Periodontal disease and cerebral atherosclerotic disease. Translational study

In the present study we found that periodontal disease (PD) was positively associated with lacunar infarct (LI) and, when present, emerged as one of the main contributors to an enhanced systemic inflamatory state promoting endothelial dysfunction with elevated levels of IL-6, PTX3, sTWEAK, and Aβ in LI patients. Moreover, moderate to severe active 1-40 PD was an independent predictor of poor functional outcome in LI patients. These findings were corroborated in a preclinical study, in which experimental PD induced with lipopolysaccharide from Porphyromonas gingivalis was associated with a mild systemic inflammatory response with disruption of the vascular endothelial function. TESE Periodontal disease and cerebral atherosclerotic disease. Translational study Tese de doutoramento Yago Leira Feijóo Periodontal disease and cerebral atherosclerotic disease. Translational study Yago Leira Feijóo 2018 PD en Ciencias Odontolóxicas ESCOLA DE DOUTORAMENTO INTERNACIONAL EN CIENCIAS DA SAÚDE DA USC TESE DE DOUTORAMENTO [PERIODONTAL DISEASE AND CEREBRAL ATHEROSCLEROTIC DISEASE. TRANSLATIONAL STUDY] [Yago Leira Feijóo] ESCOLA DE DOUTORAMENTO INTERNACIONAL PROGRAMA DE DOUTORAMENTO EN CIENCIAS ODONTOLÓXICAS SANTIAGO DE COMPOSTELA 2018 ESCOLA DE DOUTORAMENTO INTERNACIONAL EN CIENCIAS DA SAÚDE DA USC DECLARACIÓN DO AUTOR DA TESE [Periodontal disease and cerebral atherosclerotic disease. Translational study] D. Yago Leira Feijóo Presento miña tese, seguindo o procedemento adecuado ao Regulamento, e declaro que: 1) A tese abarca os resultados da elaboración do meu traballo. 2) No seu caso, na tese se fai referencia as colaboracións que tivo este traballo. 3) A tese é a versión definitiva presentada para a súa defensa e coincide ca versión enviada en formato electrónico. 4) Confirmo que a tese non incorre en ningún tipo de plaxio de outros autores nin de traballos presentados por min para a obtención de outros títulos. En Santiago de Compostela, 22 de maio de 2018 Asdo. AUTORIZACIÓN DO DIRECTOR / TITOR DA TESE [Periodontal disease and cerebral atherosclerotic disease. Translational study] D. Juan Blanco Carrión D . José Castillo Sánchez INFORMAN: Que a presente tese, correspóndese co traballo realizado por D. Yago Leira Feijóo, baixo a nosa dirección, e a utorizamos a súa presentación , considerando que reúne os r equisitos esixidos no R egulamento de Estudos de Doutoramento da USC, e que como directores desta non incorre nas causas de abstención establecidas na Lei 40/2015. En Santiago de Compostela, 22 de maio de 2018 Asdo. Asdo. El Prof. Dr. JUAN BLANCO CARRIÓN, Profesor Titular de Periodoncia de la Universidad de Santiago de Compostela y el Prof. Dr. JOSÉ CASTILLO SÁNCHEZ, Director Científico del Instituto de Investigación Sanitaria de Santiago de Compostela (IDIS), CERTIFICAN: Que el presente trabajo titulado ”ENFERMEDAD PERIODONTAL Y ENFERMEDAD CEREBRAL ATEROSCLERÓTICA. ESTUDIO TRASLACIONAL” ha sido realizado bajo su dirección por el Licenciado en Odontología YAGO LEIRA FEIJÓO, reúne todos los requisitos para optar al Doctorado Internacional. Los directores: Prof. Dr. Juan Blanco Carrión Prof. Dr. José Castillo Sánchez En Santiago de Compostela, 22 de mayo de 2018 lo he dicho más de una vez, y es que te considero como un “padre” en el mundo de la odontología en general, y de la periodoncia en particular. Gracias por tu apoyo y confianza desde el primer momento en el master de periodoncia. Eres mi referente en periodoncia y, aparte de haberme enseñado un oficio, has hecho otra cosa más importante: me has educado como persona. Tus dos palabras favoritas en el master “Vamos” y “Humildad” siempre las llevaré conmigo. Además, eres uno de los culpables de que haya realizado la Tesis, pues era un requisito obligatorio y condición sine qua non para hacer el master. En uno de mis primeros días en el master me dijiste que la periodoncia me iba a encantar. Y no te equivocabas. De hecho la amo y, en broma con mis compañeros de master, solíamos decir que nos habíamos casado con ella. Fueron años muy duros los del master y más compaginándolos con la Tesis, pero si volviese atrás, no dudaría en volver a hacer exactamente lo mismo, y sin lugar a duda, te elegiría a ti como jefe y guía. Hace unos meses estuviste hablando de investigación en un congreso junto a algunos de los investigadores referencia a nivel mundial en las diferentes áreas de la periodoncia y decías que más que pertenecer al “Dream Team” de estos investigadores, formabas parte del “Drink Team”. Puedo decir con seguridad que de largo te has ganado un puesto en el primero. Gracias por llevar la periodoncia santiaguesa y gallega al máximo nivel internacional. Eres un pionero y admiro tus ganas de seguir mejorando y formándote, porque de ti aprendí que estamos en una profesión en la que vamos a tener que estar formándonos toda la vida. Cuando voy a los congresos de periodoncia/implantes y me preguntan a qué master de periodoncia pertenezco y mi respuesta es al de Santiago, lo primero que suelen decir es: “Ah, el de Juan Blanco”. Eso te lo has ganado a pulso y créeme cuando digo que estoy muy orgulloso de haber formado parte de esa gran familia. Espero que después de estos años que hemos compartido juntos estés también orgulloso de mí. ¡Gracias de todo corazón, Jefe! Quiero agradecer al Prof. Castillo (Pepe), director de esta Tesis, por guiarme a lo largo de todo este proyecto con mano dura y crítica, pero siempre con cariño. ¡Otro de mis maestros! Cuando aceptaste ser uno de mis directores de Tesis, para mí fue todo un reto, pues rendirle cuentas a una de las personas más importantes e influyentes en la patología neurovascular, tanto a nivel nacional como internacional, no iba a ser fácil. Recuerdo que algunos me llamaban loco, diciéndome que si no tenía bastante con la periodoncia que también me iba a meter en el mundo de la neurología. Y sí, fue un reto. ¡Pero fue bonito y entretenido! Creedme cuando os digo que fue un honor ser doctorando del Prof. Castillo, pues donde realmente me di cuenta de lo que había hecho por la neurología española fue cuando empecé a revisar todas sus investigaciones. Posiblemente seas unos de los culpables de mi “gran amor” por los biomarcadores. Me conoces desde que era un niño y sabes que soy una persona impaciente y muy impulsiva, y has sabido controlar esos impulsos con tu experiencia. Decías que los datos me quemaban en las manos y que quería publicarlos demasiado rápido. Pero debes entender que, si quiero llegar a la cuarta parte de lo que has sido tú… ¡tengo que empezar a publicar cuanto antes! Además, muchas gracias por abrirme las puertas del Laboratorio de Investigación en Neurociencias Clínicas (LINC) y, consecuentemente, al mundo de la experimentación animal. Fuiste tú el que me dijo que debería hacer un estudio traslacional para mi tesis y, gracias a ello, he aprendido sobre la experimentación animal en periodoncia. En resumen Pepe, de todo corazón, muchas gracias por todo. Para mí ha sido un placer poder cerrar un círculo vital con “los Leira” que empezó con mi padre en el año 1988, cuando dirigiste tu primera tesis que resultó ser la suya, y ahora 30 años después, se cierra con la mía. Otro de mis maestros en el mundo de la investigación en odontología es el Prof. Juan Seoane. Persona que me dio la oportunidad de realizar mi primera investigación, la cual se plasmó en un artículo. Desde entonces, comenzamos una relación que fue más allá del terreno profesional. Gracias por ser la persona que eres. Al igual que el Prof. Blanco, me demostraste que se puede ser un referente y ser humilde, buena persona, generoso y atento. Recuerdo la primera vez que me reuní contigo para un tema de investigación y me dijiste que ni se me ocurriera tratarte de usted. En ese momento quedó reflejado la clase de persona que eres. Gracias por confiar en mí y tenerme siempre en cuenta considerando mis comentarios y sugerencias, ya que tengo que confesarte que al principio me sorprendía que considerases mis propuestas en ciertos temas (dado mi poca experiencia en el mundo de la investigación). Tengo que agradecerte de manera especial tus valiosas lecciones sobre la realización de una revisión sistemática, las cuales se vieron plasmadas en uno de los artículos publicados en relación con la presente Tesis. Gracias también por tus consejos y las múltiples conversaciones sobre investigación. Resumiendo Juan, muchas gracias por abrirme los ojos al maravilloso mundo de la investigación y publicación. Especial agradecimiento a la Unidad de Ictus del Servicio de Neurología del Hospital Clínico Universitario de Santiago de Compostela. En particular a la Dra. Susana Arias y al Dr. Manuel Rodríguez-Yáñez (Manu), responsables del examen clínico, neurosonológico y de neuroimagen de los pacientes con infarto lacunar. A este último, además, quiero agradecerle su disponibilidad para cualquier cosa. Eres el hombre de las 3Es (efectivo, eficaz y eficiente). ¡Manu, muchas gracias por estar ahí cuando te necesitaba! Gracias a la gente del LINC. Especial mención para Dr. Tomás Sobrino (director del LINC y “conseguidor” de kits de ELISA). Sé que no es nada fácil dirigir un laboratorio tan grande y has sufrido mi carácter impulsivo e impaciente para realizar las determinaciones de las muestras de mis pacientes pero, al final, gracias a tu gestión he cumplido los tiempos. ¡Muchas gracias, Tomás! Al Dr. Francisco Campos gracias por ayudarme a poner a punto el modelo animal de periodontitis, lo cual no fue nada fácil. ¡Fran, al final lo hemos conseguido! A Alba gracias por ser una excelente profesora de la técnica de ELISA así como consejera en el manejo de los animales. Andrés, hombre de confianza para la preparación del lipopolisacárido y otras tareas relacionadas con la experimentación animal, ¡Muchas gracias por todo! A Ramón, responsable de la resonancia al que saqué de su terreno pidiéndole un análisis de imagen de la boca pero que realizó de manera brillante. María, gracias por estar siempre disponible para responderme a cualquier duda relacionada con el manejo de los animales. Esteban, guía en el procesamiento de las muestras. Joserra, solucionador de problemas administrativos siempre con una sonrisa y con tranquilidad. A todos, muchas gracias por ayudarme durante mis años en el LINC. También especial agradecimiento merecen los investigadores de la Unidad de Imagen Molecular Preclínica, Noemí y el Dr. Pablo Aguiar. Porque también los he sacado de su terreno para realizar el examen de micro-TC de los animales y que tan brillantemente han llevado a cabo. ¡Muchas gracias a los dos! Agradecer también al Prof. Bahi Takkouche por la ayuda prestada en la realización del metaanálisis de uno de los artículos fruto de esta Tesis. Durante mi estancia en el Eastman Dental Institute en Londres, he vivido tres meses maravillosos aprendiendo de medicina periodontal. He conocido a gente increíble que me ha tratado como uno más del grupo. A todos ellos les dedico unas palabras: “I would like to acknowledge another mentor, Prof. Francesco D’Aiuto. Thank you for being so kind, helpful and friendly during my stay at Eastman Dental Institute (London). Being with you learning periodontal medicine was a dream come true. You have created an incredible atmosphere in Perio Unit with all the staff and students. This is the main reason why I want to be there in the future. My gratitude goes also to great people that I´ve met during my stay such as Fede, Basit, and Marco as well as EDI staff, MCD1, MCD2, and MCD3 students. I remember with affection our funny evenings in the golf range. Many thanks for your friendship! Again, thank you so much for three amazing months in London! And I hope to see you soon guys…” Quiero acordarme también