميزة بشكل مُحسن ضد قاعدة بيانات التوقيعات الجزيئية (MSigDB) باستخدام GSEA. لقد تلخصت العلامات بشكل فعال معظم العمليات البيولوجية ذات الصلة لمجموعات الجينات من خلال تقليل التباين والتكرار.

ميثيلation CpG في الحمض النووي البشري والتحليل المعلوماتي الحيوي

تمت دراسة الأورام وأنسجة AN من مجموعة فرعية من 38 مريضًا مصابًا بسرطان الفم والبلعوم (OSCC) تم اختيارهم عشوائيًا لتوصيف ميثيلation CpG في الحمض النووي للمضيف باستخدام تسلسل البيسلفيت. باختصار، تم إعداد الحمض النووي الكلي مسبقًا لمكتبة إلومينا باستخدام مجموعة إعداد المكتبات KAPA HTP (Kapa Biosystems، الولايات المتحدة الأمريكية). تم تعريض المكتبة لثنائي كبريتيد الصوديوم باستخدام مجموعة EpiTect DNA Bisulfite (Qiagen، الولايات المتحدة الأمريكية) ثم تم استكشافها باستخدام مجسات Methyl-Seq Capture باستخدام مجموعة SeqCap Epi CpGiant (Roche، الولايات المتحدة الأمريكية) لتسلسل Illumina NovaSeq باستخدام قراءات مزدوجة النهاية بطول 150 نقطة أساسية. تم تقليم قراءات التسلسل مزدوجة النهاية باستخدام Trimmomatic. ومطابقة للجينوم البشري المرجعي (hg38) باستخدام Bismark تم فرز ملفات BAM وإزالة التكرار؛ تم تضمين فقط المناطق الجينومية التي كانت تغطيتها من القراءات أكبر من 10. تم نمذجة مستويات الميثيل باستخدام خوارزمية قائمة على الانحدار اللوجستي مع تصحيح الإفراط في التشتت واختبار كاي-تربيع المطبق في حزمة methylKit في R. تم تحديد المناطق الميثيلية المختلفة (DMRs) في الأورام مقارنةً بأنسجة AN باستخدام حد أدنى من ونتيجة الفرق في الميث (فرط الميثلة) أو (نقص الميثيل). تم تضمين العمر والجنس والتدخين واستهلاك الكحول ومراحل N و T كعوامل مشوشة في النموذج. ثم تم تعيين مناطق DMRs إلى منطقة المحفز للجينات باستخدام دالة annotatePeak في حزمة R ChIPseeker. تم تعريف منطقة المحفز
3000 نقطة أساسية قبل و3000 نقطة أساسية بعد مواقع بدء النسخ (TSSs). عندما تم تحديد عدة مناطق ديميثيلاز (DMRs) داخل نفس منطقة المحفز، استخدمنا طريقة التصويت بالأغلبية لتحديد حالة الميثيل للجين. تم استخدام اختبار ارتباط ترتيب سبيرمان لاستكشاف العلاقة بين DMRs المرتبطة بالمحفز وأنواع البكتيريا الغنية بـ OSCC. قيمة اعتُبر كدلالة إحصائية.

التعايش مع بكتيريا الفوسوبكتيريوم نوكليتوم مع خلايا الظهارة المضيفة في المختبر

للتحقق من قدرة البكتيريا المسببة للأمراض على تعطيل النسخ الجيني للمضيف كما هو موضح في بيئة الورم في سرطان الخلايا الحرشفية الفموية، تم إنشاء نموذج قائم على الخلايا في المختبر من خلال زراعة F. nucleatum مع خلايا HGK12، وهي خط خلايا ظهارية مستمدة من الكيراتينوسيت البشرية السليمة في اللثة (هدية من البروفيسور لو، كلية العلوم الطبية الحيوية، الجامعة الصينية في هونغ كونغ) وSAS، وهو خط خلايا سرطان الخلايا الحرشفية اللسانية البشرية (Cellosaurus، RRID:CVCL_1675). باختصار، تم استخدام سلالة F. nucleatum subsp. vincentii، التي تم عزلها من مريض بسرطان الخلايا الحرشفية الفموية وتم تأكيدها من خلال تسلسل الجينوم الكامل، في نظام زراعة الخلايا مع البكتيريا عند مضاعف الإصابة (MOI) من لمدة 5 ساعات عند تحت ظروف لاهوائية. تم زراعة HGK12 و SAS معًا . نوكليتوم مرة واحدة وتمت زراعة الخلايا المصابة بعد ذلك في وسط إبيلايف أو DMEM/F12 جديد، على التوالي، عند مع بعد ليلة من الثقافة مع أو بدون بكتيريا، تم استخراج RNA الكلي من الخلايا باستخدام مجموعة RNeasy Mini (Qiagen، الولايات المتحدة الأمريكية)، تلتها إعداد مكتبة تسلسل lncRNA باستخدام Illumina. تم إجراء ثلاث تكرارات مستقلة من أجل القوة الإحصائية. باختصار، تم تعيين قراءات التسلسل المزدوجة إلى الجينوم المرجعي البشري hg38 باستخدام STAR. تم حساب عدد القراءات الموجهة إلى مناطق إكسون الجين باستخدام featureCounts تم الاحتفاظ بالجينات التي لديها متوسط Transcripts Per Kilobase Million (TPM) أكبر من 1. تم تحديد الجينات المعبر عنها بشكل مختلف (DEGs) بين خطوط الخلايا المعالجة وغير المعالجة باستخدام نموذج لوغاريتمي عام ثنائي الحدين في ، مع حد أقصى لـ abs( ) و يعتبر ذا دلالة إحصائية. إثراء وظيفي لـ تم تلخيص الجينات المعبر عنها بشكل مختلف المرتبطة بـ . nucleatum باستخدام مجموعة ToppGene أعلى 10 مصطلحات من موسوعة كيوتو للجينات والجنوم (KEGG) ذات دلالة إحصائية تم تصورها بالترتيب.

تفاعل البوليميراز المتسلسل في الوقت الحقيقي (RT-PCR) لتعبير جينات mRNA

تم استخراج RNA الكلي من خطوط خلايا HGK12 و SAS التي تم زراعتها مع F. nucleatum (نسبة الجرعة 1:100) باستخدام كاشف TRIzol. كتحكم بكتيري سلبي، تم استخدام Fusobacterium mortiferum، وهو نوع مقترح غير غازي وغير ممرض ضمن جنس الفوسوبكتيريوم. . واحد تم استخدام إجمالي RNA لإنتاج DNA مكمل بواسطة LunaScript طقم RT SuperMix (NEB، #E3010) يليه تفاعل RT-PCR القياسي مع لونا بروتوكول مزيج ماستر qPCR العالمي (NEB، #M3003). تم إدراج بادئات RTPCR المستخدمة للتحقق من تعبير الجينات في الجدول التكميلي 22.

التحليل الغربي للتعبير البروتيني

تم الحصول على عينات البروتين من خطوط خلايا HGK12 و SAS المزروعة معًا في وجود أو غياب البكتيريا المعنية. تم جمع العينات باستخدام محلول التحلل RIPA (سانتا كروز). تم إجراء تقدير البروتين باستخدام بيرس. أطقم اختبار بروتين BCA (ثيرمو فيشر ساينتيفيك). إجمالي من كل عينة تم فصل البروتين بواسطة أو SDS-PAGE وتم نقلها إلى أغشية PVDF للتلطيخ. لمنع الارتباط غير المحدد، تم حجب جميع الأغشية باستخدام حليب خالي من الدسم في 1X TBST. الأجسام المضادة الأولية ضد SNAI2 (تكنولوجيا الإشارة الخلوية، 9585T، تخفيف 1:500)، INHBA (أبكام، ab128958، تخفيف 1:500)، LAMA3 (أبكام، ab151715، تخفيف 1:500)، LAMC2 (سانتا كروز، sc-28330، تخفيف 1:1000) و الأكتين (سانتا كروز، sc-47778، 1:1000)، تم تحضيره بعد ذلك ليلاً في ، على التوالي، تلتها تطبيق الأجسام المضادة الثانوية المقابلة لمدة ساعة واحدة في درجة حرارة الغرفة. تم الكشف عن حزم البروتين المستهدفة باستخدام مادة Clarity Western ECL (Bio-Rad) وتم تصويرها باستخدام جهاز تصوير Western Blot الكيميائي AZURE 300 (Azure Biosystems). جميع
تم تحديد مستويات التعبير من خلال قياس قيم التدرج الرمادي لشرائط البروتين المستهدفة بالنسبة للبروتين المرجعي بيتا-أكتين باستخدام برنامج ImageJ. جميع البقع والهلامات تأتي من نفس التجربة وتم معالجتها بشكل متوازي.

ملخص التقرير

معلومات إضافية حول تصميم البحث متاحة في ملخص تقارير الأبحاث في مجلة Nature المرتبط بهذه المقالة.

توفر البيانات

جميع التسلسلات والرموز متاحة عند الطلب. تسلسلات الجين 16 S rRNA التي تم تحليلها في هذه الدراسة متاحة في قاعدة بيانات NCBI SRA تحت مشروع البيولوجيا PRJNA822685 (أرقام الوصول من SRR18595688 إلى SRR18595849). تم إيداع خطوط الأنابيب والبرامج النصية المستخدمة في هذه الدراسة في مستودع Github.https://github.com/lycai05/OSCC_host_bacteria_interactions).
تاريخ الاستلام: 9 أبريل 2023؛ تاريخ القبول: 27 مارس 2024؛
نُشر على الإنترنت: 08 أبريل 2024

References

  1. Li, Q. et al. Role of oral bacteria in the development of oral squamous cell carcinoma. Cancers 12, 2797 (2020).
  2. Johnson, D. E. et al. Head and neck squamous cell carcinoma. Nat. Rev. Dis. Prim. 6, 1-22 (2020).
  3. Sepich-Poore, G. D. et al. The microbiome and human cancer. Science 371 https://doi.org/10.1126/science.abc4552 (2021).
  4. Guerrero-Preston, R. et al. 16 S rRNA amplicon sequencing identifies microbiota associated with oral cancer, human papilloma virus infection and surgical treatment. Oncotarget 7, 51320-51334 (2016).
  5. Ganly, I. et al. Periodontal pathogens are a risk factor of oral cavity squamous cell carcinoma, independent of tobacco and alcohol and human papillomavirus. Int. J. Cancer 145, 775-784 (2019).
  6. Chen, Z. et al. The intersection between oral microbiota, host gene methylation and patient outcomes in head and neck squamous cell carcinoma. Cancers 12 https://doi.org/10.3390/cancers12113425 (2020).
  7. Kamarajan, P. et al. Periodontal pathogens promote cancer aggressivity via TLR/MyD88 triggered activation of Integrin/FAK signaling that is therapeutically reversible by a probiotic bacteriocin. PLoS Pathog. 16, e1008881 (2020).
  8. Michmerhuizen, N. L., Birkeland, A. C., Bradford, C. R. & Brenner, J. C. Genetic determinants in head and neck squamous cell carcinoma and their influence on global personalized medicine. Genes Cancer 7, 182-200 (2016).
  9. The Cancer Genome Atlas Network Comprehensive genomic characterization of head and neck squamous cell carcinomas. Nature 517, 576-582 (2015).
  10. Maruya, S. et al. Differential methylation status of tumorassociated genes in head and neck squamous carcinoma: incidence and potential implications. Clin. Cancer Res. 10, 3825-3830 (2004).
  11. Rosas, S. L. et al. Promoter hypermethylation patterns of p 16 , O6-methylguanine-DNA-methyltransferase, and death-associated protein kinase in tumors and saliva of head and neck cancer patients. Cancer Res. 61, 939-942 (2001).
  12. Xia, X. et al. Bacteria pathogens drive host colonic epithelial cell promoter hypermethylation of tumor suppressor genes in colorectal cancer. Microbiome 8, 108 (2020).
  13. Jena, P. K. et al. Dysregulated bile acid synthesis and dysbiosis are implicated in Western diet-induced systemic inflammation, microglial activation, and reduced neuroplasticity. FASEB J. 32, 2866-2877 (2018).
  14. Song, X. et al. Oral squamous cell carcinoma diagnosed from saliva metabolic profiling. Proc. Natl Acad. Sci. USA 117, 16167-16173 (2020).
  15. . et al. Oncometabolite 2 -hydroxyglutarate is a competitive inhibitor of alpha-ketoglutarate-dependent dioxygenases. Cancer Cell 19, 17-30 (2011).
  16. Janney, A., Powrie, F. & Mann, E. H. Host-microbiota maladaptation in colorectal cancer. Nature 585, 509-517 (2020).
  17. Dayama, G., Priya, S., Niccum, D. E., Khoruts, A. & Blekhman, R. Interactions between the gut microbiome and host gene regulation in cystic fibrosis. Genome Med. 12, 12 (2020).
  18. Helmink, B. A., Khan, M. A. W., Hermann, A., Gopalakrishnan, V. & Wargo, J. A. The microbiome, cancer, and cancer therapy. Nat. Med. 25, 377-388 (2019).
  19. Nejman, D. et al. The human tumor microbiome is composed of tumor type-specific intracellular bacteria. Science 368, 973-980 (2020).
  20. Rubinstein, M. R. et al. Fusobacterium nucleatum promotes colorectal carcinogenesis by modulating E-cadherin/beta-catenin signaling via its FadA adhesin. Cell Host Microbe 14, 195-206 (2013).
  21. Rousselle, P. & Scoazec, J. Y. Laminin 332 in cancer: when the extracellular matrix turns signals from cell anchorage to cell movement. Semin. Cancer Biol. 62, 149-165 (2020).
  22. Jung, A. R., Jung, C.-H., Noh, J. K., Lee, Y. C. & Eun, Y.-G. Epithelialmesenchymal transition gene signature is associated with prognosis and tumor microenvironment in head and neck squamous cell carcinoma. Sci. Rep. 10, 3652 (2020).
  23. Hu, Y. et al. Epigenetic suppression of E-cadherin expression by Snail2 during the metastasis of colorectal cancer. Clin. Epigenetics 10, 154 (2018).
  24. Tang, D. et al. TNF-alpha promotes invasion and metastasis via NFKappa B pathway in oral squamous cell carcinoma. Med. Sci. Monit. Basic Res. 23, 141-149 (2017).
  25. Wu, Y. et al. Stabilization of snail by NF-kappaB is required for inflammation-induced cell migration and invasion. Cancer Cell 15, 416-428 (2009).
  26. Adjuto-Saccone, M. et al. TNF-alpha induces endothelialmesenchymal transition promoting stromal development of pancreatic adenocarcinoma. Cell Death Dis. 12, 649 (2021).
  27. Coppenhagen-Glazer, S. et al. Fap2 of Fusobacterium nucleatum is a galactose-inhibitable adhesin involved in coaggregation, cell adhesion, and preterm birth. Infect. Immun. 83, 1104-1113 (2015).
  28. Yang, Y. et al. Fusobacterium nucleatum increases proliferation of colorectal cancer cells and tumor development in mice by activating toll-like receptor 4 signaling to nuclear factor-kappaB, and upregulating expression of microRNA-21. Gastroenterology 152, 851-866.e24 (2017).
  29. Long, X. et al. Peptostreptococcus anaerobius promotes colorectal carcinogenesis and modulates tumour immunity. Nat. Microbiol. 4, 2319-2330 (2019).
  30. Zhang, L., Liu, Y., Zheng, H. J. & Zhang, C. P. The oral microbiota may have influence on oral cancer. Front. Cell Infect. Microbiol. 9, 476 (2019).
  31. Huang, X. et al. Wnt7a activates canonical Wnt signaling, promotes bladder cancer cell invasion, and is suppressed by miR-370-3p. J. Biol. Chem. 293, 6693-6706 (2018).
  32. Xie, H. et al. WNT7A promotes EGF-induced migration of oral squamous cell carcinoma cells by activating -catenin/MMP9mediated signaling. Front. Pharmacol. 11, 98 (2020).
  33. Chen, L. et al. VEGF promotes migration and invasion by regulating EMT and MMPs in nasopharyngeal carcinoma. J. Cancer 11, 7291-7301 (2020).
  34. Wang, X. et al. Characterization of LIMA1 and its emerging roles and potential therapeutic prospects in cancers. Front. Oncol. 13, 1115943 (2023).
  35. Zeng, J., Jiang, W. G. & Sanders, A. J. Epithelial protein lost in neoplasm, EPLIN, the cellular and molecular prospects in cancers. Biomolecules 11 https://doi.org/10.3390/biom11071038 (2021).
  36. Feinberg, A. P., Ohlsson, R. & Henikoff, S. The epigenetic progenitor origin of human cancer. Nat. Rev. Genet. 7, 21-33 (2006).
  37. Yu, D. H. et al. Postnatal epigenetic regulation of intestinal stem cells requires DNA methylation and is guided by the microbiome. Genome Biol. 16, 211 (2015).
  38. Fellows, R. et al. Microbiota derived short chain fatty acids promote histone crotonylation in the colon through histone deacetylases. Nat. Commun. 9, 105 (2018).
  39. Ansari, I. et al. The microbiota programs DNA methylation to control intestinal homeostasis and inflammation. Nat. Microbiol. 5, 610-619 (2020).
  40. Qiao, F. et al. Downregulated PITX1 modulated by MiR-19a-3p promotes cell malignancy and predicts a poor prognosis of gastric cancer by affecting transcriptionally activated PDCD5. Cell Physiol. Biochem. 46, 2215-2231 (2018).
  41. Wang, Q., Zhao, S., Gan, L. & Zhuang, Z. Bioinformatics analysis of prognostic value of PITX1 gene in breast cancer. Biosci. Rep. 40 https://doi.org/10.1042/BSR20202537 (2020).
  42. Kolfschoten, I. G. M. et al. A genetic screen identifies PITX1 as a suppressor of RAS activity and tumorigenicity. Cell 121, 849-858 (2005).
  43. Vantaku, V. et al. Epigenetic loss of AOX1 expression via EZH2 leads to metabolic deregulations and promotes bladder cancer progression. Oncogene 39, 6265-6285 (2020).
  44. Liu, C. et al. GALNT6 promotes breast cancer metastasis by increasing mucin-type O-glycosylation of alpha2M. Aging 12, 11794-11811 (2020).
  45. Chen, S. et al. Fusobacterium nucleatum reduces METTL3-mediated m(6)A modification and contributes to colorectal cancer metastasis. Nat. Commun. 13, 1248 (2022).
  46. Xu, C. et al. Fusobacterium nucleatum promotes colorectal cancer metastasis through miR-1322/CCL20 axis and M2 polarization. Gut Microbes 13, 1980347 (2021).
  47. Wong, M. C. S. et al. Prevalence and epidemiologic profile of oral infection with alpha, beta, and gamma papillomaviruses in an Asian Chinese population. J. Infect. Dis. 218, 388-397 (2018).
  48. Chen, Z. et al. Impact of preservation method and 16S rRNA hypervariable region on gut microbiota profiling. mSystems 4 https://doi.org/10.1128/mSystems.00271-18 (2019).
  49. Matsen, F. A., Kodner, R. B. & Armbrust, E. V. pplacer: linear time maximum-likelihood and Bayesian phylogenetic placement of sequences onto a fixed reference tree. BMC Bioinforma. 11, 538 (2010).
  50. Zhu, H. et al. Convergent dysbiosis of upper aerodigestive microbiota between patients with esophageal and oral cavity squamous cell carcinoma. Int. J. Cancer 152, 1903-1915 (2023).
  51. Mirarab, S., Nguyen, N. & Warnow, T. SEPP: SATe-enabled phylogenetic placement. Pac. Symp. Biocomput. 247-258 https://doi.org/10.1142/9789814366496_0024 (2012).
  52. Hamady, M., Lozupone, C. & Knight, R. Fast UniFrac: facilitating highthroughput phylogenetic analyses of microbial communities including analysis of pyrosequencing and PhyloChip data. ISME J. 4, 17-27 (2010).
  53. Segata, N. et al. Metagenomic biomarker discovery and explanation. Genome Biol. 12, R60 (2011).
  54. Lin, H. & Peddada, S. D. Analysis of compositions of microbiomes with bias correction. Nat. Commun. 11, 3514 (2020).
  55. Fernandes, A. D. et al. Unifying the analysis of high-throughput sequencing datasets: characterizing RNA-seq, 16S rRNA gene sequencing and selective growth experiments by compositional data analysis. Microbiome 2, 15 (2014).
  56. Yang, L. & Chen, J. A comprehensive evaluation of microbial differential abundance analysis methods: current status and potential solutions. Microbiome 10, 130 (2022).
  57. Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139-140 (2010).
  58. Edgar, R. C. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat. Methods 10, 996-998 (2013).
  59. Bernard, H. U. et al. Classification of papillomaviruses (PVs) based on 189 PV types and proposal of taxonomic amendments. Virology 401, 70-79 (2010).
  60. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans Biological agents. Volume 100 B. A review of human carcinogens. IARC Monogr. Eval. Carcinog. Risks Hum. 100, 1-441 (2012).
  61. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15-21 (2013).
  62. Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923-930 (2014).
  63. Chen, J., Bardes, E. E., Aronow, B. J. & Jegga, A. G. ToppGene Suite for gene list enrichment analysis and candidate gene prioritization. Nucleic Acids Res. 37, W305-W311 (2009).
  64. Colaprico, A. et al. TCGAbiolinks: an R/Bioconductor package for integrative analysis of TCGA data. Nucleic Acids Res. 44, e71 (2016).
  65. Friedman, J. & Alm, E. J. Inferring correlation networks from genomic survey data. PLoS Comput. Biol. 8, e1002687 (2012).
  66. Saito, R. et al. A travel guide to Cytoscape plugins. Nat. Methods 9, 1069-1076 (2012).
  67. Supek, F., Bosnjak, M., Skunca, N. & Smuc, T. REVIGO summarizes and visualizes long lists of gene ontology terms. PLoS ONE 6, e21800 (2011).
  68. Subramanian, A. et al. Gene set enrichment analysis: a knowledgebased approach for interpreting genome-wide expression profiles. Proc. Natl Acad. Sci. USA 102, 15545-15550 (2005).
  69. Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114-2120 (2014).
  70. Krueger, F. & Andrews, S. R. Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications. Bioinformatics 27, 1571 (2011).
  71. Akalin, A. et al. methylKit: a comprehensive package for the analysis of genome-wide DNA methylation profiles. Genome Biol. 13, R87 (2012).
  72. Yu, G., Wang, L. G. & He, Q. Y. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics 31, 2382-2383 (2015).
  73. Manson McGuire, A. et al. Evolution of invasion in a diverse set of Fusobacterium species. mBio 5, e01864 (2014).

شكر وتقدير

يشكر المؤلفون المشاركين المجهولين الذين قدموا عينات لهذه الدراسة. كما نشكر شركة نوفوجين (HK) المحدودة، الصين، على المساعدة في تسلسل الحمض النووي الريبي غير المشفر الطويل للجينات المضيفة (IncRNAseq) وتسلسل الميثيل الحمضي النووي البشري CpG، ومرافق الجينوميات السرطانية وعلم الأمراض (CUCGP) في قسم علم التشريح وعلم الأمراض الخلوية في الجامعة الصينية في هونغ كونغ على خدمة تسلسل جين 16 S rRNA. تم دعم هذا العمل جزئيًا من قبل جهات التمويل من مؤسسة ستانلي هو الطبية (J.Y.K.C.)، ومجلس منح الأبحاث لمنطقة هونغ كونغ الإدارية الخاصة، الصين (أرقام المشاريع CUHK14108818 إلى J.Y.K.C. وCUHK14161017 إلى Z.C.)، ومكتب الغذاء والصحة، منطقة هونغ كونغ الإدارية الخاصة، الصين (رقم المرجع 19202041 إلى Z.C.)، ومنحة مطابقة الأبحاث (رقم المرجع 8601687 إلى P.K.S.C.) من مجلس منح الأبحاث لمنطقة هونغ كونغ الإدارية الخاصة، الصين. لم يكن للجهات الممولة أي دور في تصميم الدراسة، أو جمع البيانات وتحليلها، أو اتخاذ قرار النشر، أو إعداد المخطوطة.

مساهمات المؤلفين

قام P.K.S.C. و J.Y.K.C. و Z.C. بتصميم الدراسة وتصميم التجارب. جمع C.W.K.N. و K.M. و V.W.Y.L. و J.Y.K.C. عينات سريرية وبيانات وصفية. طور H.Z. و Q.M. و P.Y.W. و L.L. و P.L. الطرق وأجروا التجارب. قام L.C. و H.Z. و K.M. و A.S.L.C. و J.Y. و P.K.S.C. و J.Y.K.C. و Z.C. بتحليل البيانات ومناقشة النتائج. كتب L.C. و H.Z. و J.Y.K.C. و Z.C. المسودة الأصلية. قام جميع المؤلفين بمراجعة النسخة النهائية من المخطوطة والموافقة عليها.

المصالح المتنافسة

يعلن المؤلفون عدم وجود مصالح متنافسة.