de mis compañeros de master (Fani, Patri y Pablo). Con ellos he convivido tres intensos años en la Facultad de Odontología. Patri, somos el yin y el yang, por eso logramos compenetrarnos tan bien. Las carencias de uno se suplían con las virtudes del otro ¡Gracias por aguantarme como compañera de box todos estos años! A mi gran amigo y confidente, Pablo. Gracias por tu bondad y amistad. Hemos vivido momentos buenos y malos, pero siempre juntos. Sólo espero que acabes la tesis y, por fin, de una vez por todas, te pueda llamar Dr. Ameijeira con mayúsculas. Espero que lo que ha unido el master, no lo rompa nada. ¡Muchas gracias, coco! A la Dra. Elena Figuero, de la Universidad Complutense de Madrid. Pues ella fue la principal responsable de que hoy día sepa realizar el análisis estadístico de mis investigaciones. Gracias por explicar fácil lo que es realmente difícil de explicar. Mención aparte, aunque también miembro de la familia master de periodoncia de la Complutense es Javi Aracil, al que no considero un amigo sino que para mí es un hermano. Lo nuestro fue un flechazo en la carrera. Mucha gente dice que menos mal que nos conocimos el último año porque sino… ¡Y es que somos muy parecidos en la forma de ser! Ya sabes que el único fallo que tienes es… bueno… lo de ser del Real Madrid. No me sorprende tu inmensa bondad conmigo, porque al abrirme las puertas de tu casa en Madrid conocí a toda tu familia y, entonces lo entendí todo. Gracias a ti y a tus padres por acogerme en más de una ocasión en tu casa y hacerme sentir como en la mía. De todo corazón, ¡Gracias por tu amistad! No quisiera olvidarme de mi grupo de amigos de Santiago que nos conocemos desde hace mucho, mucho, mucho tiempo, como son: J. Veiga, Vichu, Yani, Wilson, Lewis, los dos Alex, Arce, Inesita, Elenita, Antía (Poti), Cristi, Carmencita… entre otr@s. ¡Nos vemos en los bares, chic@s! Mi grupo de “Yatris”, amigos de la carrera de Odontología: Chacho, Paiva, Yaki, Pitu, Pedrolo, Haider, Pitu y Corvi. También incluyo a María Saaaal, la cual seguro que no le importa formar parte de este gran grupo de “expertos”. Madrid siempre será nuestro punto de encuentro. ¡Gracias por estar ahí siempre que os necesito! A los Dres. José Miguel Láinez, Julio Pascual y José Vivancos, los cuales considero como parte de mi familia. Gracias por cuidarme todos estos años en las reuniones de la SEN. Además, agradecer a vuestras respectivas familias, en especial a tía Elena, siempre preocupándose por mí. De todo corazón, ¡gracias por todo! A mi hermano Pablo (PoL). Aunque estés en Madrid, sé que siempre estarás a mi lado para lo que lo necesite. Soy una persona con suerte, pues tenerte como hermano no tiene precio. Las risas cuando nos juntamos están garantizadas. Me gustaría cerrar esta sección con la otra persona más importante en mi vida junto con mi padre: Nati (mamá). Podría escribir hojas y hojas agradeciéndote todo lo que has hecho por mí. Gracias por estar siempre de mi lado, aún cuando yo no tenga razón. Sé que a veces tu amor de madre puede cegarte pero para mi es importante saber que pase lo que pase y haga lo que haga, siempre vas a estar ahí apoyándome. Has sufrido incluso más que yo mis fracasos y problemas, pero sé que también has disfrutado la que más con mis “victorias”. Al igual que con mi hermano, en esto también soy un suertudo, pues tenerte como madre es el mejor regalo que he podido tener. ¡Muchas gracias por todo, mamá! “Sólo la posibilidad teórica de que la enfermedad periodontal pueda predisponer a la aparición de un accidente cerebrovascular tiene una trascendencia alarmante para los millones de pacientes que la sufren”. “Las cosas buenas suceden a los que se dan prisa”. Anais Nin “A veces la vida te va a pegar en la cabeza con un ladrillo. Pero no pierdas la fe”. Steve Jobs “Confía en el tiempo, que suele dar dulces salidas a muchas amargas dificultades”. Miguel de Cervantes Saavedra “Cada fracaso enseña al hombre algo que necesitaba aprender”. Charles Dickens Periodontal disease and cerebral atherosclerotic disease. Translational study RESUMO Aínda que estableceuse unha asociación entre a enfermidade periodontal (PD) e o ictus isquémico de gran vaso, hai pouca evidencia sobre a relación entre a PD e infarto lacunar (LI), un subtipo de enfermidade cerebral de pequeno vaso que é responsable de aproximadamente o 25% dos casos de ictus isquémico. Polo tanto, o obxectivo deste estudo foi investigar se a PD está asociada co LI e de ser así, estudar os posibles mecanismos que podrían explicar esta asociación. No presente estudo, a PD asociouse positivamente co LI e cando está presente emerxeu como un dos principais factores que contribuiu a un estado sistémico pro-inflamatorio promovendo a disfunción endotelial con niveis elevados de IL-6, PTX3, sTWEAK, e Aβ1-40 nos pacientes con LI. Ademáis, a PD moderada/avanzada activa foi un preditor independente de mal prognóstico nos pacientes con LI. Estes resultados foron corroborados nun estudo preclínico in vivo no modelo roedor no que a PD experimental inducida por lipopolisacárido de Porphyromonas gingivalis asociouse cunha resposta inflamatoria sistémica leve coa disrupción da función endotelial vascular. PALABRAS CHAVE: enfermidade periodontal, infarto lacunar, inflamación, disfunción endotelial, enfermidade cerebral de pequeno vaso. RESUMEN Aunque se ha establecido una asociación entre enfermedad periodontal (PD) e ictus isquémico de gran vaso, existe poca evidencia sobre la relación entre PD e infarto lacunar (LI), un subtipo de enfermedad cerebral de pequeño vaso que es responsable de aproximadamente el 25% de los casos de ictus isquémico. Por lo tanto, el objetivo de nuestro trabajo fue investigar si la PD se asocia con LI y en caso de estarlo, estudiar posibles mecanismos que puedan explicar esta asociación. En el presente estudio la PD se asoció positivamente con el LI y, cuando está presente 4.8. Statistical analysis .............................................................. 106 5. Results ....................................................................................... 107 5.1. Study population ................................................................ 107 5.2. Study groups – baseline characteristics ............................. 108 5.3. Study groups – periodontal disease .................................... 109 5.4. Association between periodontal disease and its clinical parameters and the presence of lacunar infarct ......................... 111 5.5. Molecular analysis ............................................................. 113 5.5.1. IL-6 levels in sera according to periodontal status ..... 113 5.5.2. IL-10 levels in sera according to periodontal status ... 114 5.5.3. PTX3 levels in sera according to periodontal status ... 114 5.5.4. sTWEAK levels in sera according to periodontal status ............................................................................................... 115 5.5.5. Aβ1-40 levels in sera according to periodontal status ... 116 5.5.6. Aβ1-42 levels in sera according to periodontal status ... 116 5.6. Association between periodontal disease and periodontal inflamed surface area and elevated serum levels of biomarkers in lacunar infarct patients .............................................................. 117 5.6.1. IL-6.............................................................................. 117 5.6.2. PTX3 ........................................................................... 118 5.6.3. sTWEAK ..................................................................... 119 5.6.4. Aβ1-40 ........................................................................... 120 5.7. Predictors of poor prognosis in patients with lacunar infarct . 121 5.7.1. Study groups – baseline characteristics ...................... 121 5.7.2. Study groups – periodontal disease ............................. 122 5.7.3. Molecular analysis ...................................................... 124 5.7.4. Association between periodontal clinical parameters and poor outcome in patients with lacunar infarct ....................... 124 5.7.5. Correlation between periodontal inflamed surface area and significant biomarkers in patients with poor outcome ... 126 EXPERIMENTAL STUDY: complementary investigation ............ 129 1. Justification ............................................................................... 131 2. Hypothesis ................................................................................ 131 3. Objective ................................................................................... 131 4. Material and Methods ............................................................... 132 4.1. Experimental periodontitis model ..................................... 132 4.1.1. Porphyromonas gingivalis lipopolysaccharide-induced periodontitis .......................................................................... 132 4.1.2. Experimental design ................................................... 134 4.1.3. Animals and Anaesthesia ........................................... 135 4.1.4. Porphyromonas gingivalis lipopolysaccharide preparation .............................................................................................. 136 4.1.5. Periodontal induction procedure ................................. 136 4.2. MRI analysis ...................................................................... 137 4.2.1. MRI examination ........................................................ 138 4.2.2. MRI assessment .......................................................... 139 4.2.3. Data processing .......................................................... 139 4.3. μCT analysis ...................................................................... 141 4.3.1. μCT examination ........................................................ 141 4.3.2. Alveolar bone loss measurement ................................ 141 4.4. Serum collection and laboratory tests ................................ 142 4.5. Euthanasia .......................................................................... 143 4.6. Statistical analysis ............................................................. 144 5. Results ...................................................................................... 144 5.1. Periodontal inflammation .................................................. 144 5.2. Alveolar bone loss ............................................................. 147 5.3. Biomarkers ......................................................................... 149 5.3.1. Systemic inflammation ............................................... 149 5.3.2. Endothelial dysfunction .............................................. 150 5.3.3. Aβ peptides ................................................................. 151 DISCUSSION ................................................................................... 153 CONCLUSIONS ............................................................................... 169 IMPLICATIONS FOR FUTURE RESEARCH ............................... 173 REFERENCES.................................................................................. 177 RESUMEN ....................................................................................... 229 CONFLICT OF INTEREST STATEMENT .................................... 