معلومات إضافية

معلومات إضافية النسخة الإلكترونية تحتوي على
المواد التكميلية متاحة على
https://doi.org/10.1038/s41522-024-00511-x.
يجب توجيه المراسلات وطلبات المواد إلى جيسون واي. ك. تشان أو زيغوي تشين.
معلومات إعادة الطبع والتصاريح متاحة على http://www.nature.com/reprints
ملاحظة الناشر: تظل شركة سبرينغر ناتشر محايدة فيما يتعلق بالمطالبات القضائية في الخرائط المنشورة والانتماءات المؤسسية.
الوصول المفتوح: هذه المقالة مرخصة بموجب رخصة المشاع الإبداعي للاستخدام والمشاركة والتكيف والتوزيع وإعادة الإنتاج في أي وسيلة أو صيغة، طالما أنك تعطي الائتمان المناسب للمؤلفين الأصليين والمصدر، وتوفر رابطًا لرخصة المشاع الإبداعي، وتوضح ما إذا كانت هناك تغييرات قد تم إجراؤها. الصور أو المواد الأخرى من طرف ثالث في هذه المقالة مشمولة في رخصة المشاع الإبداعي للمقالة، ما لم يُذكر خلاف ذلك في سطر الائتمان للمادة. إذا لم تكن المادة مشمولة في رخصة المشاع الإبداعي للمقالة وكان استخدامك المقصود غير مسموح به بموجب اللوائح القانونية أو يتجاوز الاستخدام المسموح به، ستحتاج إلى الحصول على إذن مباشرة من صاحب حقوق الطبع والنشر. لعرض نسخة من هذه الرخصة، قم بزيارة http://creativecommons.org/licenses/by/4.0/.
© المؤلفون 2024

  1. قسم علم الأحياء الدقيقة، الجامعة الصينية في هونغ كونغ، منطقة هونغ كونغ الإدارية الخاصة، الصين. قسم الأنف والأذن والحنجرة، جراحة الرأس والعنق، الجامعة الصينية في هونغ كونغ، منطقة هونغ كونغ الإدارية الخاصة، الصين. قسم علم الأمراض الكيميائية، الجامعة الصينية في هونغ كونغ، منطقة هونغ كونغ الإدارية الخاصة، الصين. مركز جورجيا للسرطان، أوغستا، GA 30912، الولايات المتحدة الأمريكية. قسم الطب، كلية الطب في جورجيا، جامعة أوغستا، أوغستا، GA 30912، الولايات المتحدة الأمريكية. كلية العلوم الطبية الحيوية، الجامعة الصينية في هونغ كونغ، منطقة هونغ كونغ الإدارية الخاصة، الصين. قسم الطب والعلاج، الجامعة الصينية في هونغ كونغ، منطقة هونغ كونغ الإدارية الخاصة، الصين. ساهم هؤلاء المؤلفون بالتساوي: ليويانغ كاي، هينغيان زو، تشيان تشيان مو.

Journal: npj Biofilms and Microbiomes, Volume: 10, Issue: 1
DOI: https://doi.org/10.1038/s41522-024-00511-x
PMID: https://pubmed.ncbi.nlm.nih.gov/38589501
Publication Date: 2024-04-08

Integrative analysis reveals associations between oral microbiota dysbiosis and host genetic and epigenetic aberrations in oral cavity squamous cell carcinoma

Liuyang Cai , Hengyan Zhu , Qianqian Mou , Po Yee Wong , Linlin Lan , Cherrie W. K. Ng , Pu Lei , Man Kit Cheung © , Daijuanru Wang , Eddy W. Y. Wong , Eric H. L. Lau , Zenon W. C. Yeung , Ronald Lai , Katie Meehan , Sherwood Fung , Kwan Chee A. Chan , Vivian W. Y. Lui , Alfred S. L. Cheng , Jun Yu , Paul K. S. Chan , Jason Y. K. Chan & Zigui Chen

(a) Check for updates

Abstract

Dysbiosis of the human oral microbiota has been reported to be associated with oral cavity squamous cell carcinoma (OSCC) while the host-microbiota interactions with respect to the potential impact of pathogenic bacteria on host genomic and epigenomic abnormalities remain poorly studied. In this study, the mucosal bacterial community, host genome-wide transcriptome and DNA CpG methylation were simultaneously profiled in tumors and their adjacent normal tissues of OSCC patients. Significant enrichment in the relative abundance of seven bacteria species (Fusobacterium nucleatum, Treponema medium, Peptostreptococcus stomatis, Gemella morbillorum, Catonella morbi, Peptoanaerobacter yurli and Peptococcus simiae) were observed in OSCC tumor microenvironment. These tumor-enriched bacteria formed 254 positive correlations with 206 up-regulated host genes, mainly involving signaling pathways related to cell adhesion, migration and proliferation. Integrative analysis of bacteria-transcriptome and bacteria-methylation correlations identified at least 20 dysregulated host genes with inverted CpG methylation in their promoter regions associated with enrichment of bacterial pathogens, implying a potential of pathogenic bacteria to regulate gene expression, in part, through epigenetic alterations. An in vitro model further confirmed that Fusobacterium nucleatum might contribute to cellular invasion via crosstalk with E -cadherin -catenin signaling, TNF pathway and extracellular matrix remodeling by up-regulating SNAI2 gene, a key transcription factor of epithelial-mesenchymal transition (EMT). Our work using multi-omics approaches explored complex host-microbiota interactions and provided important insights into genetic and functional basis in OSCC tumorigenesis, which may serve as a precursor for hypothesisdriven study to better understand the causational relationship of pathogenic bacteria in this deadly cancer.

Oral cavity squamous cell carcinoma (OSCC) is the most prevalent oral malignancy worldwide and is associated with significant mortality and morbidity rates . About of oral cancers are diagnosed at an advanced stage, resulting in a 5 -year survival rates of less than . Known etiological factors of OSCC tumorigenesis include tobacco consumption, alcohol abuse
and betel nut chewing; however, the prevalence of OSCC without traditional risk factors has been increasing in recent years . There is therefore an urgent need to identify other underlying etiologies with prognostic relevance for early diagnosis and improved treatment in OSCC patients. Given the increasing evidence indicating the carcinogenic potential of the human
microbiome in cancers , elucidating the human microbiome in OSCC may explain, in part, the fact that a subset of patients who are not exposed to traditional risk factors eventually develop cancer. In fact, a significant loss in microbial diversity has been reported in OSCC patients ; several periodontal pathogens, for example, Fusobacterium, Peptostreptococcus and Treponema, were significantly enriched in OSCC tumor microenvironment . These pathogenic bacteria were reported to promote oral cancer aggressivity via TLR/MyD88 triggered activation of integrin/FAK signaling pathway , making the understanding of pathogenic bacteria in cancer of great importance.
There is a growing awareness that global patterns of genetic and epigenetic changes play a critical role in the molecular characteristics of oral cancer. For example, EGFR, PIK3CA, NOTCH pathways and TP53 gene are among the most frequently altered in head and neck squamous cell carcinoma (HNSCC) . Several studies of head and neck cancer have identified promoter methylation of CDKN2A (p16), DAP kinase (DAPK), and DNA repair genes and . Bacteria could trigger human epigenetic modification, thereby silencing tumor suppressor gene expression . Thus, profiling DNA genetic and epigenetic events that might be affected by dysbiosis of the oral microbiota may provide us with a key understanding of its role in gene regulation and enable us in establishing specific marker for early diagnosis and therapeutic strategies.
In this study, OSCC tumors and their adjacent normal (AN) tissues from a cohort of HPV-negative patients were simultaneously profiled for oral mucosal microbiota, host genome-wide transcriptome and DNA CpG methylation to explore the genetic basis of host-microbiota interactions. Integrative analysis using multi-omics approaches revealed complex networks between oral microbiota dysbiosis and host genetic and epigenetic abnormalities. Our findings provide important insights into genetic and functional basis for better understanding the role of oral bacteria in the pathogenesis of OSCC.

Results
Study subjects

A total of 98 OSCC patients who provided paired tumor and AN tissues were recruited in this study. Among them, eight patients infected with highrisk HPV types ( 7 with HPV16 and one with HPV18) were excluded. This retained cohort consisted of 57 males and 33 females, with a mean age of 65 years (sd: 12 years). Detailed demographic and clinical information are available in Supplementary Tables 1 and 2.

Oral microbiota dysbiosis in OSCC

High quality bacterial 16 S reads from paired tumor and AN tissues were available from 81 OSCC patients (mean reads of ) (Supplementary Table 3 and Supplementary Data 1-3), with Firmicutes (mean relative abundance of standard deviation of ) as the most predominant bacterial phylum, followed by Fusobacteria ( ), Proteobacteria ( ), Bacteroidetes ( ) and seven other phyla (Supplementary Fig. 1a). At the amplicon sequence variant (ASV) level, significantly reduced alpha diversity of the oral mucosal microbial community was observed in tumor tissues as measured by Richness, Shannon and Simpson indices compared to AN controls (Mann-Whitney test, ) (Fig. 1a); similarly, a principal coordinate analysis (PCoA) based on the weighted GUniFrac matrix showed distinct separation between tumor and AN groups ( ) (Fig. 1b). The finding underscores the distinct microbial community associated with OSCC tumor sites compared to adjacent normal tissues.
We used permutational multivariate analysis of variance (PERMANOVA) to quantify the contribution of factors (disease status, gender, aging, smoking, alcohol consumption, and T and N cancer stages) to the differences of microbial composition in patients with OSCC. This analysis identified disease status as the independent factor that had the largest effect on the overall structure of the oral microbiota (weighted GUniFrac ; unweighted GUniFrac ; Bray-Curtis ) (Supplementary Fig. 1b). In addition,
other factors, including T stage (T1&T2 vs. T3&T4) (weighted GUniFrac ; unweighted GUniFrac ; Bray-Curtis ) and smoking (Yes vs. No) (weighted GUniFrac ; Bray-Curtis ) were also significant, albeit with a smaller effect. The influence of disease status on oral microbiota composition highlights its potential role in OSCC pathogenesis.
Using a linear discriminant analysis effect size (LEfSe) test (LDA > 3, ), which was further verified by at least two of the three compositional aware tools, ANCOM-BC2, ALDEx2 and ZicoSeq tests adjusted for the covariates of T stage and smoking ( ), we applied a phylogenetic placement algorithm, pplacer, to map ASV reads against a NCBI 16S rRNA reference database with maximum likelihood and observed 6 and 13 bacterial genera significantly enriched and depressed in the relative abundance in OSCC tumor microenvironment, respectively (Supplementary Fig. 1c, Table 1 and Supplementary Table 4a). Moreover, seven mucosal bacterial species (Fusobacterium nucleatum, Treponema medium, Peptostreptococcus stomatis, Gemella morbillorum, Catonella morbi, Peptoanaerobacter yurli and Peptococcus simiae) were found to be predominantly colonized in the oral cavity of cancer patients (Fig. 1c, Table 1 and Supplementary Table 4b), suggesting a potential of tumor-enriched bacteria in the pathogenesis of OSCC.
A hierarchical clustering inferred from 6 discriminative bacterial species (LDA > 3, AUC > 0.65) was able to differentiate the surveyed samples into two clades, composed of the majority of tumors (62/92) and controls (51/70), with an odds ratio of 5.48 (95% CI 2.66-11.68, ) (Fig. 1d). These bacterial species achieved a combined area under the receiver operating characteristic (ROC) curve (AUC) of 0.858 (95% CI 0.803-0.914). This discrimination may serve as a basis for developing diagnostic markers for OSCC.

Fusobacterium nucleatum associated with OSCC patients without traditional risk factors

Since F. nucleatum was the most predominant bacterial species in OSCC tumor microenvironment ( vs. ) (Supplementary Fig. 2a and Table 1), we further analyzed its colonization associated with patient features. The enrichment of . nucleatum was significantly associated with non-smokers (mean log2-fold change of 1.48 vs. ) (Supplementary Fig. 2b) and non-drinkers (1.35 vs. ) (Supplementary Fig. 2c). Since all recruited patients were high-risk HPV-negative, the finding implies the potential of pathogenetic role of . nucleatum in OSCC patients negative for HPV infection, smoking and alcohol consumption (triple negative). F. nucleatum was also associated with better cancer-specific survival ( 1.42 vs. ) (Supplementary Fig. 2d), which was likely supported by a better 3 -year disease-specific survival ( vs. at 36 months, ) (Supplementary Fig. 2e). These results suggest that . nucleatum could be a distinctive marker in HPV-negative OSCC cases, particularly among non-smokers and non-drinkers, and may also be indicative of a more favorable prognosis in these patients.

Differentially expressed host gene transcriptome in OSCC

A subset of 40 tumors and 22 AN tissues were profiled for host genome-wide transcriptome (Supplementary Tables 2 and 5), which revealed a globally reconfigured host gene dysregulation in OSCC tumor microenvironment (Supplementary Fig. 3a and Supplementary Table 6). Among 17,225 transcripts with mean transcripts per kilobase million greater than 1 (TPM > 1), we identified 1407 up-regulated (tumor-associated) and 1525 downregulated (AN-associated) genes in the surveyed OSCC cohort (log2FC > 1 or ) (Supplementary Figs. 3a and 3b and Supplementary Table 6), highly consistent with the transcriptome profile of The Cancer Genome Atlas (TCGA) OSCC data (Spearman cor ) (Supplementary Fig. 3c). For example, the expression of LAMC2 (involved in epithelial cell migration) was dramatically elevated in both , ) (Fig. 2b) and TCGA-OSCC datasets (Supplementary Fig. 3d). Likewise, MMP1 (involved in tumor cell invasion) was ranked as one of
Fig. 1 | Oral microbiota dysbiosis associated with oral cavity squamous cell carcinoma (OSCC). a Comparison of the oral microbiota alpha diversity between OSCC tumor (tumor) and adjacent normal (AN) tissues at the amplicon sequence variant (ASV) level. The boxplot’s center line indicates the median value, the box bounds represent the first and third quartiles, and the whiskers extend to the smallest and largest values in the data, respectively. b Principal coordinate analysis plot based on weighted GUniFrac distance matrix inferred from ASVs. c Discriminative bacterial species as detected by linear discriminant analysis (LDA) effect size (LEfSe)
analysis (score ), which was further validated by at least two of the three compositional aware tools, ANCOM-BC2, ALDEx2 and ZicoSeq tests adjusted for the covariates of T stage and smoking ( ). The bar length represents log10 LDA score. Differences in the relative abundance were further tested by pairwise Mann-Whitney test and Tukey HSD post hoc as shown on the right panels. ; ; . d Hierarchical cluster analysis using distance matrix of six discriminative bacterial species (LDA , AUC ) classified the surveyed tissue samples into two clades.
the most up-regulated genes ( ). Examples of down-regulated genes included tumor suppressors CRISP3 ( ) and EMP1 ( ).
An enrichment analysis categorizing 1292 annotated tumor-associated differentially expressed genes (DEGs) ( ) into the Kyoto Encyclopedia of Genes and Genomes (KEGG) database found that most altered canonical pathways were related to carcinogenic signaling and immune regulation ( ) (Fig. 2c and Supplementary Table 7). For example, extracellular matrix-receptor (ECM-receptor) interaction ( ) involving the regulation of cellular movement, proliferation and differentiation was the most enriched pathway, in which a variety of DEGs that encode collagen, integrin and laminin were highly expressed ( ) (Fig. 2d). Disorders of a variety of metabolism pathways were observed when the annotated down-regulated DEGs
( ) were summarized for functional analysis (Supplementary Fig. 3e and Supplementary Table 7). For instance, the downregulation of tyrosine metabolism has been implicated in tumor progression of several types of cancers such as esophageal cancer , oral cancer , and hepatocellular carcinoma .
Survival analysis further identified 62 tumor- and 48 AN associated DEGs with prognostic values for OSCC ( ) (Supplementary Table 8). These included MMP1 (involved in proliferation and differentiation) ( ), SERPINE1 (involved in tumor migration and tissue remodeling) ( ), COL5A2 (involved in cellular proliferation and invasion) ( ), and FAP (involved in familial adenomatous polyposis) ( ) that were significantly associated with poor disease-specific survival (Fig. 2e). The identification of these DEGs provides a prognostic landscape that could be
Table 1 | Discriminative bacterial genus and species observed in OSCC tumor tissues compared with the paired adjacent normal tissues
Taxa Mean abundance ( sd) (%) LEfSe test Tukey HSD test MWU test ANCOM-BC2 test ALDEx2 test ZicoSeq test Representative ASV
AN Tumor LDA group LDA score LDA value Difference (%) value p value value value p value value value p value value value value
Genus
Fusobacterium Tumor 4.78 0.0000 12.45 (8.71, 16.20) 0.0000 0.0000 0.0000 1.12 4.27 0.0000 0.0002 5.72 0.0000 0.0001 0.0600 0.0979
Treponema Tumor 4.29 0.0180 4.07 (1.85, 6.29) 0.0004 0.0001 0.0028 0.81 2.57 0.0114 0.0235 2.57 0.0156 0.9819 0.0100 0.0017
Peptoanaerobacter Tumor 3.82 0.0017 1.19 (0.47, 1.92) 0.0014 0.0000 0.0001 1.46 6.16 0.0000 0.0000 3.58 0.0012 0.3104 0.0100 0.0070
Peptostreptococcus Tumor 3.80 0.0019 1.20 (0.26, 2.14) 0.0125 0.0004 0.0066 0.71 2.74 0.0070 0.0151 3.77 0.0004 0.1432 0.0300 0.0468
Catonella Tumor 3.54 0.0001 0.62 (0.29, 0.95) 0.0003 0.0000 0.0003 0.72 3.23 0.0017 0.0044 3.94 0.0003 0.1222 0.0100 0.0063
Peptococcus Tumor 3.06 0.0002 0.23 (0.10, 0.36) 0.0008 0.0001 0.0018 0.87 4.23 0.0001 0.0003 3.78 0.0008 0.2074 0.0100 0.0017
Streptococcus AN 4.55 0.0000 -7.39 (-10.66, -4.13) 0.0000 0.0000 0.0000 -0.91 -3.64 0.0004 0.0013 -3.04 0.0046 0.7240 0.0100 0.0313
Veillonella AN 4.30 0.0000 -3.97 (-5.22, -2.71) 0.0000 0.0000 0.0000 -1.11 -3.95 0.0001 0.0005 -3.98 0.0002 0.0660 0.0100 0.0067
Rothia AN 3.98 0.0000 -1.80 (-3.62, 0.01) 0.0518 0.0000 0.0000 -1.27 -4.40 0.0000 0.0001 -4.38 0.0001 0.0220 0.0100 0.0208
Granulicatella AN 3.69 0.0000 -0.96 (-1.58, -0.34) 0.0027 0.0000 0.0000 -0.85 -3.21 0.0017 0.0044 -4.02 0.0002 0.0768 0.0100 0.0428
Schaalia AN 3.64 0.0000 -0.85 (-1.26, -0.45) 0.0000 0.0000 0.0000 -0.98 -4.03 0.0001 0.0004 -4.00 0.0003 0.0867 0.0100 0.0013
Actinomyces AN 3.57 0.0000 -0.68 (-0.93, -0.43) 0.0000 0.0000 0.0000 -1.08 -5.32 0.0000 0.0000 -4.92 0.0000 0.0073 0.0100 0.0081
Lautropia AN 3.36 0.0002 -0.42 (-0.68, -0.15) 0.0023 0.0000 0.0005 -0.87 -4.08 0.0002 0.0007 -3.18 0.0059 0.5965 0.0100 0.0013
Corynebacterium AN 3.30 0.0000 -0.36 (-0.59, -0.14) 0.0018 0.0000 0.0002 -0.75 -3.53 0.0008 0.0026 -3.96 0.0006 0.1612 0.0100 0.0013
Eubacterium AN 3.28 0.0393 -0.38 (-0.67, -0.09) 0.0117 0.0007 0.0113 -0.56 -2.42 0.0174 0.0337 -0.63 0.5459 1.0000 0.0100 0.0208
Halomonas AN 3.22 0.0024 -0.28 (-0.65, 0.09) 0.1305 0.0045 0.0545 -1.15 -6.82 0.0000 0.0001 -1.99 0.1134 0.9503 0.0100 0.0470
Megasphaera AN 3.16 0.0028 -0.27 (-0.47, -0.08) 0.0063 0.0003 0.0060 -0.63 -2.98 0.0042 0.0096 -2.42 0.0295 0.9625 0.0100 0.0072
Lancefieldella AN 3.11 0.0115 -0.24 (-0.55, 0.08) 0.1447 0.0001 0.0018 -1.06 -4.66 0.0000 0.0001 -1.20 0.2663 1.0000 0.0100 0.0070
Phocaeicola AN 3.04 0.0040 -0.22 (-0.41, -0.03) 0.0212 0.0006 0.0109 -0.74 -3.57 0.0008 0.0027 -2.49 0.0387 0.8908 0.0100 0.0070
Species
Fusobacterium nucleatum Tumor 4.68 0.0000 9.75 (5.91, 13.59) 0.0000 0.0000 0.0000 1.15 4.54 0.0000 0.0001 3.92 0.0002 0.1623 0.0100 0.0229 4c305539242bc00f72118f6a70d5d103
Treponema medium Tumor 4.02 0.0030 2.06 (0.97, 3.16) 0.0003 0.0000 0.0013 1.01 4.01 0.0001 0.0005 3.15 0.0036 0.8050 0.0100 0.0043 88d5e4fcfd68b0cef9f267bd04d6b934
Peptostreptococcus stomatis Tumor 3.67 0.0021 0.95 (0.08, 1.82) 0.0323 0.0005 0.0094 0.67 2.78 0.0062 0.0132 3.79 0.0004 0.2569 0.0100 0.0183 feb281c58c97b847d8e32aa9dae2174b
Gemella morbillorum Tumor 3.52 0.0000 0.67 (0.23, 1.11) 0.0034 0.0000 0.0011 0.39 1.93 0.0569 0.0879 4.73 0.0000 0.0140 0.0100 0.0488 66358cc06be2067caaeab6047c327466
Catonella morbi Tumor 3.51 0.0001 0.62 (0.29, 0.95) 0.0003 0.0000 0.0005 0.71 3.45 0.0008 0.0024 4.00 0.0003 0.1628 0.0100 0.0092 de1e79ff05735a3eb21b59e27188410c
Peptoanaerobacter yurli Tumor 3.38 0.0145 0.47 (0.00, 0.95) 0.0514 0.0008 0.0133 1.08 5.64 0.0000 0.0000 2.61 0.0246 0.9239 0.0100 0.0213 74a1666eb102cd17c22028f87467e2e4
Peptococcus simiae Tumor 3.06 0.0002 0.23 (0.10, 0.36) 0.0008 0.0001 0.0019 0.88 4.68 0.0000 0.0001 3.76 0.0011 0.3176 0.0100 0.0025 e5944afd1dc39fe0c43f907049fd6ac4
Porphyromonas gingivalis AN 3.96 0.0016 -1.60 (-2.77, -0.42) 0.0080 0.0000 0.0003 -0.78 -3.33 0.0013 0.0037 -2.62 0.0179 0.9642 0.0100 0.0051 204ae45c366a21148f2675b55295831c
Schaalia odontolytica AN 3.57 0.0000 -0.72 (-1.11, -0.32) 0.0004 0.0000 0.0000 -0.78 -3.63 0.0004 0.0015 -3.86 0.0004 0.2285 0.0200 0.0043 845954a9086d53d7993b092f560b4fcc
Lautropia mirabilis AN 3.32 0.0004 -0.42 (-0.66, -0.19) 0.0005 0.0000 0.0007 -0.93 -5.00 0.0000 0.0001 -2.86 0.0166 0.8482 0.0100 0.0025 14e61b2c1523a6a67df3d576d3c01fdc
Actinomyces oris AN 3.31 0.0000 -0.41 (-0.57, -0.25) 0.0000 0.0000 0.0000 -0.86 -5.31 0.0000 0.0000 -4.51 0.0001 0.0855 0.0100 0.0025 37e783541d8270c9329ce720d8331512
Campylobacter gracilis AN 3.00 0.0009 -0.18 (-0.37, 0.01) 0.0679 0.0000 0.0013 -0.72 -4.73 0.0000 0.0001 -2.50 0.0324 0.9517 0.0100 0.0124 42cb1a145d5871ce13065a619775cab6
Genomes (KEGG) pathways significantly enriched by up-regulated DEGs in OSCC . Sizes of the circles indicate the number of DEGs in the pathway. d KEGG gene network visualizing up-regulated DEGs in OSCC ( ).
e Kaplan-Meier estimate for 3-year disease-specific survival based on the abundance levels of four DEGs (MMP1, SERPINE1, COL5A2, and FAP). The log-rank test was used to determine significance.
Fig. 2 | Host transcriptome profiling in OSCC by lncRNA-seq. a Volcano plot showing differentially expressed genes (DEGs; abs( ) ) between OSCC tumor and tissues. Dots in red and blue indicate up- and down-regulated DEGs in tumors when compared to AN tissues, respectively. b Expression of four representative genes (LAMC2, MMP1, EMP1, and CRISP2) involved in OSCC tumorigenesis. c Top 10 Kyoto Encyclopedia of Genes and
instrumental for risk stratification and personalized treatment planning in OSCC.

Interactions between oral microbiota and host transcriptome in OSCC

In our study, we aim to explore the potential associations between pathogenic bacteria and the transcriptome in OSCC by investigating the correlations between differentially expressed genes (DEGs) and the relative abundance of tumor-enriched bacterial species in OSCC tumor microenvironment. Using Spearman correlations analysis, we noted associations between 814 host DEGs ( 370 tumor- and 444 AN-associated) and 7 tumorenriched bacterial species that were frequently found in 34 tumor tissues with available bacterial 16 S and host lncRNA-seq data, resulting in 990 statistically significant bacteria-transcriptome associations ( ) (Fig. 3a and Supplementary Table 9a). Notably, among the most significant of these associations ( ), displayed coherent patterns; that is, a positive correlation of tumor-enriched bacteria with upregulated genes (Up & Positive) or a negative correlation with downregulated genes (Down & Negative) (Fig. 3b, c). For example, we observed a significant positive correlation between the abundance of C. morbi and the expression of HSPH1 (involved in upregulation of Wnt signaling pathway related to cellular proliferation and migration) (rho ) (Fig. 3d). Similarly, a positive correlation was found between . stomatis abundance and GALNT6 expression (involved in tumor progression and metastasis) (rho ). Conversely, tumor suppressor genes such as PRKN (rho ) and (rho ) exhibited negative correlation with C. morbi abundance. Additionally, COLGALT2, a gene downregulated in breast cancer, showed a negative correlation with multiple bacterial species including C. morbi (rho ) and T. medium (rho ). It was also noteworthy to mention the positive correlation observed between C. morbi and T. medium in the surveyed OSCC tumors (SparCC correlation ) (Fig. 3c and Supplementary Table 9b). We would like to clarify that the identification of these correlations may contribute to a better understanding of the complex interactions between the microbiota and the host response, but they require subsequent functional studies to elucidate any potential causal relationships. Due to the inherent limitations in taxonomic resolution at the species level with the V3-V4 region of the 16 S rRNA gene, we also conducted correlation analyses at the genus level (Supplementary Table 10a, b). These analyses serve as a basis for more targeted hypotheses that could inform future experimental work designed to explore the intricacies of hostmicrobiota interactions in OSCC.
Up-regulated host DEGs positively associated with tumor-enriched bacterial species (Up & Pos, N = 206, ) were summarized for Gene Ontology (GO) biological processes to characterize the potential of pathogenic bacteria in OSCC pathogenesis (Supplementary Table 11a). As shown in Fig. 3e, the top enriched GO terms included cell cycle, motility, adhesion, proliferation and migration. This was consistent with the Gene Ontology (GO) enrichment results obtained at the genus level (Supplementary Table 11b). For example, CDK6 was positively associated with T. medium; this gene was able to phosphorylate retinoblastoma (Rb) in the G1 phase of the cell cycle by derepressing E2F to promote cellular proliferation. Similarly, several genes encoding proteins involved in tumor shedding, adhesion and migration, such as collagens (COL4A5, COL4A6), integrins (ITGB4) and laminins (LAMA3), were overexpressed and positively associated with pathogenic bacteria in OSCC tumor microenvironment.