241 ABBREVIATIONS Abbreviations 35 ABBREVIATIONS Aa: Aggregatibacter actinomycetemcomitans AAP: American Academy of Periodontology Aβ: amyloid-beta Aβ1-40: amyloid-beta 1-40 Aβ1-42: amyloid-beta 1-42 AD: Alzheimer´s disease AIC: Akaike information criterion ANOVA: one-way analysis of variance APP: amyloid precursor protein ApoE: apolipoprotein E APRs: acute-phase reactants ARIC: Atherosclerosis Risk in Communities ARRIVE: Animal Research Reporting of In Vivo Experiments BBB: blood-brain barrier BMI: body mass index BoP: bleeding on probing CAL: clinical attachment level CDC: Center for Disease Control and Prevention CEJ: cemento-enamel junction CI: confidence interval CKD: chronic kidney disease CMBs: cerebral microbleeds CRP: C-reactive protein CSVD: cerebral small vessel disease CT: computed tomography CV: coefficient of variation CVD: cardiovascular disease DALYs: disability-adjusted life years DAMPs: damage associated molecular peptides DMFT: decayed, missing and filled teeth index YAGO LEIRA FEIJÓO 36 ECs: endothelial cells EDV: endothelium-dependent vasodilation ELISA: enzyme-linked immunosorbent assay eNOS: endothelial nitric oxide synthase ET-1: endothelin-1 EcT: echo time FA: flip angle FimA: fimbriae A FLAIR: fluid-attenuated inversion recovery FMBS: full-mouth bleeding score FMD: flow-mediated dilation FMPS: full-mouth plaque score Fn: Fusobacterium nucleatum Fn14: fibroblast growth factor-inducible 14 GCF: gingival crevicular fluid GI: gingival index GR: gingival recession HagA: hemagglutinin A HDLs: high-density lipoproteins HPA: hypothalamic-pituitary-adrenal axis HR: hazard ratio HSPs: heat-shock proteins HSP60: heat-shock protein-60 IL-1A: Interleukin-1A IL-1B: Interleukin-1B IL-1β: Interleukin-1β IL-6: Interleukin-6 IL-8: Interleukin-8 IL-10: Interleukin-10 IL-6R: cellular interleukin-6 receptor ICAM-1: intercellular adhesion molecule-1 ICD: International Classification of Diseases Ig-A: immunoglobulin-A IKK: Iκβ kinase IMT: intima-media thickness LBP: LPS binding protein Abbreviations 37 LDLs: low-density lipoproteins LI: lacunar infarct LPS: lipopolysaccharide μCT: micro-CT MCAO: middle cerebral artery occlusion MCP-1: monocyte chemotactic protein-1 mm2: square millimetres MMPs: matrix metalloproteinases MMP-9: matrix metalloproteinase-9 MMP-2: matrix metalloproteinase-2 MRI: magnetic resonance imaging mRS: modified Rankin scale NA: number of averages NF-κB: nuclear factor-κB NHANES: National Health and Nutrition Examination Survey NIH: National Health Insurance NO: nitric oxide NO2: nitrous oxide NSAIDs: non-steroidal anti-inflammatory drugs O2: oxygen OPG: osteoprotegerin OR: odds ratio PAI-1: plasminogen-activator inhibitor-1 PD: periodontal disease/periodontitis PESA: periodontal epithelial surface area Pg: Porphyromonas gingivalis PGE2: prostaglandin E2 Pi: Prevotella intermedia PISA: periodontal inflamed surface area PMNs: polymorphonuclear neutrophils PPCs: periodontal profile classes PPD: probing pocket depth PTXs: pentraxins PTX3: pentraxin 3 PVS: perivascular spaces Qs: quartiles YAGO LEIRA FEIJÓO 38 r: Pearson´s correlation coefficient RANKL: receptor activator of nuclear factor-κB ligand ROC: receiver operating characteristic ROIs: regions of interest ROS: reactive oxygen species RR: relative risk RT: repetition time SAA: serum amyloid protein A sCD14: soluble CD14 SgI: signal intensity SI: silent infarct s-ICAM: soluble intercellular adhesion molecule sIL-6R: soluble interleukin-6 receptor SMC: smooth muscle cell STROBE: Strengthening the Reporting of Observational Studies in Epidemiology sTWEAK: soluble tumor necrosis factor-like weak inducer of apoptosis SW: spectral bandwidth T1-w: T1-weighted T2-w: T2-weighted Td: Treponema denticola Tf: Tannerella forsythia TF: tissue factor TGs: triglycerides TIA: transient ischemic attack TLRs: toll-like receptors TLR-2: toll-like receptor 2 TLR-4: toll-like receptor 4 TNF-α: tumor necrosis factor-α TOAST: Trial of Org 10172 in Acute Stroke Treatment tPA: tissue plasminogen-activator TWEAK: tumor necrosis factor-like weak inducer of apoptosis US: United States VCAM-1: vascular cell adhesion molecule-1 WHO: World Health Organization WMD: weighted mean difference INTRODUCTION YAGO LEIRA FEIJÓO 46 symbiotic relationship should exist between the microbiota and the host response. Accordingly, the host can provide key nutrients through the gingival crevicular fluid (GCF), and the proteins and peptides that are released by microorganisms elicit a proportionate and resolving host response (Marsh 2003; Van Dyke 2008). In case of biofilm accumulation, some periodontopathogens such as Fusobacterium nucleatum (Fn) that are capable of sensing and influencing their own environment via chemical signals start to emerge and evoke a stronger host response leading to the onset of gingival inflammation and increase the supply of certain nutrients (e.g. heme), which are key to the proliferation of periodontopathogens such as Porphyromonas gingivalis (Pg) (Kolenbrander et al. 2002). This process is called “incipient dysbiosis” and in non-susceptible subjects it does not progress beyond gingivitis. On contrary, in susceptible individuals, an inappropriate and excessive response occurs in the host, in which a great number of cytokines, reactive oxygen species (ROS) and MMPs are produced and along with their antagonist produce periodontal tissue breakdown. The so-called damage associated molecular peptides (DAMPs) are released and the inflammatory response is enhanced. Due to poor innate inflammation resolving response, the periodontal inflammatory lesion becomes chronic. In addition, some viruses are capable of creating dysregulation in the immune system. When the chronic inflammatory state is established, a rich nutritional environment is created for sustaining dysbiosis due to healing process and inflammation leading to a pathogenic biofilm. The periodontal lesion is dominated by plasma cells and polymorphonuclear neutrophils (PMNs). The latter is responsible for dysregulating chemotactic and microbial processes and, as a result, failing to release pro-resolving lipid mediators. At this phase, because the natural process of inflammatory resolution is inactive, disruption of the biofilm is warranted to an extent in which health-promoting microorganisms can re-establish themselves and try to reduce the inflammatory process. As a result of the interrelation between the health-promoting biofilm and the host response, a balanced well-regulated inflammatory immune repertoire is restored. Nevertheless, a great variety exists with regards to the degree of biofilm reduction necessary to achieve in order to Introduction 47 establish symbiosis. On the one side, there are patients at a low risk of developing PD (disease-resistant patients). On the other side, high-risk patients are those who develop immediately PD, even if only mild plaque accumulation is present but is enough to trigger a destructive host response and, subsequently, periodontal tissue damage (Meyle and Chapple 2015). Figure 2. Current model of PD pathogenesis. Adapted from Meyle and Chapple (2015) and reproduced with permission from Wiley. 1.4. Risk factors 1.4.1. Aging Epidemiologic studies showed more PD in older age groups as compared to younger groups (Grossi et al. 1994, 1995). Also, evidence demonstrated that plaque accumulation is more frequent and, as a result, more severe gingivitis could be observed in elderly people as compared to the younger ones, which suggest an age-related effect (Abdellatif and Burt 1987). However, PD could be more severe in the elderly due to Healthy gingiva Symbiosis Host response Biomass Acute resolution of inflammation PMNs Complement Antigens Bacterial DNA Gingivitis Incipient dysbiosis Host response Biomass Chronic resolution of inflammation PMNs T & B cells Antigens LPS Antibodies Periodontitis Established dysbiosis Biomass Chronic nonresolving inflammation PMNs Plasma cells Antigens LPSAntibodies Host response Connective tissue breakdown & Bone resorption DAMPs GCF Failed inflammatory resolution Cytokines MMPs ROS Prostanoids •Risk factors (i.e., behavioural, environmental and genetic) •Epigenetic effects •Risk factors (i.e., behavioural, environmental and genetic) •Epigenetic effects Gingipains YAGO LEIRA FEIJÓO 48 cumulative periodontal tissue destruction over a lifetime rather than an age-related issue affecting host susceptibility. 1.4.2. Gender Male sex has been considered as one of the main risk factors for PD. Although being a man or a woman is genetically determined, no inherent difference between the two genders may exist in terms of susceptibility to PD (Genco and Borgnakke 2013). Nevertheless, gender-related lifestyle could be the responsible for the higher prevalence of PD in males in comparison to females in all groups of age, race/ethnic, and geographic locations (Grossi et al. 1994; Haas et al. 2012; Eke et al. 2012). Results from a nationally representative sample of non-institutionalized civilians in the United States (US) showed that men ha about 50% higher prevalence of PD (Eke et al. 2012). In fact, men had 180% more severe PD than women, with over half the men being affected (56.4%). In Spain, epidemiologic data demonstrated that males had more periodontal pockets compared with females (43.2% versus 31.6%) (Carasol et al. 2016). 1.4.3. Smoking Since cigarette smoking was associated with acute necrotizing ulcerative gingivitis in the 50´s (Pindborg 1947), tobacco consumption has been studied as a risk factor for PD. Historically, the association between smoking and the presence of PD was controversial due to the fact that smokers had higher levels of plaque than non-smokers and, therefore, oral hygiene alone would account for the differences in the periodontal status between smokers and non-smokers. Nevertheless, it has been demonstrated that this statement was wrong because smoking was strongly associated with PD independent of plaque accumulation (Grossi et al. 1994, 1995). Currently, smoking is considered an important risk factor for PD. A meta-analysis of 2361 subjects showed that smoking was associated with the presence of PD, with an overall estimated odds ratio (OR) of 2.82 [95% confidence interval (CI): 2.363.39] (Papapanou 1996). The number of pack years of smoking was Introduction 49 positively correlated with the amount of attachment and bone loss, showing a clear dose-response (Grossi et al. 1994, 1995). Following non-surgical periodontal therapy, patients who smoked experienced less reduction in PPD than non-smokers (Labriola et al. 2005) and a causal relationship was found between poor periodontal wound healing and smoking habit (Heasman et al. 2006). If smoking cessation is implemented, periodontal patients will experience less progression of alveolar bone loss (Bolin et al. 1993) as well as a significant reduction in PPDs (Preshaw et al. 2005). There are three hypotheses why cigarette smoking is detrimental to the periodontium. The first one is that smoking select for specific periodontopathogens [i.e., Pg, Treponema denticola (Td) and Tannerella forsythia (Tf)] leading to increased risk for development and progression of PD (Zambon et al. 1996). The second hypothesis suggests that smoking may result in peripheral vasoconstriction, probably associated with low doses of nicotine (Bergström et al. 2001; Morozumi et al. 2004). This fact leads to reduced gingival bleeding in smokers (Bergström et al. 2001) and less oxygen tension within the periodontal pocket and, thus, favours the overgrowth of Pg and Td (Genco and Borgnakke 2013). The third hypothesis is the impairment of neutrophil function by smoking via nicotine. This substance makes neutrophils more sensitive to bacterial challenge (Soder et al. 1999). Furthermore, nicotine inhibits the proliferation, chemotaxis, and attachment of fibroblasts from the periodontium leading to worse periodontal healing and regeneration (Cuff et al. 1989; James et al. 1999). 