Transcriptome deregulation by DNA methylation

Paired tumor and AN tissues from 38 randomly selected OSCC patients were profiled for DNA CpG methylation (Supplementary Tables 2 and 12). A total of 4,584,317 CpG sites with reads per sample were characterized, of which were located within the promoter regions (Supplementary Fig. 4a). We quantified methylation signals on each chromosome using a 1 kb sliding window to smooth the distribution, which clearly differentiated the surveyed tumors from AN tissues by principal component analysis
(PCA, ) (Supplementary Fig. 4b). We focused on differentially methylated regions (DMRs) mapped to gene promoter regions (up to 3 kb ) because of their association with gene expression. Overall, 17,056 hyper(increased CpG methylation, meth.diff > 1.0) and 26,474 hypo-methylated (decreased CpG methylation, meth.diff ) promoter regions that regulate 9052 and 13,087 genes, respectively, were identified when comparing OSCC tumors with AN tissues ( ) (Supplementary Fig. 4c and Supplementary Table 13). Hyper-methylated genes included previously reported targets of recurrent hyper-methylation in OSCC, such as DDAH2, CCNA1, DCC, as well as some cancer-associated genes including HRAS (Supplementary Fig. 4d). Similarly, examples of hypo-methylated genes included PI3, AIM2, PTHLH, IFNG and CEACAM1. We designated 477 suppressed and 636 overexpressed DEGs with CpG hyper- (Hyper-Down) and hypo-methylation (Hypo-Up) in their promoter regions, respectively, given the potential of host transcriptome dysregulation by epigenetic modifications (Supplementary Fig. 4e, f and Supplementary Table 14).

Bacteria-associated epigenetic aberrance on host gene dysregulation in OSCC

Enrichment of pathogenic bacteria in OSCC tumor microenvironment may lead to dysregulation of epigenetic modifications. To test this hypothesis, we established the association between DMRs and tumor-enriched bacterial species using Spearman correlation (Supplementary Table 15a). Overall, 13,172 DMRs (responsible for 8690 genes) and 7 tumor-enriched bacterial species formed 16,630 bacteria-methylation associated pairs ( ), with P. simiae and F. nucleatum contributing to ( ) of positive associations with hypermethylation (Hyper & Pos) and 50.4% (1071/2125) of negative associations with hypomethylation (Hypo & Neg), respectively (Fig. 4a, b). An integrative analysis of bacteria-methylation and bacteriatranscriptome correlations further identified 15 suppressed DEGs that might be silenced by bacteria-associated CpG hypermethylation in their promoter regions (Fig. 4c and Supplementary Table 15b). For example, the enrichment of C. morbi was simultaneously correlated with the hypermethylation (Hyper & Pos) and inhibited gene expression (Down & Neg) of AOX1 and PITX1 (Fig. 4d). Other examples of tumor suppressors include DHRS3 and CES1. Meanwhile, inverse correlations of NRG1 and ITGB4 between DNA CpG hypomethylation and gene overexpression were found to be linked to T. medium and C. morbi abundance, respectively (Fig. 4e), implying the potential of tumor-enriched bacteria to alter DNA methylation that might translate into activated gene expression. NRG1 as an oncogene binds to and activates members of the ErbB family of receptor tyrosine kinases, triggering downstream signaling pathways, such as the PI3K/AKT and MAPK/ERK pathways, which are involved in cell growth, survival, and proliferation. ITGB4 is aberrantly expressed in several cancers including breast, colorectal, and lung cancers. The bacteria-methylation correlations at the genus level were established and shown in Supplementary Table .
To better understand the potential of F. nucleatum in OSCC pathogenesis by altering host gene regulation, we first applied 384 up-regulated DEGs positively associated with 7 tumor-enriched bacterial species (Up & Pos, ) for a Gene Set Enrichment Analysis (GSEA) (Supplementary Table 17a). Altogether 368 connections were observed (Supplementary Fig. 5 and Supplementary Table 17b), with F. nucleatum interacting with 20 upregulated DEGs that involved 18 hallmark pathways (Fig. 5a). Among them, for example, SNAI2 ( ), potentially triggered by F. nucleatum and C. morbi, is one of the key transcription factors that regulates the Epithelial-Mesenchymal Transition (EMT) process. Cells undergoing EMT exhibit a loss of epithelial markers such as E-cadherin and claudins, thereby detach from the primary tumor, invade surrounding tissues, and eventually metastasize to distant sites. As F. nucleatum and F. periodonticum, the two species with close phylogenetic relationship and similar biological functions, accounted for and of
Fusobacterium 16S rRNA reads, we also performed a GSEA inferred from the genus-level bacterial-transcriptome connections and observed 40 upregulated DEGs involving 22 enriched hallmark carcinogenic pathways that were positively associated with Fusobacterium (Supplementary Table 17c, d), including six hallmark genes (LAMA3, INHBA, SNAI2, NT5E, MYLK and TPM1) within the EMT process (Supplementary Fig. 6).
We then applied an in vitro cell-based model by co-culturing F. nucleatum with cancer (SAS) and non-cancer (HGK12) human oral epithelial cells, respectively, to profile cellular gene transcriptomes using lncRNA-seq (Fig. 5b and Supplementary Table 18). These two cell lines showed differentially
Fig. 3 | Interactions between tumor-enriched bacterial species and host DEGs associated with OSCC. a Bar plot of bacteria-transcriptome Spearman correlation showing the abundance of host DEGs in OSCC tumor tissues associated with tumorenriched bacterial species ( ). Up and Down indicate up-regulated and downregulated DEGs, respectively; Pos and Neg indicate positive and negative associations, respectively. C. morbi Catonella morbi, G. morbillorum Gemella morbillorum, P. yurli Peptoanaerobacter yurli, P. simiae Peptococcus simiae, P. stomatis Peptostreptococcus stomatis, F. nucleatum Fusobacterium nucleatum, T. medium Treponema medium. b Heat map depicting highly significant bacteria-transcriptome correlations ( ). Color panel indicates the Spearman correlation coefficients. Asterisks indicate significance level (** ; ). Columns are ranked by the average correlation coefficients of DEGs with tumor-enriched bacteria.
c Network visualizing significant correlations between bacteria-transcriptome (solid line, Spearman ) and bacteria-bacteria (dashed line, SparCC correlation ). d Scatter plots showing four representative correlations between bacteria and DEG in OSCC tumor tissues, where the strength of correlation (Spearman rho) and significance ( value) are shown in each plot. Marginal boxplots depict overall gene expression (right) and bacterial 16S rRNA gene abundance (top), with Mann-Whitney test between tumor and AN tissues. e Gene Ontology (GO) pathways significantly enriched by up-regulated host DEGs associated with tumorenriched bacteria. GO terms with functional similarity are clustered in the semantic space by REVIGO. B & H corrected values are indicated by color panel. The size of the circles indicates the number of DEGs associated with the pathways.
expressed transcriptomes, with 820 up- and 478 down-regulated DEGs commonly detected (Supplementary Fig. 7a and Supplementary Table 19). While activated genes in HGK12 were involved in neutrophil mediated immune response, cytokine-mediated signaling pathway and regulation of autophagy, the gene expression in SAS infected with F. nucleatum was mainly enriched in the regulation of signal transduction, programmed cell death and apoptotic process (Supplementary Fig. 7b and Supplementary Tables 20 and 21), which may be in part due to that fact that these two cells have completely different properties. Interestingly, among 148 DEGs coherently associated with F. nucleatum 16S reads in the surveyed OSCC tumor tissues (Up & Pos and Down & Neg, ), at least 15 and 10 genes were confirmed to be overexpressed or suppressed in the two co-cultured cell lines, respectively , ) (Fig. 5b, c and Supplementary Table 19). These include the overexpression of the hallmark genes SNAI2, LIMA1, PLEK2 and RAB31 involved in multiple GSEA pathways (Fig. 5c, d). In contrast, SLC16A7, a gene encoding a membrane transport solute carrier (family 16 member 7) and commonly reduced in various types of cancer, was negatively associated with . nucleatum (rho ) and significantly suppressed in situ ( ) and in vitro ( HGK 12 : ; SAS: ). Using a real-time PCR, we verified the overexpression of 5 representative hallmark genes involved in the EMT pathway in vitro, including SNAI2 (F. nucleatumassociated), INHBA and LAMA3 (Fusobacterium-associated), and LAMC2 (Fig. 5e and Supplementary Table 23), although their protein levels, as measured by western blot, were controversial (Supplementary Fig. 7c and Supplementary Table 24). Our findings in vitro indicate a significant increase in the mRNA expression levels of three EMT hallmark genes (LAMA3, INHBA, SNAI2) upon F. nucleatum infection. Additionally, we included LAMC2, the protein-coding gene for the gamma subunit of Laminin332, which is part of the same protein complex as LAMA3, as supplementary evidence. We also investigated the protein laminin332 and found that the protein expressions of two subunits were potentially increased by . nucleatum infection in both HGK12 and SAS cell lines. This observation aligns with the results obtained from RNAseq analysis of cocultured cell lines but the underlying mechanisms of the carcinogenetic pathways in which these genes are involved in OSCC warrants further investigation.

Discussion

Host-microbiota maladaptation has been reported in various types of cancer . Our study through integrative analysis of bacteriatranscriptome and bacteria-methylation correlations in HPV-negative OSCC patients revealed complex networks between oral microbiota dysbiosis and host genetic and epigenetic abnormalities, functionally implicated in various cancer-related pathways. Our data also suggests that dysregulation of host gene transcriptome might be influenced by tumor-enriched bacteria or by bacteria-associated epigenetic modifications. An in vitro model further confirmed that F. nucleatum, an oral bacterial species closely associated with OSCC tumorigenesis, could activate hallmark genes involved in multiple cancer-associated pathways. These findings may serve
as a precursor for hypothesis-driven study to better understand the molecular mechanisms of pathogenic bacteria underlying OSCC pathogenesis.
Carcinogenesis might arise from changes in chronic host-microbe interactions upon colonization by key pathogens. Our multi-omics data analysis highlighted several cancer-associated pathways that might be directly or indirectly triggered by tumor-enriched oral bacteria in the OSCC tumor microenvironment. For example, our integrative analysis in situ and in vitro both implied the regulatory role of F. nucleatum on SNAI2 in OSCC, in line with the role of . nucleatum virulence factor FadA to modulate the Ecadherin/ -catenin signaling via activating TNF proinflammatory pathway, ECM remodeling, and Wnt signaling (Fig. 6) . Laminin is the main component of the extracellular matrix (ECM) and plays key functions in cell adhesion, cell migration and signal transduction , among which Laminin-332 is a primary member of the laminin family and contains three chains encoded by the LAMA3, LAMB3 and LAMC2. Meanwhile, -catenin could be translocated to the nucleus to form a complex with T-cell factor protein (TCF) and activate EMT-related genes, such as INHBA, LAMA3, LAMC2, NT5E and MMP1, thereby contributing to the inflammatory and oncogenic responses. Cells undergoing EMT exhibit loss of epithelial markers such as occludins, claudins, and E-cadherin, and acquire concomitant expression of Vimentin, N-cadherin, and Fibronectin , in which SLUG protein (SNAI2) has a fundamental role by suppressing several cellcell adhesion genes including E-cadherin expression . Additionally, the pro-inflammatory cytokine acts as an inflammatory mediator to trigger EMT of tumor cells and promote metastasis . It could also induce SNAI family stabilization through activation of the NF- pathway .
In addition to F. nucleatum, several other oral bacteria, such as Treponema denticola, could promote cancer aggressivity via crosstalk between the integrin/FAK and TLR/MyD88 signaling pathways . Integrin-associated PI3K/Akt signaling has been reported to be activated by Peptostreptococcus anaerobius, which regulates cell cycle progression . Both bacterial species were found to be enriched in the surveyed OSCC tumor tissues with limited statistical significance. Interestingly, T. medium and P. stomatis, two bacterial species phylogenetically related to . denticola and . anaerobius, respectively, significantly increased in relative abundance in the OSCC tumor microenvironment. P. stomatis has been associated with various oral diseases even cancer . For example, associated with . stomatis could activate the canonical Wnt pathway through -catenin/ MMP9-mediated signaling that regulates cell proliferation, differentiation and EMT . Three hallmark genes related to EMT, including GJA1, MMP1, and SNAI2, which are capable to promote cancer cell migration and invasion, were also positively associated with . stomatis in the surveyed OSCC tumor tissues . It is important to note that the dysregulation of host gene expression might be influenced by the contributions from several pathogenic bacteria. LIMA1 (belonging to apical junctions) initially identified as a differentially expressed gene in oral epithelial cell carcinogenesis , for example, were positively associated with C. morbi, F. nucleatum and . stomatis, highlighting the complexity of the host-microbiota interactions in the OSCC tumor microenvironment. The effects of