1.4.4. Diabetes A two-year follow-up radiographic study among Pima Indians diagnosed with type 2 diabetes showed that poor glycaemic control could lead to both an increased risk for alveolar bone loss and more severe progression of PD than those without type 2 diabetes (Taylor et al. 1998). In addition, patients with a better control of the disease were less prone for bone-loss progression compared with patients who had YAGO LEIRA FEIJÓO 50 worse metabolic control (Taylor et al. 1998; Bandyopadhyay et al. 2010). What is more, it seems that diabetes precedes PD as it was observed an adjusted relative risk (RR) of incident PD of 2.6 (95% CI: 1.0-6.6) in patients with type 2 diabetes in comparison with nondiabetics (Nelson et al. 1990). Pre-diabetes has also been associated with PD. In a case-control study, matched by age and sex, people with pre-diabetes (measured as impaired blood glucose levels) had moderate PD (Löesche et al. 2000). Cross-sectional data showed that the prevalence of subjects with alveolar bone loss ≥ 6 mm was significantly higher in those with impaired fasting glucose than among individuals without it (Zadik et al. 2010). Similarly, women with a history of gestational diabetes may be at a higher risk of having PD. In fact, results from the National Health and Nutrition Examination Survey (NHANES) carried out in the US population demonstrated that the prevalence of PD was significantly higher in women with a history of gestational diabetes compared to those without this condition (30.5% versus 4.8%) (Novak et al. 2006). The biological mechanisms through which diabetes is associated with PD are mainly based on the inflammatory process. Inflammation is a central feature of both diseases. In periodontal tissues of diabetics, the inflammatory processes are up-regulated. Interleukin-1β (IL-1β) and prostaglandin E2 (PGE2) levels measures in GCF are higher in patients with diabetes than in non-diabetics with the same periodontal condition (Salvi et al. 1997a). Monocytes from patients with diabetes enhance the production of tumor necrosis factor-α (TNF-α), IL-1β and PGE2 compared with those free of diabetes (Salvi et al. 1997a, 1997b; Engebretson et al. 2004). Additionally, elevated systemic inflammatory markers were found in diabetics (Dandona et al. 2004). There is another hypothesis suggesting that diabetes is responsible for a hyperreactive inflammatory response to the bacterial challenge and, thus, could enhance severity of PD. Accordingly, gingival mice tissues with experimental diabetes showed increased vascular permeability and impaired neutrophil chemotaxis, both of which can lead to worse Introduction 51 periodontal status given a similar bacterial challenge (Gyurko et al. 2006; Sima et al. 2010). 1.4.5. Obesity and metabolic syndrome Several studies over the last decade showed a relationship between overweight/obesity and PD. A meta-analysis including mostly crosssectional and case-control studies, demonstrated that the OR of having PD if an individual was either obese or overweight were 1.81 (95% CI: 1.42-2.30) and 1.27 (95% CI: 1.06-1.51), respectively, compared to those subjects with a normal body mass index (BMI) (Suvan et al. 2011). It has been hypothesized that overgrowth of periodontopathogens such as Tf could occur in individuals with a healthy periodontium who are obese, putting them at a higher risk for the onset and progression of PD (Haffajee and Sokransky 2009). It is well known that an elevated inflammatory response may be seen in obese individuals due to the production of numerous inflammation-related factors by adipose tissue. Data from 1221 adults, those in the highest quartile of BMI had the highest levels of TNF-α and soluble TNF-α receptors. Furthermore, a positive correlation was found between TNFα levels with PD only in subjects included in the lowest quartile of BMI (Genco et al. 2005). Adipocytokines such as adiponectin or leptin measured in serum are elevated in patients with PD (Saito et al. 2008; Karthikeyan and Pradeep 2007). The latter, was suggested to be a proinflammatory mediator in the association between PD and acute myocardial infarct (Gundala et al. 2014) or chronic migraine (Leira et al. 2017a). Experimental data demonstrated that systemic low-grade inflammation along with elevated gene expression for TNF-α and Creactive protein (CRP) in obese animals might lead obese rats to be more prone to develop PD (Endo et al. 2010). Additionally, it has been shown that after uncontrolled Pg infection in obese animals could predispose to increased alveolar bone loss via impaired immune response (Amar et al. 2007). Regarding metabolic syndrome, a secondary analysis of the NHANES III showed that females that had 3 or more metabolic YAGO LEIRA FEIJÓO 52 syndrome components were at a 2-fold increased risk for developing PD (OR=2.1, 95% CI: 1.2-3.7). Moreover, it seems that abdominal obesity appeared to be the most significant contributor in this association (Andriankaja et al. 2010). This could be explained because there is an enhanced chronic systemic inflammatory response in subjects who have some metabolic syndrome components leading to impaired immunopathologic response to periodontopathogens and, as a result, leading to greater periodontal tissue destruction (Genco and Borgnakke 2013). 1.4.6. Osteoporosis, dietary calcium and vitamin D A systematic review that included 35 studies showed an association between systemic osteoporosis and the presence of PD based on radiological parameters (e.g. alveolar bone loss or alveolar crest height). Nevertheless, when PD was assessed by clinical parameters (e.g., periodontal attachment loss), results were controversial (Martínez-Maestre et al. 2010). Results from the NHANES III revealed that subjects, especially women, with a low dietary calcium intake (< half of the recommended dietary allowance) had more severe PD (Nishida et al. 2000). In addition, a clinical trial demonstrated that individuals receiving periodontal maintenance therapy who took calcium and vitamin D supplements tend to have a better periodontal condition than those non-supplement takers (Miley et al. 2009). 1.4.7. Stress A positive relationship was found between psychosocial stress status and PD. Accordingly, the more severe the stress in patients, the greater level of PD) (Peruzzo et al. 2007). A cross-sectional analysis showed that financial strain, which is a measure of chronic stress, was associated with a greater severity of PD measures either as clinical attachment loss or bone loss (Genco et al. 1999). However, subjects with the ability to cope with stressful/traumatic life events are less prone to develop more severe PD (Hugoson et al. 2002). Based on this, results Introduction 53 from a 2-year follow-up clinical trial demonstrated that patients with active coping behaviour had a milder disease levels than those with passive coping behaviour strategies (Wimmer et al. 2005). In terms of biological plausibility, higher cortisol levels were positively associated with more severe PD (Hilgert et al. 2006). Moreover, salivary cortisol was a predictor of periodontal attachment loss ≥ 5 mm (Rai et al. 2011; Rosania et al. 2009). On the other hand, stress can have a harmful effect on behaviour. As a result, subjects with stress may show poor oral hygiene, increased tobacco consumption, irregular dental check-ups as well as dietary changes, thus, worsening periodontal conditions. 1.4.8. Genetics It is believed that some genes could modify PD. Genetics factors such as gene-gene interactions or gene-environmental interactions (i.e., epigenetic factors) may also be an important issue in the development of PD. It has been shown that familial aggregation of aggressive forms of PD is often high among certain families (40-50%) (Meng et al. 2001). A Brazilian study showed familial aggregation of PD in three generations of families (Rapp et al. 2011). Classical studies in twins showed that a substantial part of chronic PD might be attributed to genetics (Michalowicz et al. 1991, 2000). On contrary, a lack of correlation between both mono and dizygotic twins and clinical attachment loss and bone loss was found (Torres de Heens et al. 2010). Hence, in chronic forms of PD, the genetic background is not clear and remains controversial. Regarding gene polymorphisms, up to date, there is no polymorphism that can be considered as a risk factor for PD. However, there is some interesting data in relation to IL-1 polymorphisms, in which it was found that IL-1A and IL-1B genetic variations were significant contributors to chronic PD, mainly in Caucasians (Karimbux et al. 2012). Evidence from an emerging brand of periodontal research, socalled epigenetics, hypothesized that the methylation status of genes YAGO LEIRA FEIJÓO 54 affecting PGE2 levels might be changed in periodontally-affected tissues, which suggests an epigenetic contribution to the inflammatory response posed by PD (Zhang et al. 2010). 1.5. Periodontal disease and atherosclerosis 1.5.1. Overview of periodontal disease as a chronic low-grade inflammatory condition Besides the localized nature of PD, a plethora of systemic markers of this condition have been reported and speculated to contribute to systemic diseases (Loos 2005). In periodontal health, the epithelial barrier in the oral cavity along with the protective innate immune molecules inhibits periodontopathogens from entering into the periodontal tissues and the bloodstream. Hence, in a healthy gingiva, only small counts of bacteria (mostly facultative) enter into the circulation (Li et al. 2000). When the periodontal infection is present within the gingiva, it is hypothesized that the inflamed and ulcerated subgingival pocket epithelium provides an easy entrance for periodontal bacteria, many of which are gram-negative anaerobic. Bacteraemia may occur in PD, as periodontopathogens are capable of colonizing distant sites (Haraszthy et al. 2000). In addition, bacterial components such as LPS may also be disseminated into the bloodstream. These LPS together with bacterial antigens can trigger significant systemic inflammatory processes. Accordingly, white blood cells (e.g. PMNs) and acute-phase reactants (APRs) from endothelial cells (ECs) and hepatocytes may produce pro-inflammatory mediators. Moreover, locally produced pro-inflammatory molecules (i.e., IL-1β, TNF-α, IL-6, and PGE2) are dumped into the systemic circulation and exert effects on distant organ systems (Moutsopoulos and Madianos 2006). Therefore, PD elicits a low-grade systemic inflammatory state. Introduction 55 1.5.2. Inflammatory mechanisms 1.5.2.1. Pro-inflammatory state Several inflammatory biomarkers are present in elevated levels in the systemic circulation of periodontal patients than those without PD. It is hypothesized that pro-inflammatory cytokines and other markers are produced in the periodontal lesion (Preshaw and Taylor 2011). These mediators may be dumped into the bloodstream. If this happens and the biomarkers reach a level in which bioactivity can be preserved, they would impact tissues and distant organs. For example, they can affect the liver leading to an acute-phase response that would impact other organs. As a result, a number of chained events take place such as inflammatory changes in the endothelium, up-regulation of vascular adhesion molecules, promotion of cytokine production and, finally, initiation and acceleration of atheroma development (Schenkein and Loos 2013). As stated before, periodontal patients have frequent bacteremic episodes with LPS being present in the systemic circulation. Additionally, experimental data demonstrated that Pg infection could promote inflammatory responses in distant organs from the oral cavity (i.e., atheroma) (Gibson et al. 2006; Gibson and Genco 2007). A metaanalysis of 702 patients with PD and 902 non-PD subjects showed a statistically significant weighted mean difference (WMD) between groups of 1.65 mg/L (95% CI: 1.05-2.24) in serum/plasma CRP levels, which is an APR produced in the liver in response to, among other cytokines, IL-6 (Paraskevas et al. 2008). In addition, it has been shown that periodontal treatment might reduce the levels of CRP or IL-6 (D’Aiuto et al. 2013). There is also evidence that MMPs, which play a pivotal role in both periodontal destruction (Meyle and Chapple 2015) and atherosclerotic plaque rupture, and can be induced by oral bacterial products (Hajishengallis et al. 2002). Gingipains, which are Pg proteases, can stimulate the production of MMPs as well as activate these markers when they are latent (Inamura et al. 2003). It has been demonstrated that high plasma MMP concentrations, especially MMP-9, in the acute YAGO LEIRA FEIJÓO 62 two main groups exist that are brain infarction and transient ischemic attack (TIA). While brain infarct is defined as a qualitative or quantitative of the blood supply to a part of the brain with a neurologic dysfunction lasting > 24 hours due to cell death, TIA is a transient episode of neurologic dysfunction caused by ischemia (loss of blood supply), without acute infarction lasting < 24 hours. From a clinical point of view, the most widely used classification for ischemic stroke is the Trial of Org 10172 in Acute Stroke Treatment (TOAST) classification system, which is based on the ischemic stroke etiology (Adams et al. 1993). The TOAST criteria includes the following categories:  Large-artery atherosclerosis: these patients will have clinical and brain imaging findings of either significant (> 50%) stenosis or occlusion of a major brain artery or branch cortical artery, presumably due to atherosclerosis. Diagnostic studies should exclude potential sources of cardiogenic embolism.  