transcriptome expression (DEG) are represented by filled circles, with sizes corresponding to the values. The bar plot shows Spearman correlation coefficients of CpG methylation and transcriptome expression with the abundance of tumorenriched bacteria. d Scatter plots depicting AOX1 and PITX1 hypermethylation (top panel) and down-regulated expression (bottom panel) in relation to C. morbi, where the strength of correlation (Spearman rho) and significance ( value) are shown in each plot. Marginal boxplots depict overall gene expression/methylation (right) and bacterial 16S rRNA gene abundance (top), with Mann-Whitney test between tumor and AN tissues. e Scatter plots depicting NRG1 and ITGB4 hypomethylation (top panel) and up-regulated expression (bottom panel) in relation to T. medium and C. morbi, respectively.
Fig. 5 | OSCC-related carcinogenic potential of Fusobacterium nucleatum in situ and in vitro. a A Gene Set Enrichment Analysis (GSEA) inferred from 384 upregulated DEGs positively associated with 7 tumor-enriched bacterial species (Up & Pos, ). The network displays 20 DEGs involving 18 hallmark pathways associated withs F. nucleatum. b Schematic illustration of an in vitro cell model by co-culturing F. nucleatum in human oral epithelial cells HGK12 and SAS to profile the cellular gene transcriptome using lncRNA-seq. The number of DEGs in situ (abs( ) with coherent association with . nucleatum 16S reads (Up & Pos and Down & Neg, ); the number of DEGs in situ with coherent dysregulation in at least one cell line (abs( ) and confirmed by another cell line (abs(log2FC ). c . nucleatum-associated DEGs observed simultaneously in situ and in vitro. Filled circle indicates status of transcriptome expression (DEG) in vitro and in situ, respectively, with sizes
corresponding to the values. The bar plot shows Spearman correlation coefficients between DEG transcriptome expression and . nucleatum abundance in OSCC tumor tissues. Five . nucleatum-associated GSEA hallmark genes are highlighted in bold. d Scatter plots depicting up-regulation of SNAI2 and LIMA1 in relation to F. nucleatum in OSCC tumor tissues, where the strength of correlation (Spearman rho) and significance ( value) are shown in each plot. Marginal boxplots depict overall gene expression (right) and bacterial 16S rRNA gene abundance (top), with Mann-Whitney test between tumor and AN tissues. e RT-PCR analysis for the indicated overexpression of four GSEA hallmark genes (SNAI2, INHBA, LAMA3 and LAMC2) in HGK12 and SAS cells co-cultured with F. nucleatum or F. mortiferum. Data are presented as mean SD and significance ( ) are determined by two-tailed unpaired Student’s test based on .
Fig. 6 | Schematic model hypothesizing the carcinogenetic potential of pathogenic bacteria in OSCC pathogenesis. F. nucleatum virulence factor FadA has been reported to activate the E -cadherin -catenin associated cell signaling, which further activates the TNF proinflammatory pathway, ECM remodeling, and Wnt signaling. Subsequently, -catenin can be translocated to the nucleus to form a complex with T-cell factor protein (TCF) and activate EMT-related genes, such as SNAI2, LAMA3, INHBA, and MMP1. Moreover, F. nucleatum and T. denticola have
been reported to promote cancer aggressivity via crosstalk between the integrin/FAK and TLR/MyD88 signaling pathways. Integrin-associated PI3K/Akt signaling can be activated by . anaerobius, which regulates cell cycle progression. Gal-GalNAc on tumor cells is the receptor of Fap2 to recruit F. nucleatum to the tumor site. In addition, hypomethylation of NRG1 may activate the PI3K/AKT pathway while hypermethylation of AOX1 may promote cellular invasion and metastasis.
LIMA1 can extend from cell migration and cytoskeleton dynamics to cell cycle, gene regulation, angiogenesis, and lipid metabolism, providing new ideas for future exploration of cancer treatment strategies targeting this gene .
Recent studies suggest that epigenetic alterations may play important roles in the initiation and propagation of cancer . Epigenetic mechanisms have also been recognized as a critical player at the interface between the human microbiome and the intestinal epithelial cell . For example, exposure to commensal microbiota induced localized DNA methylation changes, which is necessary for proper intestinal homeostasis . In contrast, F. nucleatum and H. hathewayi, two bacterial pathogens associated with colorectal cancer, are able to drive host colonic epithelial cell promoter hypermethylation of tumor suppressor genes . Our data suggest that dysregulation of host gene transcriptome could be influenced by bacteriaassociated epigenetic abnormality. For example, hypermethylation and downregulation of PITX1 and AOX1 were both associated with C. morbi. PITX1 has been identified as a potential tumor-suppressor gene related to cell apoptosis and was down-regulated by DNA hypermethylation in gastric cancer and esophageal squamous cell carcinoma (ESCC) . PITX1 suppresses tumorigenicity by downregulating the RAS pathway through RASAL1, a RAS-GTPase-activating protein . Similarly, loss of , probably linked to its hypermethylated promoter, contributes to the transition from low-grade to high-grade during bladder carcinoma progression, thereby promoting invasion and metastasis . Meanwhile, hypomethylation of several oncogenes such as NRG1, ITGB4 and GALNT6 associated with tumor-enriched bacteria suggests the potential of microbiota to promote host gene expression through epigenetic activation. For example, GALNT6 has been shown to promote tumorigenesis and metastasis by catalyzing mucin-type O-glycosylation-mediated stabilization of MUCl and fibronectin (FN) in breast cancer cells . Besides, recent study has reported that . nucleatum could induce the significant decline of modifications in CRC resulting in the improvement on CRC aggressiveness . The metastasis of colorectal cancer has also been demonstrated to be promoted by . nucleatum through miR-1322/CCL20 axis and M2 polarization, which
indicating the importance of host-microbiome interactions . However, the mechanisms by which oral microbiota programmed gene promoter methylation leads to dysregulated expression in OSCC remains to be elucidated.
Our study has several limitations. First, oral cavity biopsy from healthy participants were not included. Although AN tissues from OSCC patients well serve as controls, the joint effects of environmental exposures and clinical variables between cancer patients and healthy individuals are challenging to account for long-term confounding. Also, some examination results related to oral condition were not available, which may overlook the synergistic effect of some confounders on oral microbial dysbiosis and need to be remedied in subsequent studies. Second, prognostic evaluation for OSCC is based on clinical TNM classification, but this staging system is not sufficient for optimal prognostication and may be supplemented by other methods such as histological grading. Third, this study lacks a direct link between host gene transcriptome and DNA CpG methylation since only ten overlapping tumors were available, which warrants a more comprehensive insight into the epigenetic programming inferred from a larger cohort. However, it is interesting to observe coherent correlations through integrative analysis of bacteria-transcriptome and bacteria-methylation correlations. Especially, the dysregulation of several F. nucleatum-associated hallmark genes involving EMT pathways could be well verified in vitro using lncRNAseq, RT-PCR and western blot. Fourth, our study reveals association but not causality of the oral microbiota in the pathogenies of OSCC; further studies using in vivo and in vitro models will be helpful to identify specific microbiota-host connections underlying the mechanism. Last but not least, the resolution in defining bacterial species inferred from 16S rRNA V3-V4 region reads may be limited and bacterial strains cannot be taken into account in the current study. However, we provide the most representative ASV sequences and encourage further analysis based on different levels of taxonomic classification as well as traditional bacterial culture.
In summary, we applied multi-omics approaches to reveal complex networks between oral mucosal microbiota and host gene dysregulation in the pathogenesis of OSCC. We found differentially abundant oral bacteria,
dysregulated host gene expression, and aberrant DNA CpG methylation in tumor microenvironment, with enriched functions related to various cancer-related pathways. Integrative analysis between bacteriatranscriptome and bacteria-methylation correlations reveals that dysregulation of host gene transcriptome might be influenced by tumor-enriched bacteria or by bacteria-associated epigenetic abnormality. Our findings extend the current understanding of the host-microbiota interactions in the pathogenesis of OSCC, which provides important insights into the genetic and functional basis as potential diagnostic markers and therapeutic targets for formulating more effective prevention and intervention strategies to manage this morbid entity.

Methods

Patient recruitment

A total of 98 primary OSCC patients of Han Chinese ethnicity administrated to two hospitals in Hong Kong (Prince of Wales Hospital and United Christian Hospital) were recruited between October 2015 and January 2020. Demographic and clinical information at enrollment, including age, gender, smoking, alcohol consumption, location of the tumor, T stage and N stage, were collected in Supplementary Table 2. No cases of metastasis were included in this study.

Ethics approval

This study was approved by The Joint Chinese University of Hong KongNew Territories East Cluster Clinical Research Ethics Committee (CREC reference numbers 2015.396 and 2017.143). All patients agreed to participate with written informed consent and were reviewed by a pathologist.

Tissue specimen collection and DNA/RNA extraction

Tumor and AN tissues ( away from the margin of the tumor) were collected at the time of surgery as per our previous methodology . These specimens were briefly irrigated with sterile saline to wash away surface contamination and then stored at until further use. Around of fresh frozen tissues were manually homogenized into small pieces and treated with proteinase K at overnight. DNA and RNA from the disaggregated samples were simultaneously extracted using AllPrep DNA/RNA Mini Kit (Qiagen, Valencia, CA, USA) and eluted into elution buffer separately following the manufacturer’s protocol. DNA from tumor tissues were tested for HPV infection using a PCR-based amplicon sequencing assay as previously described .

Microbiota 16S rRNA gene V3-V4 amplicon sequencing

Extracted DNA was used for oral microbiota profiling by sequencing the bacterial 16S rRNA gene hypervariable V3-V4 region (341F: -CCT ACG GGN GGC WGC AG-3′, 806R: -GGA CTA CNV GGG TWT CTA . A pair of dual unique 12 bp barcodes was indexed to each amplicon set through the forward and reverse primers; successful amplicons were equally pooled and sequenced on an Illumina MiSeq using paired-end 300 bp reads. For quality control, each sequencing batch included a mock DNA community, negative controls and technical replicates.

Microbiota 16S sequence data bioinformatics and statistical analysis

Demultiplexed short 16S reads passing quality filtering were imported into QIIME2 (v2021.4) to generate an amplicon sequence variant (ASV) table as previously described . The representative reads were further assigned at the species level using pplacer by inserting into a phylogenetic tree inferred from complete sequences of the bacterial 16S rRNA gene to maximize phylogenetic likelihood . ASVs with a total count after removing reads assigned to archaea, mitochondria or chloroplasts were retained, with operational taxonomic unit (OTU) count tables showing the bacterial reads per sample at different taxonomic levels. Paired tumor and AN tissues with more than 2000 reads were retained for further analysis. A phylogenetic tree was generated by inserting the representative ASV reads into the SLIVA 128 reference database using the SATe-enabled phylogenetic placement (SEPP)
method . Alpha diversity of the observed bacterial ASV reads based on richness, Shannon and Simpson indexes were calculated using diversity in the Vegan R package. GUniFrac and Bray-Curtis distance metrics were computed inferred from the ASV profile to differentiate community compositions (beta diversity) between cancer and control groups using permutational multivariate analysis of variance (PERMANOVA) with 9999 permutations using the adonis2 in the Vegan R package . In the effect size analysis, disease status (tumor vs. AN) controlled association between clinical variables, including gender, age, smoking, alcohol consumption, N and T stages, were performed. Discriminative bacterial taxa between tumor and AN tissues based on OTU count tables were estimated using linear discriminant analysis (LDA) effect size (LEfSe) analysis , with a cutoff of LDA , which was further validated by different compositional aware tools with bias correction, including ANCOM-BC2 and ZicoSeq , adjusted for the covariates of T stage and smoking ( ). In addition, a linear regression analysis test, non-parametric Mann-Whitney Wilcoxon rank sum test (MWU), Tukey’s honest significant difference (Tukey HSD) post hoc test, and a binomial generalized log-linear model in EdgeR were applied. Heatmaps of the most discriminative bacterial genera were generated using the heatmap. 2 in the gplots R package, with hierarchical clustering setting of “hclust(method=”ward.D”) and dist(method= “euclidean”)”. Logistic regression and receiver operating characteristic (ROC) curve with the calculation of area under the ROC curve (AUC) were used to assess the potential biomarkers identified for cancer case screening. Kaplan-Meier analysis was used for univariate survival analysis with log-rank test to determine statistical differences in survival outcomes. An optimal cutoff point inferred from the relative abundance of individual bacterial taxa in the surveyed samples was calculated to divide tumor tissues into “high” and ‘low” groups using the cutpointr R script. The Cox proportional hazard regression method in stepwise manner controlling for gender, age, post-surgery treatment and other factors reported significant in the univariate analysis was used for multivariate analysis of survival. A two-sided value and/or a false discovery rate (FDR)adjusted value or was used as the threshold for significance.