Cardioembolism: This category includes patients with arterial occlusions presumably due to an embolus arising in the heart (i.e., mechanical prosthetic valve, atrial fibrillation, atrial thrombus, sick sinus syndrome, recent myocardial infarction, left ventricular thrombus, dilated cardiomyopathy, akinetic left ventricular segment, atrial myxoma, and infective endocarditis). Potential large-artery atherosclerotic sources of thrombosis should be eliminated.  Small-artery occlusion: Patients included in this category are those whose strokes are often labelled as lacunar infarcts (LIs) in other classifications (Bamford et al. 1987). The patient should have one of the traditional clinical lacunar syndromes and should not have evidence of cerebral cortical dysfunction. A history of diabetes mellitus or hypertension supports the clinical diagnosis. The patient should also have a normal computed tomography/magnetic resonance imaging (CT/MRI) examination or a relevant brain stem or subcortical hemispheric lesion with a diameter < 1.5 cm demonstrated. Potential cardiac Introduction 63 sources for embolism should be absent, and evaluation of the large extracranial arteries should not demonstrate a stenosis > 50% in an ipsilateral artery.  Acute stroke of other determined etiology: This category includes patients with rare causes of stroke, such as nonatherosclerotic vasculopathies, hypercoagulable states, or hematologic disorders. These patients should have clinical and CT or MRI findings of an acute ischemic stroke, regardless of the size or location. Diagnostic studies should reveal one of these unusual causes of stroke. Cardiac sources of embolism and large-artery atherosclerosis should be excluded by other studies.  Stroke of undetermined etiology: Some patients will have no likely etiology determined despite an extensive evaluation. In others, no cause is found but the evaluation was cursory. This category also includes patients with two or more potential causes of stroke so that the physician in unable to make a final diagnosis. 3. CEREBRAL SMALL VESSEL DISEASE 3.1. Definition The term cerebral small vessel disease (CSVD) refers to a syndrome of clinical and imaging findings that are suggested to result from pathologies in perforating cerebral arterioles, capillaries and venules. CSVD is responsible for almost 45% of dementia, and accounts for about 20% of all stroke worldwide, 25% of ischemic strokes (or lacunar strokes), and about 20% of these cases present some level of disability (Pantoni 2010). CSVD includes small subcortical infarcts or LIs, silent infarcts (SIs), leukoaraiosis (white matter hyperintensities), prominent perivascular spaces (PVS) or VirchowRobin spaces, and cerebral microbleeds (CMBs) (Figure 3). Normally, LIs cause acute stroke symptoms, while other CSVD lesions are clinically more insidious and, therefore, are referred to as “silent lesions”. YAGO LEIRA FEIJÓO 64 Figure 3. Neuroimaging examples of different types of CSVD. a: LI. b: SI. c: Leukoaraiosis. d: PVS. e: CMBs. 3.2. Subtypes 3.2.1. Lacunar infarct Ischemic LI is defined as a stroke within a small deep perforating artery that is attributable to a recent small infarct (< 1.5 cm diameter) in the white matter, basal ganglia, pons or brainstem, and is consistent with a lacunar clinical syndrome (Wardlaw et al. 2013a). The main causes of LIs are atheroma of parent arteries or perforating arterioles, embolism from the heart or carotid arteries, and lipohyalinosis or fibrinoid necrosis (Bailey et al. 2012). Atheroma in the middle cerebral artery is responsible for almost 20% of LI cases. Evidence regarding embolism as a common cause for LI is limited (< 10% of LI cases). Although the exact mechanism remains unclear, intrinsic CSVD such as lipohyalinosis (referred to hyaline deposition in the perforating arteries together with diffuse arteriopathy) is considered Introduction 65 as the most common cause of LI (Figure 4). In this sense, it has been suggested that endothelial dysfunction is present in the pathogenesis of atherothrombotic neurovascular disease such as LI (Blanco et al. 2005; Knottnerus et al. 2009). In a group of patients with different types of ischemic stroke, impaired FMD was more conspicuous in LI patients than those diagnosed with other types of ischemic stroke (Chen et al. 2006); and patients diagnosed with LI had a diminished FMD compared to healthy controls and to those with similar vascular risk factors (Pretnar-Oblak et al. 2006). In addition, it was found that patients with LI had significantly higher serum levels of ICAM-1 than controls (Hassan et al. 2003). Risk factors for LI may include aging, diabetes mellitus or hypertension (Bezerra et al. 2012; Mast et al. 1995), and they are also considered to be predictors of poor outcome in these patients (Norrving 2003; Blanco et al. 2006). Moreover, there is an association between increased inflammatory response and early neurologic worsen in subjects with LI, thus, elevated concentrations of inflammatory markers such as IL-6 or TNF-α are related also to poor prognosis (Castellanos et al. 2002). Figure 4. Schematics of LI physiopathology. Lipohyalinosis Endothelial dysfunction BBB dysruption Risk factors LACUNAR INFARCT Arteriolar lumen narrowing Plasma fluids components deposition Vascular integrity loss Fibrinoid deposition and hypertrophy of smooth muscle cells YAGO LEIRA FEIJÓO 66 3.2.2. Silent infarcts SIs of presumed vascular origin are round or ovoid, subcortical, fluid-filled cavities with a diameter ranging from 3 to 15 mm. SIs can occur without any prior symptoms, but can also be the result of a previous acute small subcortical infarct (i.e., LI) or hemorrhage (Wardlaw et al. 2013b). At fluid-attenuated inversion recovery (FLAIR) MRI, a SI is usually presented as a hypointense hole sometimes surrounded by a hyperintense rim. Although SIs might have lacked acute symptoms, when present in large numbers they are related to dementia, cognitive impairment and increased risk of stroke (Vermeer et al. 2003, 2007). Results from the Cardiovascular Health Study showed that SIs are much more prevalent than LIs and they tend to be associated with ageing and high systolic blood pressure (Longstreth et al. 1998). In addition to these variables, basal carotid atheromatosis or low HDL were found to be associated with incident SIs (van Dijk et al. 2008; Gouw et al. 2008). 3.2.3. Leukoaraiosis (white matter hyperintensities) Leukoaraiosis of presumed vascular origin are very common in older subjects and regarded as typical signs of CSVD. Symptoms of leukoaraiosis develop insidiously, such as cognitive impairment, dementia and depression (Pantoni 2010; Jiménez et al. 2008), but it increases near 3-fold the risk for developing stroke, 2-fold the risk for developing dementia and is also associated with a higher risk of overall death (Debette and Markus 2010). Leukoaraiosis are usually symmetrically and bilaterally distributed in the white matter. Examination by MRI shows these lesions as hyperintense in relation to the normal brain, and can be patchy or confluent depending on their severity. Mechanisms explaining leukoaraiosis may include chronic partial ischemia secondary to diffuse atherosclerosis, hypotensive episodes Introduction 67 (e.g., cardiac arrhythmia or postural hypointension) (McQuinn and O´Leary 1987; Sulkava and Erkinjuntti 1987), breakage of the BBB, leakage of toxic fluids into the white matter (Pantoni et al. 1993), SIs (Conklin et al. 2014), and venous collagenosis (Moody et al. 1995). Besides ageing, several studies showed an association between hypertension and leukoaraiosis (Liao et al. 1997; de Leeuw et al. 2002; Dufouil et al. 2001) and that effective hypertensive therapy was related with lower risk for developing this condition (de Leeuw et al. 2002; Dufouil et al. 2001). With regards to risk factors for progression of leukoaraiosis, again ageing and hypertension along with tobacco consumption were predictors of leukoaraiosis progression (van Dijk et al. 2008). Furthermore, ICAM-1 levels were associated with progression of leukoaraiosis, further supporting a potential causal role of endothelial activation in the pathogenesis of leukoaraiosis (Markus et al. 2005). 3.2.4. Prominent perivascular spaces or Virchow-Robin spaces PVS are the extension of subarachnoid spaces that surround cerebral microvessels (Braffman et al. 1988). When enlarged, are commonly seen as hyperintense on T2 MRI either punctuate or linear depending on how the image is situated with respect to the course of the vessel and with a diameter < 3 mm (Potter et al. 2015), but sometimes can be larger (Hernández-Mdel et al. 2013). Usually, PVS do not ha hyperintense rim on T2-weighted (T2-w) or FLAIR unless passing through a white matter hyperintensity area, which can help to discriminate between PVS and SIs. It is suggested that more PVS could be associated with hypertension, cognitive decline and increased risk of dementia (Hernández-Mdel et al. 2013; Zhu et al. 2010; Maclullich et al. 2004). Although it has been hypothesized that impaired BBB or blockage of drainage of interstitial fluid might be responsible for enlarged PVS, the exact mechanisms underlying the physiopathology of prominent PVS remains unclear (Wardlaw et al. 2009; Weller et al. 2009). YAGO LEIRA FEIJÓO 68 3.2.5. Cerebral microbleeds CMBs are small perivascular hemosiderin deposits (usually with macrophages), which presumably result from leakage through cerebral small vessels that can be visualized as small, rounded, homogeneous, and hypointense lesions on T2-w (gradient-recalled echo) MRI or susceptibility-weighted imaging. Besides perivascular hemosiderinladen macrophages, other pathological correlations with CMBs are old haematomas, intact erythrocytes, and microaneurysms (Shoamanesh et al. 2011). Lipofibrohyalinosis and amyloid angiopathy are the most common vascular findings in CMBs. Although most CMBs are asymptomatic, they can be associated with amyloid deposition due to its potential relationship with stroke and dementia (Cordonnier et al. 2007; Martínez-Ramirez et al. 2014). Ageing and hypertension as well as diabetes are considered to be associated with both previous and incident cases of CMBs (Cordonnier et al. 2007, Poels et al. 2011). 