HPV detection and genotyping

HPV genotyping was performed using a PCR-based amplicon sequencing assay targeting the conserved L1 open reading frame (ORF) of HPV as previously described . In brief, a pair of dual 12-bp barcodes was indexed to the PCR amplicon using forward and reverse primers for demultiplexing. Short reads generated by Illumina MiSeq PE150 were blasted against a comprehensive PV reference database using UPARSE . An operational taxonomic unit (OTU) count table was created using a identity threshold, assigning each OTU with a PV type . Based on the IARC classification, 12 HPV types (HPV16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59) were ranked as high risk (HR) and considered oncogenic .

Host gene long non-coding RNA sequencing (IncRNA-seq) and transcriptome profiling

Tumor and AN tissues from a subset of 40 randomly selected OSCC patients were profiled for host genome-wide transcriptome using lncRNA-seq. In brief, total RNA was depleted for rRNA using QIAseq FastSelect rRNA remove kit and converted to directional RNA library using NEBNext Ultra II Directional RNA Library Prep Kit for Illumina NovaSeq using paired-end 150 bp reads. Pair-end sequencing reads were mapped to the hg38 human reference genome using STAR . Number of reads mapped to the gene exon regions was counted using featureCounts ; genes with a mean Transcripts Per Kilobase Million (TPM) greater than 1 were retained. Differentially expressed genes (DEGs) between tumor and AN tissues were identified using a binomial generalized log-linear model in using a cutoff of and (fold change) (up-regulated genes) or (downregulated genes). Age, smoking, alcohol consumption, and gender were included as confounding factors in the model. Functional enrichment of OSCC-associated DEGs were summarized using ToppGene Suite . The top 10 Kyoto Encyclopedia of Genes and Genomes (KEGG) terms with
statistical significance ( ) were visualized in order. To determine DEGs associated with disease-specific survival, the expression of each gene was labeled with a binary value of “High” or “Low” using “optimal” cutpoints determined by cutpointr function in R . The parameter of “metric” in cutpointr was set as “sum_sens_spec” to maximize the sum of sensitivity and specificity. The Kaplan-Meier method with CI for Cox proportional hazards ratio was used for univariate survival analysis using coxph function in R .

The Cancer Genome Atlas (TCGA) OSCC RNA-seq data analysis

A TCGA OSCC RNA-seq HTSeq raw read counts matrix along with metadata information (age, smoking, alcohol consumption, gender, and site of resection or biopsy) were downloaded using the TCGAbiolinks R package . This dataset contains tissue samples from 502 OSCC patients and 44 non-OSCC controls. Samples were further filtered using the following criteria: (1) one sample was removed due to the lack of “age” information, and (2) 206 samples were discarded because the “site of resection or biopsy” didn’t belong to oral cavity, resulting in 309 cases and 30 controls for the comparison with the Hong Kong (HK) OSCC cohort.

Network analysis between bacterial abundance and host gene expression

Host DEGs and OSCC-associated bacterial species were explored for bacteria-transcriptome interactions with Spearman correlation coefficients calculated using the cor.test function in R. Bacterial count tables were normalized in percentage for the correlation analysis. A value was used as the cutoff for significance. Sparse Correlations for Compositional Data (SparCC) was used to explore bacteria-bacteria correlations, with a correlation coefficient for significance. Heatmaps and networks were plotted using corrplots function in R and Cytoscape v3.7.1 , respectively. To characterize bacteria target pathways in OSCC pathogenesis, up-regulated host DEGs with positive association with tumor-enriched bacterial species were summarized for Gene Ontology (GO) biological processes using ToppGene suite . The obtained values were then corrected for multiple testing using the Benjamini & Hochberg (B & H) method. Enriched GO terms ( ) and corresponding values were used as the input for to reduce GO term redundancy and visualize semantic clustering of the identified GO terms. Meanwhile, DEGs associated with tumorenriched bacteria and their correlation coefficients were merged for a refined hallmark gene set enrichment analysis against the Molecular Signatures Database (MSigDB) using GSEA . The hallmarks effectively summarized most of the relevant biological processes of gene sets by reducing variation and redundancy.

Human DNA CpG methylation and bioinformatic analysis

Tumor and AN tissues from a subset of 38 randomly selected OSCC patients were characterized for host DNA CpG methylation using bisulfite sequencing. In brief, the total DNA was pre-prepared for illumina library using KAPA HTP Library Preparation Kit (Kapa Biosystems, USA). The library was exposed to sodium bisulfite using EpiTect DNA Bisulfite Kit (Qiagen, USA) and then probed with Methyl-Seq Capture probes using SeqCap Epi CpGiant Kit (Roche, USA) for Illumina NovaSeq using paired-end 150 bp reads. Paired-end sequencing reads were trimmed using Trimmomatic and aligned to the human reference genome (hg38) using Bismark . BAM files were sorted and deduplicated; only genomic regions with read coverage larger than 10 were included. Methylation levels were modeled using a logistic regression-based algorithm with overdispersion correction and Chi-square test implemented in the methylKit R package . Differentially methylated regions (DMRs) in tumors compared with AN tissues were identified using a cutoff of and meth.diff score (hypermethylation) or (hypomethylation). Age, gender, smoking, alcohol consumption, N and T stages were included as confounding factors in the model. DMRs were then assigned to the promoter region of genes using the annotatePeak function in the ChIPseeker R package . The promoter region was defined
as 3000 bp upstream and 3000 bp downstream of the transcription start sites (TSSs). When multiple DMRs were located within the same promoter region, we used a majority voting method to determine the methylation status of the gene. Spearman’s rank-order correlation test was used to explore the association between promoter-associated DMRs and OSCC-enriched bacterial species. A value was considered as statistical significance.

Fusobacterium nucleatum co-culture with host epithelial cell in vitro

To validate the potential of pathogenic bacteria to dysregulate host transcriptome revealed in OSCC tumor microenvironment, an in vitro cellbased model was established by co-culturing F. nucleatum in HGK12 cell, a healthy human gingival keratinocyte immortalized head and neck derived epithelial cell line (gifts from Prof. Lui, School of Biomedical Sciences, The Chinese University of Hong Kong) and SAS, a human oral tongue squamous cell carcinoma cell line (Cellosaurus, RRID:CVCL_1675). In brief, F. nucleatum subsp. vincentii, a strain isolated from an OSCC patient and confirmed by whole genome sequencing, was applied for cell-bacteria coculture system at a multiplicity of infection (MOI) of for 5 h at under anaerobic condition. HGK12 and SAS were co-cultured with . nucleatum once and infected cells were then incubated with fresh Epilife or DMEM/F12 medium, respectively, at with . After overnight of culture with or without bacteria, total RNA from cells was extracted using RNeasy Mini Kit (Qiagen, USA), followed by Illumina lncRNA-seq library preparation. Three independent repeats were performed for statistical power. In brief, pair-end sequencing reads were mapped to the hg38 human reference genome using STAR . Number of reads mapped to the gene exon regions was counted using featureCounts ; genes with a mean Transcripts Per Kilobase Million (TPM) greater than 1 were retained. Differentially expressed genes (DEGs) between treated and untreated cell lines were identified using a binomial generalized log-linear model in , with a cutoff of abs( ) and considered statistically significant. Functional enrichment of . nucleatum-associated DEGs were summarized using ToppGene Suite . The top 10 Kyoto Encyclopedia of Genes and Genomes (KEGG) terms with statistical significance ( ) were visualized in order.

Real-time PCR (RT-PCR) of mRNA gene expression

Total RNA was extracted from HGK12 and SAS cell lines co-cultured with F. nucleatum (MOI 1:100) by using TRIzol reagent. As a negative-bacterial control, Fusobacterium mortiferum, a suggested non-invasive and nonpathogenic species within the Fusobacterium genus, was utilized . One of total RNA was applied for generating complementary DNA by LunaScript RT SuperMix Kit (NEB, #E3010) followed by standard RT-PCR reaction with Luna Universal qPCR Master Mix Protocol (NEB, #M3003). The RTPCR primers used for gene expression validation are listed in Supplementary Table 22.

Western blot of protein expression

Protein samples were obtained from co-cultured HGK12 and SAS cell lines in the presence or absence of respective bacteria. The samples were collected using RIPA lysis buffer (Santa Cruz). The protein quantification was performed using Pierce BCA Protein Assay Kits (Thermo Fisher Scientific). A total of of protein from each sample was separated by or SDS-PAGE and transferred onto PVDF membranes for blotting. To prevent nonspecific binding, all membranes were blocked using non-fat milk in 1X TBST. Primary antibodies against SNAI2 (Cell Signaling Technology, 9585T, 1:500 dilution), INHBA (Abcam, ab128958, 1:500), LAMA3 (Abcam, ab151715, 1:500), LAMC2 (Santa Cruz, sc-28330, 1:1000) and actin (Santa Cruz, sc-47778, 1:1000), were then incubated overnight at , respectively, followed by the application of corresponding secondary antibodies for 1 h at room temperature. The target protein bands were detected using Clarity Western ECL Substrate (Bio-Rad) and imaged using Chemiluminescent Western Blot Imager AZURE 300 (Azure Biosystems). All
expression levels were determined by quantifying the grayscale values of the target protein bands relative to the reference beta-actin using ImageJ software. All blots and gels derive from the same experiment and they were processed in parallel.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.

Data availability

All sequences and codes are available upon request. The 16 S rRNA gene amplicon next-generation sequences analyzed in this study are available in NCBI SRA database under BioProject PRJNA822685 (accession numbers from SRR18595688 to SRR18595849). The bioinformatics pipelines and scripts used in this study have been deposited to the Github repository (https://github.com/lycai05/OSCC_host_bacteria_interactions).
Received: 9 April 2023; Accepted: 27 March 2024;
Published online: 08 April 2024