4. PERIODONTAL DISEASE AS A RISK FACTOR FOR STROKE There is plenty of evidence with regards to the association between PD and stroke (Table 2, 3, and 4). The vast majority of studies show a positive relationship between both diseases, but the lack of consensus regarding definitions of both diseases makes difficult to establish the accurate magnitude of the effect. Firstly, some studies include stroke as the outcome without discriminating between ischemic and hemorrhagic stroke (Morrison et al. 1999; Diouf et al. 2015; Lee et al. 2006; Beukers et al. 2017). This is a critical issue since hemorrhagic stroke has a different aetiology from cerebral ischemia and, thus, biologically is less feasible that PD could predispose to hemorrhage. In fact, when they are analysed separately, no association was found between PD and hemorrhagic stroke (Wu et al. 2000). Secondly, self-reported diagnosis of PD rather than clinical periodontal records can lead to bias when PD is defined (Howell et al. 2001; Joshipura et al. 2003; Yu et al. 2015). Similarly, stroke diagnosis should be based on CT/MRI examination and, several studies include self-reported measurements of stroke (Elter et al. 2003; Lee et al. 2006). As a result, meta-analyses available in this topic have similar Introduction 69 methodological problems (Janket et al. 2003; Khader et al. 2004; Sfyroeras et al. 2012; Lafon et al. 2014a). In order do avoid these issues, recently, a systematic review and meta-analysis was published with the aim to investigate the potential association between PD and ischemic stroke (Leira et al. 2017b). All the included studies (3 prospective cohort studies and 5 case-control studies) (Beck et al. 1999; Wu et al. 2000; Jimenez et al. 2009; Grau et al. 2004; Dörfer et al. 2004; Sim et al. 2008; Pradeep et al. 2010; Lafon et al. 2014b) defined PD with clinical measurements (i.e., CAL and PPD) and cerebral ischemia was based on reliable examinations (i.e., acute ischemic lesion on brain imaging and/or neurological deficit) and valid classifications such as the International Classification of Diseases (ICD) (Kokotailo and Hill 2005) and TOAST criteria (Adams et al. 1993). Results from a meta-analysis showed that, overall, patients with severe PD had 2.8-fold increased risk for developing ischemic stroke. If studies were analysed according to study design, the RR for cohort studies was 2.52 (95% CI: 1.77-3.58) and for case-control studies was slightly higher (RR=3.04; 95% CI: 1.10-8.43) (Figure 5). Therefore, it seems that subjects with severe PD are at a higher risk for having ischemic stroke. First author, year Type of study Study population Country Stroke diagnosis PD diagnosis Follow -up (years) Confounders adjusted RR (95% CI) Beck, 1996 Prospective 1,147 men United States of America Ischemic stroke: history and physical examination detecting sustained neurological consistent with cerebral thrombosis (ICD-8, 432-436) Alveolar bone loss at interproximal tooth surfaces and worst clinical probing depth per tooth 25 Age, smoking, family history of heart disease, diastolic blood pressure, education (high school education or less) and type 2 diabetes Whole mouth bone loss > 20%: 2.80 (1.45-5.48) Morrison, 1999 Retrospective 10,120 subjects (35-84 years of age) Canada Fatal cerebrovascular disease (ICD-8, 430-438) Obvious pockets and loose teeth 23 Age, serum total cholesterol, smoking status, diabetes, hypertensive status and province of residence 1.63 (0.72 - 3.67) Wu et al W u et al 9962 adultos Edad, 10 10 10 10 Periodontitis e Wu, 2000 Prospective 9,962 subjects (25-74 years of age) United States of America Ischemic stroke (ICD-9, 433-434 and 436-438) 4 or more teeth with overt pockets or 10 Sex, age, race, education, poverty index, diabetes status, Ischemic stroke (incident YAGO LEIRA FEIJÓO 70 Hemorrhagic stroke (ICD-9, 430-432) worse conditions hypertension, smoking status, average alcohol use, BMI and serum cholesterol events): 2.11 (1.30-3.42) Ischemic stroke (fatal events): 2.90 (1.49-5.62) Hemorrhagic stroke (incident events): 1.22 (0.53-2.83) Hemorrhagic stroke (fatal events): 1.12 (0.32-3.89) Howell, 2001 Prospective 22,071 U.S. male physicians (40-84 years of age) United States of America Nonfatal stroke (typical neurological deficit, either sudden or rapid in onset, that lasted >24 h and was attributed to a cerebrovascular event) Self - reported 12.3 Age, aspirin and beta-carotene treatment assignment, cigarette smoking, alcohol use, history of hypertension, BMI, reported history of diabetes, physical activity, parental history of myocardial infarct and angina 1.01 (0.81 - 1.27) Joshipura, 2003 Prospective 41,380 male health professional s (40-75 years of age) United States of America Ischemic stroke (according to the National Survey of Stroke) Self - reported 12 Age, smoking, alcohol consumptions, BMI, physical activity, family history of myocardial infarct, multivitamin supplement use, vitamin E use, history of hypertension, diabetes, hypercholesterolemia and professions 1.33 (1 .03 - 1.70) Jimenez , 2009 Prospective 1,137 men United States of America Ischemic stroke: history and physical examination detecting sustained neurological consistent with Radiographic alveolar bone loss and cumulative PPD 24 Age, BMI, HDL, total cholesterol, triglycerides, diagnosis of hypertension, mean systolic and diastolic blood, diabetes diagnosis, Mean bone loss > 20%: 3.52 (1.597.81) Cumulative PPD > 30 mm: Introduction 71 cerebral thrombosis (ICD-8, 432-436) daily alcohol consumption, comprehensive smoking index, marital status, and baseline measures of education, occupation and income 1.07 (0.59 - 1.93) Yu, 2015 Prospective 39,863 female healthcare professional s (≥ 45 years of age) United States of America Ischemic stroke: focal neurologic deficit of sudden onset that persisted > 24 h (clinical information, CT and MRI was used in order to distinguish hemorrhagic from ischemic events) Self - reported 15.7 Age, race/ethnicity, BMI, education, smoking, diabetes, hypertension, hypercholesterolemia, family history of myocardial infarct and physical activities Prevalent PD: 1.12 (0.911.37) Incident PD: 1.41 (1.021.95) Sen, 2018 Prospective 6,736 dentates (45-64 years of age) United States of America Ischemic stroke (according to the National Survey of Stroke) Perio dontal profile classes (PPCs): PPC-A: periodontal health PPC-B: mild PD PPC-C: high gingival index score PPC-D: tooth loss PPC-E: posterior PD PPC-F: severe tooth loss PPC-G: severe PD 15 Age, sex, race/center, BMI, hypertension, diabetes mellitus, LDL, smoking, pack years, education Incident ischemic stroke (overall): 2.06 (1.41-3.01) Cardioembolic subtype: 2.62 (1.22-5.63) Lacunar subtype: 1.34 (0.64-2.79) Thrombotic subtype: 2.18 (1.26-3.78) Table 2. Summary of the most relevant cohort studies evaluating the association between PD and stroke. YAGO LEIRA FEIJÓO 78 carotid artery atherosclerosis (Hosomi et al. 2012). Recently, subanalysis from the ARIC study confirmed these results, in which it was concluded that PD is independently associated with incident ischemic stroke (Sen et al. 2018). Regarding intervention studies, up to date, there is a lack of prospective clinical trials evaluating the effect of periodontal therapy in patients with stroke and surrogates markers of cerebral ischemia. However, results from retrospective analyses from the Taiwanese National Health Insurance (NIH) research database showed that dental prophylaxis and periodontal treatment was a protective factor for having incident ischemic stroke [Hazard ratio (HR)=0.79; 95% CI: 0.75-0.81 and HR=0.95; 95% CI: 0.91-0.99, respectively] (Lee et al. 2013). In addition, those subjects who received ≥ 1 tooth scaling within 2 years were less prone to develop a stroke (HR=0.81; 95% CI: 0.730.92) (Cheng et al. 2012). 5. BIOMARKERS 5.1. Systemic inflammation 5.1.1. Interleukin-6 IL-6 is a pluripotent cytokine with 26 kDa molecular mass, containing 185 amino acids and four-helix bundles structure, which is primarily derived from macrophages. This cytokine acts through its soluble receptor (sIL-6R) or its cellular receptor (IL-6R) (Jones et al. 2001). The main role of IL-6 is to be a pro-inflammatory cytokine considering that it can amplify inflammatory responses. Other roles of this cytokine include autoimmunity induction, differentiation of B and neuronal cells, APRs [e.g., CRP, serum amyloid protein A (SAA), haptoglobin, or fibrinogen] induction in hepatocytes, and induction of receptor activator of NF-κB ligand (RANKL) that will activate osteoclasts and bone resorption (Hirano et al. 1988; Ridker et al. 1997; Guo et al. 2011). Besides pro-inflammatory cells such as lymphocytes, macrophages, monocytes and neutrophils, other non-proinflammatory Introduction 79 cells are responsible for the production of IL-6 (i.e., fibroblasts, myocytes, osteoblasts, and ECs) (Matsuki et al. 1992). This cytokine is considered to be an upstream regulator that plays a central role in propagating the downstream inflammatory response related to atherosclerosis. Indeed, IL-6 acts at different stages in the process of atherosclerosis (Hartman and Frishman 2014). In the development of an atherosclerotic plaque, IL-6 is largely responsible for coordinating the influx of inflammatory cells. ECs respond to the binding of a complex of IL-6 and sIL-6R by elaborating chemokines and increasing ICAM-1 expression, leading to leucocyte recruitment and transmigration (Romano et al. 1997). IL-6 stimulates the hypothalamic-pituitary-adrenal (HPA) axis through the peripheral synthesis of corticotropin-realising factor, thus, promoting proinflammatory effects (Späth-Schwalbe et al. 1994; Karalis et al. 1997). IL-6 activity upon naïve T-lymphocytes results in their differentiation into T-helper cells, which are able to continue propagation of the inflammatory cascade (Eddahri et al. 2009). In addition, IL-6 has procoagulant effects, primarily by tissue factor (TF) induction within monocytes (Neumann et al. 1997). The production of TF allows thrombus formation, and as a result, vascular event such as ischemic stroke (Rodríguez-Yáñez and Castillo 2008) at the site of atherosclerosis. Inflammation is considered to play an important role in cerebral infarct deterioration (Castillo and Leira 2001). In patients with ischemic stroke, elevated levels of plasmatic IL-6 have been associated with the presence of early neurological deterioration (Castellanos et al. 2002), poor functional outcome (Blanco et al. 2006; Castellanos et al. 2002; Rodríguez-Yáñez et al. 2006), early CT ischemic changes (RodríguezYáñez et al. 2008) as well as greater volume infarct, which is one of the main factors implicated in functional outcome of these patients (Leira et al. 2006). Results from a multicentre study demonstrated that elevated plasmatic levels of IL-6 in baseline increased the risk of new vascular disease event or death from vascular disease in patients with ischemic stroke, who were not anti-coagulated (Castillo et al. 2009). On YAGO LEIRA FEIJÓO 80 the other hand, lower levels of IL-6 are associated with ischemic tolerance in patients with ischemic stroke and a previous TIA, hence, with a good functional outcome (Castillo et al. 2003). In PD, IL-6 seems to play a role in modulating the response to periodontal bacteria, leading to both local and systemic inflammation. An excessive IL-6 response may contribute to the development of a chronic inflammatory lesion, which can result in periodontal tissue breakdown and alveolar bone loss via MMPs and osteoclasts activity, T cells activation and amplification o the inflammatory cascade (Nibali et al. 2012). It has been shown that IL-6 levels were elevated both locally (i.e., GCF and saliva) and systemically (i.e., serum) in PD patients (Geivelis et al. 1993; Costa et al. 2010; Shimada et al. 2010). Moreover, periodontal treatment seems to reduce in the long-term serum IL-6 levels in severe periodontal patients (D’Aiuto et al. 2004; Shimada et al. 2010). 