References

  1. Li, Q. et al. Role of oral bacteria in the development of oral squamous cell carcinoma. Cancers 12, 2797 (2020).
  2. Johnson, D. E. et al. Head and neck squamous cell carcinoma. Nat. Rev. Dis. Prim. 6, 1-22 (2020).
  3. Sepich-Poore, G. D. et al. The microbiome and human cancer. Science 371 https://doi.org/10.1126/science.abc4552 (2021).
  4. Guerrero-Preston, R. et al. 16 S rRNA amplicon sequencing identifies microbiota associated with oral cancer, human papilloma virus infection and surgical treatment. Oncotarget 7, 51320-51334 (2016).
  5. Ganly, I. et al. Periodontal pathogens are a risk factor of oral cavity squamous cell carcinoma, independent of tobacco and alcohol and human papillomavirus. Int. J. Cancer 145, 775-784 (2019).
  6. Chen, Z. et al. The intersection between oral microbiota, host gene methylation and patient outcomes in head and neck squamous cell carcinoma. Cancers 12 https://doi.org/10.3390/cancers12113425 (2020).
  7. Kamarajan, P. et al. Periodontal pathogens promote cancer aggressivity via TLR/MyD88 triggered activation of Integrin/FAK signaling that is therapeutically reversible by a probiotic bacteriocin. PLoS Pathog. 16, e1008881 (2020).
  8. Michmerhuizen, N. L., Birkeland, A. C., Bradford, C. R. & Brenner, J. C. Genetic determinants in head and neck squamous cell carcinoma and their influence on global personalized medicine. Genes Cancer 7, 182-200 (2016).
  9. The Cancer Genome Atlas Network Comprehensive genomic characterization of head and neck squamous cell carcinomas. Nature 517, 576-582 (2015).
  10. Maruya, S. et al. Differential methylation status of tumorassociated genes in head and neck squamous carcinoma: incidence and potential implications. Clin. Cancer Res. 10, 3825-3830 (2004).
  11. Rosas, S. L. et al. Promoter hypermethylation patterns of p 16 , O6-methylguanine-DNA-methyltransferase, and death-associated protein kinase in tumors and saliva of head and neck cancer patients. Cancer Res. 61, 939-942 (2001).
  12. Xia, X. et al. Bacteria pathogens drive host colonic epithelial cell promoter hypermethylation of tumor suppressor genes in colorectal cancer. Microbiome 8, 108 (2020).
  13. Jena, P. K. et al. Dysregulated bile acid synthesis and dysbiosis are implicated in Western diet-induced systemic inflammation, microglial activation, and reduced neuroplasticity. FASEB J. 32, 2866-2877 (2018).
  14. Song, X. et al. Oral squamous cell carcinoma diagnosed from saliva metabolic profiling. Proc. Natl Acad. Sci. USA 117, 16167-16173 (2020).
  15. . et al. Oncometabolite 2 -hydroxyglutarate is a competitive inhibitor of alpha-ketoglutarate-dependent dioxygenases. Cancer Cell 19, 17-30 (2011).
  16. Janney, A., Powrie, F. & Mann, E. H. Host-microbiota maladaptation in colorectal cancer. Nature 585, 509-517 (2020).
  17. Dayama, G., Priya, S., Niccum, D. E., Khoruts, A. & Blekhman, R. Interactions between the gut microbiome and host gene regulation in cystic fibrosis. Genome Med. 12, 12 (2020).
  18. Helmink, B. A., Khan, M. A. W., Hermann, A., Gopalakrishnan, V. & Wargo, J. A. The microbiome, cancer, and cancer therapy. Nat. Med. 25, 377-388 (2019).
  19. Nejman, D. et al. The human tumor microbiome is composed of tumor type-specific intracellular bacteria. Science 368, 973-980 (2020).
  20. Rubinstein, M. R. et al. Fusobacterium nucleatum promotes colorectal carcinogenesis by modulating E-cadherin/beta-catenin signaling via its FadA adhesin. Cell Host Microbe 14, 195-206 (2013).
  21. Rousselle, P. & Scoazec, J. Y. Laminin 332 in cancer: when the extracellular matrix turns signals from cell anchorage to cell movement. Semin. Cancer Biol. 62, 149-165 (2020).
  22. Jung, A. R., Jung, C.-H., Noh, J. K., Lee, Y. C. & Eun, Y.-G. Epithelialmesenchymal transition gene signature is associated with prognosis and tumor microenvironment in head and neck squamous cell carcinoma. Sci. Rep. 10, 3652 (2020).
  23. Hu, Y. et al. Epigenetic suppression of E-cadherin expression by Snail2 during the metastasis of colorectal cancer. Clin. Epigenetics 10, 154 (2018).
  24. Tang, D. et al. TNF-alpha promotes invasion and metastasis via NFKappa B pathway in oral squamous cell carcinoma. Med. Sci. Monit. Basic Res. 23, 141-149 (2017).
  25. Wu, Y. et al. Stabilization of snail by NF-kappaB is required for inflammation-induced cell migration and invasion. Cancer Cell 15, 416-428 (2009).
  26. Adjuto-Saccone, M. et al. TNF-alpha induces endothelialmesenchymal transition promoting stromal development of pancreatic adenocarcinoma. Cell Death Dis. 12, 649 (2021).
  27. Coppenhagen-Glazer, S. et al. Fap2 of Fusobacterium nucleatum is a galactose-inhibitable adhesin involved in coaggregation, cell adhesion, and preterm birth. Infect. Immun. 83, 1104-1113 (2015).
  28. Yang, Y. et al. Fusobacterium nucleatum increases proliferation of colorectal cancer cells and tumor development in mice by activating toll-like receptor 4 signaling to nuclear factor-kappaB, and upregulating expression of microRNA-21. Gastroenterology 152, 851-866.e24 (2017).
  29. Long, X. et al. Peptostreptococcus anaerobius promotes colorectal carcinogenesis and modulates tumour immunity. Nat. Microbiol. 4, 2319-2330 (2019).
  30. Zhang, L., Liu, Y., Zheng, H. J. & Zhang, C. P. The oral microbiota may have influence on oral cancer. Front. Cell Infect. Microbiol. 9, 476 (2019).
  31. Huang, X. et al. Wnt7a activates canonical Wnt signaling, promotes bladder cancer cell invasion, and is suppressed by miR-370-3p. J. Biol. Chem. 293, 6693-6706 (2018).
  32. Xie, H. et al. WNT7A promotes EGF-induced migration of oral squamous cell carcinoma cells by activating -catenin/MMP9mediated signaling. Front. Pharmacol. 11, 98 (2020).
  33. Chen, L. et al. VEGF promotes migration and invasion by regulating EMT and MMPs in nasopharyngeal carcinoma. J. Cancer 11, 7291-7301 (2020).
  34. Wang, X. et al. Characterization of LIMA1 and its emerging roles and potential therapeutic prospects in cancers. Front. Oncol. 13, 1115943 (2023).
  35. Zeng, J., Jiang, W. G. & Sanders, A. J. Epithelial protein lost in neoplasm, EPLIN, the cellular and molecular prospects in cancers. Biomolecules 11 https://doi.org/10.3390/biom11071038 (2021).
  36. Feinberg, A. P., Ohlsson, R. & Henikoff, S. The epigenetic progenitor origin of human cancer. Nat. Rev. Genet. 7, 21-33 (2006).
  37. Yu, D. H. et al. Postnatal epigenetic regulation of intestinal stem cells requires DNA methylation and is guided by the microbiome. Genome Biol. 16, 211 (2015).
  38. Fellows, R. et al. Microbiota derived short chain fatty acids promote histone crotonylation in the colon through histone deacetylases. Nat. Commun. 9, 105 (2018).
  39. Ansari, I. et al. The microbiota programs DNA methylation to control intestinal homeostasis and inflammation. Nat. Microbiol. 5, 610-619 (2020).
  40. Qiao, F. et al. Downregulated PITX1 modulated by MiR-19a-3p promotes cell malignancy and predicts a poor prognosis of gastric cancer by affecting transcriptionally activated PDCD5. Cell Physiol. Biochem. 46, 2215-2231 (2018).
  41. Wang, Q., Zhao, S., Gan, L. & Zhuang, Z. Bioinformatics analysis of prognostic value of PITX1 gene in breast cancer. Biosci. Rep. 40 https://doi.org/10.1042/BSR20202537 (2020).
  42. Kolfschoten, I. G. M. et al. A genetic screen identifies PITX1 as a suppressor of RAS activity and tumorigenicity. Cell 121, 849-858 (2005).
  43. Vantaku, V. et al. Epigenetic loss of AOX1 expression via EZH2 leads to metabolic deregulations and promotes bladder cancer progression. Oncogene 39, 6265-6285 (2020).
  44. Liu, C. et al. GALNT6 promotes breast cancer metastasis by increasing mucin-type O-glycosylation of alpha2M. Aging 12, 11794-11811 (2020).
  45. Chen, S. et al. Fusobacterium nucleatum reduces METTL3-mediated m(6)A modification and contributes to colorectal cancer metastasis. Nat. Commun. 13, 1248 (2022).
  46. Xu, C. et al. Fusobacterium nucleatum promotes colorectal cancer metastasis through miR-1322/CCL20 axis and M2 polarization. Gut Microbes 13, 1980347 (2021).
  47. Wong, M. C. S. et al. Prevalence and epidemiologic profile of oral infection with alpha, beta, and gamma papillomaviruses in an Asian Chinese population. J. Infect. Dis. 218, 388-397 (2018).
  48. Chen, Z. et al. Impact of preservation method and 16S rRNA hypervariable region on gut microbiota profiling. mSystems 4 https://doi.org/10.1128/mSystems.00271-18 (2019).
  49. Matsen, F. A., Kodner, R. B. & Armbrust, E. V. pplacer: linear time maximum-likelihood and Bayesian phylogenetic placement of sequences onto a fixed reference tree. BMC Bioinforma. 11, 538 (2010).
  50. Zhu, H. et al. Convergent dysbiosis of upper aerodigestive microbiota between patients with esophageal and oral cavity squamous cell carcinoma. Int. J. Cancer 152, 1903-1915 (2023).
  51. Mirarab, S., Nguyen, N. & Warnow, T. SEPP: SATe-enabled phylogenetic placement. Pac. Symp. Biocomput. 247-258 https://doi.org/10.1142/9789814366496_0024 (2012).
  52. Hamady, M., Lozupone, C. & Knight, R. Fast UniFrac: facilitating highthroughput phylogenetic analyses of microbial communities including analysis of pyrosequencing and PhyloChip data. ISME J. 4, 17-27 (2010).
  53. Segata, N. et al. Metagenomic biomarker discovery and explanation. Genome Biol. 12, R60 (2011).
  54. Lin, H. & Peddada, S. D. Analysis of compositions of microbiomes with bias correction. Nat. Commun. 11, 3514 (2020).
  55. Fernandes, A. D. et al. Unifying the analysis of high-throughput sequencing datasets: characterizing RNA-seq, 16S rRNA gene sequencing and selective growth experiments by compositional data analysis. Microbiome 2, 15 (2014).
  56. Yang, L. & Chen, J. A comprehensive evaluation of microbial differential abundance analysis methods: current status and potential solutions. Microbiome 10, 130 (2022).
  57. Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139-140 (2010).
  58. Edgar, R. C. UPARSE: highly accurate OTU sequences from microbial amplicon reads. Nat. Methods 10, 996-998 (2013).
  59. Bernard, H. U. et al. Classification of papillomaviruses (PVs) based on 189 PV types and proposal of taxonomic amendments. Virology 401, 70-79 (2010).
  60. IARC Working Group on the Evaluation of Carcinogenic Risks to Humans Biological agents. Volume 100 B. A review of human carcinogens. IARC Monogr. Eval. Carcinog. Risks Hum. 100, 1-441 (2012).
  61. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15-21 (2013).
  62. Liao, Y., Smyth, G. K. & Shi, W. featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30, 923-930 (2014).
  63. Chen, J., Bardes, E. E., Aronow, B. J. & Jegga, A. G. ToppGene Suite for gene list enrichment analysis and candidate gene prioritization. Nucleic Acids Res. 37, W305-W311 (2009).
  64. Colaprico, A. et al. TCGAbiolinks: an R/Bioconductor package for integrative analysis of TCGA data. Nucleic Acids Res. 44, e71 (2016).
  65. Friedman, J. & Alm, E. J. Inferring correlation networks from genomic survey data. PLoS Comput. Biol. 8, e1002687 (2012).
  66. Saito, R. et al. A travel guide to Cytoscape plugins. Nat. Methods 9, 1069-1076 (2012).
  67. Supek, F., Bosnjak, M., Skunca, N. & Smuc, T. REVIGO summarizes and visualizes long lists of gene ontology terms. PLoS ONE 6, e21800 (2011).
  68. Subramanian, A. et al. Gene set enrichment analysis: a knowledgebased approach for interpreting genome-wide expression profiles. Proc. Natl Acad. Sci. USA 102, 15545-15550 (2005).
  69. Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114-2120 (2014).
  70. Krueger, F. & Andrews, S. R. Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications. Bioinformatics 27, 1571 (2011).
  71. Akalin, A. et al. methylKit: a comprehensive package for the analysis of genome-wide DNA methylation profiles. Genome Biol. 13, R87 (2012).
  72. Yu, G., Wang, L. G. & He, Q. Y. ChIPseeker: an R/Bioconductor package for ChIP peak annotation, comparison and visualization. Bioinformatics 31, 2382-2383 (2015).
  73. Manson McGuire, A. et al. Evolution of invasion in a diverse set of Fusobacterium species. mBio 5, e01864 (2014).

Acknowledgements

The authors thank the anonymous participants who provided samples for this study. We also thank Novogene (HK) Co., Ltd., China for the assistance with host gene long non-coding RNA sequencing (IncRNAseq) and human DNA CpG methylation sequencing, and the Core Utilities of Cancer Genomics and Pathobiology (CUCGP) at Department of Anatomical and Cellular Pathology of the Chinese University of Hong Kong for the service of 16 S rRNA gene sequencing. This work was partially supported by funding agents from the Stanley Ho Medical Foundation (J.Y.K.C.), the Research Grants Council of the Hong Kong Special Administrative Region, China (project numbers CUHK14108818 to J.Y.K.C. and CUHK14161017 to Z.C.), the Food and Health Bureau, Hong Kong SAR, China (reference no. 19202041 to Z.C.), and the Research Matching Grant (reference no. 8601687 to P.K.S.C.) from the Research Grants Council of the Hong Kong Special Administrative Region, China. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Author contributions

P.K.S.C., J.Y.K.C. and Z.C. conceived the study and designed the experiments. C.W.K.N., K.M., V.W.Y.L. and J.Y.K.C. collected clinical samples and metadata. H.Z., Q.M., P.Y.W., L.L., P.L. and S.F. developed the methods and performed the experiments. L.C., H.Z., K.M., A.S.L.C., J.Y., P.K.S.C., J.Y.K.C. and Z.C. analysed the data and discussed the results. L.C., H.Z., J.Y.K.C. and Z.C. wrote the original draft. All authors revised and approved the final version of the manuscript.

Competing interests

The authors declare no competing interests.

Additional information

Supplementary information The online version contains
supplementary material available at
https://doi.org/10.1038/s41522-024-00511-x.
Correspondence and requests for materials should be addressed to Jason Y. K. Chan or Zigui Chen.
Reprints and permissions information is available at http://www.nature.com/reprints
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.
© The Author(s) 2024

  1. Department of Microbiology, The Chinese University of Hong Kong, Hong Kong SAR, China. Department of Otorhinolaryngology, Head and Neck Surgery, The Chinese University of Hong Kong, Hong Kong SAR, China. Department of Chemical Pathology, The Chinese University of Hong Kong, Hong Kong SAR, China. Georgia Cancer Center, Augusta, GA 30912, USA. Department of Medicine, Medical College of Georgia, Augusta University, Augusta, GA 30912, USA. School of Biomedical Sciences, The Chinese University of Hong Kong, Hong Kong SAR, China. Department of Medicine and Therapeutics, The Chinese University of Hong Kong, Hong Kong SAR, China. These authors contributed equally: Liuyang Cai, Hengyan Zhu, Qianqian Mou.