5.1.2. Interleukin-10 IL-10 modulates expression of cytokines, soluble mediators and cell surface molecules by cells of myeloid origin, with important consequences for their ability to activate and sustain immune and inflammatory responses (Moore et al. 2001). The effect of IL-10 on cytokine production and function of human macrophages are generally similar to those on monocytes, but less pronounced (Wilkes et al. 1995; Armstrong et al. 1996; Thomassen et al. 1996; Nicoid et al. 1995). The inhibitory effects of IL-10 on IL-1 and TNF production are crucial to its anti-inflammatory activities, because these cytokines often have synergistic activities on inflammatory pathways and processes, and amplify these responses by inducing secondary mediators such as chemokines and PGs (Niiro et al. 1994, 1995; Mertz et al. 1994). Furthermore, IL-10 downregulates the expression of TLR-4, which is the signal transducing receptor for bacterial LPS (Muzio et al. 2000). Introduction 81 Low or absent IL-10 leads to several changes in gene expression that ultimately results in deleterious vascular remodelling and impaired vascular relaxation in response to physiologic mediators, which would exacerbate secondary brain damage following acute brain injury (Garcia et al. 2017). Accordingly, patients with LIs and neurological deterioration had significantly lower IL-10 levels in plasma than those without worsening (Vila et al. 2003). In addition, plasmatic levels of IL-10 ≥ 30 pg/mL were independently associated with a good outcome in ischemic stroke patients who underwent systemic thrombolytic treatment with tPA (Rodríguez-Yáñez et al. 2013). Due to its anti-inflammatory properties, IL-10 also plays an important role in PD (Sasaki et al. 2004). The lockout of PD may result in accelerating alveolar bone resorption and decreasing bone formation (Zhang et al. 2014). In fact, it has been demonstrated that IL-10 upregulated osteoprotegerin (OPG) expression but downregulated RANKL expression (Liu et al. 2006). Moreover, IL-10 may inhibits pro-inflammatory cytokines that are involved in osteoclasts proliferation, thus, inhibiting bone resorption (Boyle et al. 2003). On the other hand, IL-10 has the ability of promoting osteoblast differentiation eventually via pro-inflammatory cytokines downregulation (Dresner-Pollak et al. 2004). 5.2. Endothelial dysfunction 5.2.1 Pentraxin 3 Pentraxins (PTXs) belongs to the superfamily of APRs, which are divided into short and long PTXs. Short constituents include CRP and SAA, which are synthesized in the liver mostly upon IL-6 stimulation. On contrary, long constituent PTX3 is produced by neutrophils, fibroblasts, dendritic cells, macrophages, epithelial cells, and ECs, in response to pro-inflammatory signals such as bacterial products, TNFα, and IL-1β and by TLR receptor engagement (Mantovani et al. 2008; Vilahur and Badimon 2015). Indeed, plasma PTX3 levels are associated with vascular endothelial dysfunction, atherosclerosis, inflammation, or YAGO LEIRA FEIJÓO 82 damage and reach the peak much earlier than CRP (Norata et al. 2010; Peri et al. 2000). Evidence suggests that cardiovascular disease (CVD) (Fornai et al. 2016), chronic kidney disease (CKD) (Sjöberg et al. 2016), or increased risk of mortality after ischemic stroke (Ryu et al. 2012) are linked to higher PTX3 levels. With regards to PD, it has been hypothesized that PTX3 concentrations measured in GCF or serum could be a valuable diagnostic molecule for periodontal tissue destruction (Kathariya et al. 2013). In small-vessel vasculitides, PTX3 seems to be an indicator of disease activity, considering that endothelial cells from active skin lesions showed enhanced production of PTX3 (Fazzini et al. 2001). One of the key features of atherosclerosis, which is cholesterol accumulation in the intima of the vessels, is related to an immuneinflammatory response leading to the recruitment and activation of different cellular types (i.e., macrophages, monocytes, PMNs and ECs). These cells are responsible for producing PTX3 in response to inflammatory stimuli commonly associated with the process of atherosclerosis (Bonacina et al. 2013; Fornai et al. 2016). In fact, a higher expression of PTX3 could be observed in human advanced atherosclerotic plaques (Rolph et al. 2002) and it has been suggested that PTX3 could be an indicator of carotid plaque vulnerability (Shindo et al. 2014) by binding the fibroblast growth factor 2, which play a role in the proliferation and migration of smooth muscle cells (Bassi et al. 2009). Experimental data demonstrated that PTX3 null/Apolipoprotein E (ApoE) null mice presented larger atherosclerotic lesions compared to those ApoE null only, and this fact was due to raised bone marrow monocytosis, elevated macrophage accumulation, and increased expression of adhesion molecules, chemokines, and cytokines in the vessel wall (Norata et al. 2009). PTX3 has also been studied as a potential therapeutic target against atherosclerotic progression. Based on this, PTX3 suppression was associated with a reduction in the inflammatory and apoptosis process that was mediated by the IKK/IκB/NF-κB pathway (Qiu et al. 2015). However, to date, it remains Introduction 83 unclear which are the mechanisms underlying the role of this protein in the onset and progression of atheromatosis (Casula et al. 2017). PTX3 has also been studied as a contributor to vascular endothelial dysfunction. In the rodent model, exogenous administration of PTX3 significantly blunted NO production through the MMP-1 and P-selectin pathway leading to morphological alterations of ECs (Carrizzo et al. 2015). In addition, there is clinical evidence suggesting that plasma PTX3 is a more potent predictor of endothelial dysfunction compared to CRP (Yasunaga et al. 2014). Accordingly, an association was found between PTX3 levels and FMD, which is a direct measure of endothelial dysfunction. In patients with coronary artery disease (Yanusaga et al. 2014), CKD (Yilmaz et al. 2009a), or obstructive sleep apnea syndrome (Kanbay et al. 2015), PTX3 levels are negatively correlated with FMD. Therefore, giving insight of the potential role of this APR in endothelial impairment. A study carried out in chronic periodontal patients has shown a positive correlation between periodontal clinical parameters such as CAL (a measure of prolonged exposure to PD) and PPD (an indicator of current PD) and PTX3 levels in general (Pradeep et al. 2011) and specific infection site (Pradeep et al. 2011; Fujita et al. 2012). Furthermore, cross-sectional clinical data showed that in subjects diagnosed with periodontitis, PTX3 concentrations in GCF were significantly elevated in periodontally affected sites compared to nonPD sites (Fujita et al. 2012). Regarding saliva samples of PTX3 levels, significant differences could only be seen at the subgroup level, namely generalized aggressive PD subgroup, when comparisons were made with controls (Gümüş et al. 2014) (Table 5). YAGO LEIRA FEIJÓO 84 First author, year Study design Diagnosis Sample size Age (mean) PTX3 levels (ng/mL) Method Pradeep, 2011 Case-control (ageand sexmatched) Chronic PD: Gingival Index (GI) >1; PPD ≥5 mm; CAL ≥3 mm; and radiographic evidence of bone loss Patients: 15 Controls: 10 Patients: 33 Controls: 23 Patients: 3.07 ± 0.71 Controls: 1.60 ± 1.12 Serum (ELISA) Patients: 3.37 ± 1.45 Controls: 1.95 ± 0.91 GCF Pradeep, 2012 Case - control (ageand sexmatched) Ch r onic PD: GI >1; PPD ≥5 mm; CAL ≥3 mm; and radiographic evidence of bone loss. Cases were also diagnosed with CKD Patients: 20 Controls: 20 Patients: 37 Controls: 30 Patients: 6.33 ± 2.74 Controls: 1.83 ± 0.75 Serum (ELISA) Fujita, 2012 Cross - sectional A l l patients were diagnosed with chronic PD. -Diseased sites: PPD ≥5 mm; CAL ≥3 mm and BoP -Healthy sites: PPD ≤3 mm and no BoP 50 59 Diseased sites: 0.64 ± 0.39 Control sites: 0.06 ± 0.10 GCF Gümüş, 2014 Case-control (age-matched) Subgroup generalized chronic PD: presence of ≥4 teeth in each jaw with a PPD ≥5 mm; CAL ≥4 mm; ≥50% bone loss in at least 2 quadrants; and BoP in >80% of proximal sites Patients: 25 Controls: 22 Patients: 50 Controls: 47 Patients: 47.74 ± 28.16 Controls: 36.66 ± 23.37 Saliva S ubgroup generalized aggressive PD: presence of at least 6 permanent teeth (including incisors and/or 1st molars) with at least one site with PPD and CAL ≥5 mm and 6 teeth other than 1st molars and incisors with similar PPD and CAL measurements; familial aggregation; and radiographic bone loss ≥30% of root length affecting ≥3 permanent teeth other than incisors and 1st molars Patients: 25 Controls: 22 Patients: 36 Controls: 36 Patients: 51.95 ± 29.04 Controls: 29.85 ± 25.21 Saliva Table 5. Summary of observational studies reporting PTX3 levels in patients with PD. Introduction 85 5.2.2. Soluble tumor necrosis factor-like weak inducer of apoptosis Tumor necrosis factor-like weak inducer of apoptosis (TWEAK) is a member of the TNF superfamily of cytokines that is synthesized as a type II transmembrane protein from which a soluble 17 kDa ligand factor with biological activity can be released [soluble TWEAK (sTWEAK)] (Chicheportiche et al. 1997). TWEAK is expressed in several tissues (i.e., heart, lung, and brain) (Wiley and Winkles 2003) and cells (i.e., ECs and SMCs). Different cellular responses are induced by sTWEAK such as cell proliferation, migration, differentiation, and angiogenesis (Lynch et al. 1999; Donohue et al. 2003; Desplat-Jégo et al. 2002; Tran et al. 2003). In addition, it is involved in the expression of pro-inflammatory mediators and adhesion molecules in ECs and astrocytes (Harada et al. 2002; Saas et al. 2000). TWEAK activity is mainly mediated via binding to fibroblast growth factor-inducible 14 (Fn14), which is a 14 kDa member of the TNF receptor family (Wiley et al. 2001) and is expressed by fibroblasts, ECs, epithelial cells, and tumor cells of non-lymphoid origin. The interaction between TWEAK and Fn14 has been associated with endothelial dysfunction and the process of atherosclerosis by several mechanisms (Liu et al. 2017). TWEAK induces expression of adhesion molecules such as E-selectin and ICAM-1 on the cell surface of human umbilical vein ECs in vitro (Harada et al. 2002) and the TWEAK-Fn14 axis also induce MCP-1 and IL-8 secretion, which predominantly recruits monocytes and neutrophils (Blanco-Colio et al. 2007a). Vascular SMCs, which normally are in a contractile phenotype to maintain the vascular tone and diameter, in atherosclerosis is transformed into a pro-inflammatory phenotype that could be elicited through TWEAK/Fn14 interaction (Harada et al. 2002; Lynch et al. 1999). Furthermore, is has been demonstrated that in aortic SMCs, TWEAK was able to induce thrombotic molecules such as TF and PAI1, which are involved in the process of thrombosis (Muñoz-García et al. 2011). TWEAK/Fn14 axis also showed some effects on monocytes and macrophages involved in the inflammatory process of atherosclerosis. Through Fn14 inhibition, the interaction TWEAK/ YAGO LEIRA FEIJÓO 86 Fn14 may alter macrophage trafficking and increase lipid uptake of macrophages (Schapira et al. 2009). In addition, TWEAK was found to induce several pro-inflammatory mediators of atherogenesis such as IL6, MCP-1 and IL-8 in activated monocytes (Kim et al. 2004) as well as able to enhance MMP-9 and MMP-2 activity in ApoE null mice (Sastre et al. 2014). Human studies showed that patients with CKD, sTWEAK levels were positively correlated with FMD (Yilmaz et al. 2009b, 2011) and associated with atheromatosis progression (Fernández-Laso et al. 2017). It was also found that in artery samples from patients with atherosclerosis, sTWEAK concentrations were negatively correlated with carotid IMT (Blanco-Colio et al. 2007b). In cerebral ischemia, experimental data showed increased levels of TWEAK and Fn14 in the area surrounding the necrotic core, also known as ischemic penumbra region, after middle cerebral artery occlusion (MCAO) (Yepes et al. 2005). Moreover, intracerebroventicular injection of a soluble Fn14-Fc decoy receptor immediately after MCAO significantly reduced infarct volume as well as the apoptotic cell death in the ischemic penumbra region (Yepes et al. 2005). Results from an animal study showed that in response to ischemic signal during stroke, there was a release of TWEAK from astrocytes and its binds to Fn14 receptor activated the NF-κB pathway and also induced pro-inflammatory cytokines and MMP-9 leading to BBB disruption and increased permeability (Polavarapu et al. 2005). Additionally, administration of Fn14-Fc decoy receptor immediately or 1 hour after MCAO improved neurovascular permeability (Zhang et al. 2007). In humans, serum levels of TWEAK from ischemic stroke patients were significantly elevated compared to healthy controls (Inta et al. 2008). In brain tissues of patients who died due to ischemic stroke, the levels of Fn14 from infarct tissues were significantly higher than those brain tissues from the contralateral region without infarct (Inta et al. 2008). Regarding PD, there is scarce literature of the potential role of TWEAK in the pathogenesis of the disease. However, it was found that in periodontally affected tissues, TWEAK and Fn14 were overexpressed compared to healthy periodontal tissues (Kataria et al. Introduction 87 2010) and that the TWEAK/Fn14 axis was responsible for the induction of IL-1β, ICAM-1 and VCAM-1 within human gingival fibroblasts (Hosokawa et al. 2006, 2012). 5.3. Amyloid-beta peptides Amyloid-beta (Aβ) is a 38-to 43 amino-acid peptide, which is produced by proteolytic cleavage of amyloid precursor protein (APP). Plasma Aβ includes both Aβ1-40 and Aβ1-42. Whereas parenchymal Aβ deposition is considered to play a pivotal role in neuronal loss and cognitive impairment in Alzheimer´s disease (AD) (Hardy and Allsop 1991), progressive Aβ accumulation within the walls of cortical and leptomeningeal small arterioles is the pathological hallmark of cerebral amyloid angiopathy (Charidimou et al. 2012). Senile plaque amyloid is primarily comprised of Aβ1-42, whilst amyloid of vascular origin consists of Aβ1-40 species. It has been hypothesized that exposure to Aβ1-40, but not Aβ1-42, produced a profound and selective alteration in the regulation of the cerebral circulation by ECs (Niwa et al. 2000a). Based on this, endothelial dysfunction by Aβ deposition has been related to LIs and leukoaraiosis (van Dijk et al. 2004; Gurol et a. 2006). Accordingly, diffuse-CSVD (i.e., LIs with leukoaraiosis) was independently associated with elevated plasma levels of Aβ1-40 (Gomis et al. 2009). It has been postulated that PD is involved in the synthesis and accumulation of Aβ in the brain. By means of positron emission tomography imaging techniques, PD was associated with Aβ deposition in brain areas known to be susceptible to AD (Kamer et al. 2015). Periodontally affected tissues from healthy patients also demonstrated overexpression of APP, thus, giving insight on the potential relationship between PD and Aβ accumulation (Kubota et al. 2014). Recently, it was found that serum Aβ1-42 levels from patients diagnosed with cognitive impairment who had severe PD were significantly higher than those without cognitive decline or those with lower levels of PD (GilMontoya et al. 2017). Clinical study 95 CLINICAL STUDY 1. JUSTIFICATION PD is among the ten most prevalent diseases affecting human beings. A growing body of evidence suggests that PD not only have local effects (i.e., gingiva) but also is capable of producing systemic side effects in distant organs (Loos 2005). In fact, it is speculated that the inflamed and ulcerated subgingival pocket epithelium forms an easy port of entry for periodontal bacteria. Bacteremia in PD has been reported after oral examination (Daly et al. 2001) and dental procedures (Heimdahl et al. 1990) and, interestingly, periodontopathogens have been identified in atheromatous plaques (Haraszthy et al. 2000). Additionally, bacterial components such as LPS may also be disseminated into the blood circulation (Geerts et al. 2002). These LPS together with bacterial antigens can trigger significant systemic inflammatory processes. Accordingly, white blood cells and acutephase proteins from ECs and hepatocytes may produce proinflammatory mediators. Moreover, locally produced pro-inflammatory molecules may spill into the systemic circulation and exert effects on distant organ systems (Moutsopoulos and Madianos 2006). Based on this, PD has been regarded as a systemic inflammatory and endothelial vascular stressor that can act as an independent risk factor of largevessel ischemic stroke (Leira et al. 2017c). LI, a type of CSVD, is responsible for approximately 25% of the cases of ischemic stroke (Pantoni 2010). Atheromatous stenosis, thrombosis occlusion (Fisher 1979) or inflammation (Wardlaw 2005) have been suggested as the main causes of this disease. Although the exact mechanism remains unclear, intrinsic CSVD such as lipohyalinosis is considered as the most common cause of LI. Accordingly, endothelial dysfunction is present in the pathogenesis of YAGO LEIRA FEIJÓO 96 LI (Blanco et al. 2005; Knottnerus et al. 2009). Furthermore, there is an association between increased inflammatory response and early neurologic worsen in patients with LI showing elevated concentrations of pro-inflammatory markers such as IL-6 or TNF-α in those with a poor functional prognosis (Castellanos et al. 2002). To date, little is known about the relationship between PD and small-vessel ischemic stroke (i.e., LI) with conflicting results (Leira et al. 2016; Sen et al. 2018). To the best of our knowledge, there is a lack of mechanistic studies investigating the potential role of PD as a source of systemic inflammation and endothelial dysfunction in patients with LI as well as if it could be a predictor of poor functional outcome in these patients. Therefore, identifying conditions that could contribute to an enhanced systemic inflammatory state as well as promote endothelial dysfunction may have significant prognostic and treatment implications in patients with LI. 2. HYPOTHESIS Our hypothesis is that PD is common in patients with LI and is associated with the presence of LI independently of well-known vascular risk factors. When PD is present in patients with LI, may result in an enhanced systemic inflammatory response with vascular dysfunction of the endothelium expressed by higher serum levels of both pro-inflammatory and endothelial dysfunction biomarkers. We also hypothesized that periodontal inflammation could be an independent predictor of poor functional prognosis in LI patients. Clinical study 97 3. OBJECTIVES 3.1. Primary objective  The main objective of this work is to investigate the association between PD and its clinical parameters and the presence of LI. 3.2. Secondary objectives  To analyse possible physiopathological mechanisms through which PD could contribute to a higher risk for developing LI (i.e., systemic inflammation and endothelial dysfunction).  To investigate whether periodontal inflammation predicts poor functional outcome in patients diagnosed with LI. 4. MATERIAL AND METHODS 4.1 Study design A case-control study was carried out by the Periodontology Unit of the University of Santiago de Compostela in collaboration with the Stroke Unit of the University Clinical Hospital of Santiago de Compostela by following the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines (Von Elm et al. 2008). This research was performed in accordance with the Declaration of Helsinki of the World Medical Association (2008) and approved by the Ethics Committee of the Servizo Galego de Saúde (2016/399). Informed consent was obtained from each patient or their relatives after full explanation of the procedures. YAGO LEIRA FEIJÓO 98 4.2. Study population 4.2.1. Case group Patients who had attended the Stroke Unit of the University Clinical Hospital of Santiago de Compostela between January 2014 and January 2015 were asked by telephone to participate in this study as cases. Cases were those with a diagnosis of LI based on the TOAST criteria (Adams et al. 1993) and they were included in the study if they fulfilled the following inclusion criteria: (i) >18 years of age; (ii) at least 15 teeth (excluding third molars); and (iii) written informed consent. Exclusion criteria were as follows: (i) patient who have received periodontal treatment in the previous 12 months; (ii) history of neurovascular and/or neuroinflammatory disease; (iii) systemic antibiotics, corticosteroids, and/or immunosuppressant therapy within 3 months prior to periodontal assessment; and (iv) chronic use of nonsteroidal anti-inflammatory drugs (NSAIDs). 4.2.2. Control group Healthy control subjects, matched by age and gender, were selected from the hospital database of the Service of Neurology. In order to include individuals without any neurological disorder, we reviewed 194 CT/MRI scans of subjects who were referred to the Service of Neurology with a suspicious diagnosis of certain neurological diseases such as non-specific headache, vestibular syndromes, brain tumours or altered level of consciousness between 2009-2013. Of these, 12 presented some subtype of asymptomatic CSVD [SI (n=4), leukoaraiosis (n=8)] and, thus, were excluded from the study. Therefore, 182 subjects free from any neurological disease were contacted by telephone and asked to participate. Inclusion and exclusion criteria were the same as for the case group. Control individuals were clinically examined and interviewed in parallel with patient recruitment. Clinical study 99 4.3. Periodontal examination The periodontal examination was performed by a single calibrated periodontist (YL). The calibration was completed before the start of the study in the Periodontology Unit of the Faculty of Odontology (University of Santiago de Compostela) using 10 non-study patients suffering from moderate or severe PD. Intra-examiner reliability was assessed by the intraclass correlation coefficients (for PPD, GR and CAL), which were 0.79, 0.87 and 0.79, respectively, demonstrating a high degree of reliability in the measurements (Leira et al. 2016). In the present study the following periodontal parameters were evaluated in all teeth (except 3º molars): (i) PPD, measured from the free gingival margin to the bottom of the sulcus; (ii) CAL, measured from the cemento-enamel junction to the bottom of the sulcus or pocket; (iii) GR, measured as the distance from the free gingival margin to the exposed cemento-enamel junction; (iv) full-mouth plaque score (FMPS), defined as the number of sites with detectable supragingival dental plaque divided by the total number of sites per mouth, multiplied by 100 (O’Leary et al. 1972); (v) full-mouth bleeding score (FMBS), defined as the number of sites with gingival bleeding on probing divided by the total number of sites per mouth, multiplied by 100 (Ainamo and Bay 1975); and (vi) the number of missing teeth (excluding 3º molars). All measurements were recorded at six sites per tooth (mesiobuccal, disto-buccal, mid-buccal, mesio-lingual, disto-lingual, and midlingual), except for FMPS (four sites/tooth) using a sterile mouth mirror and with a calibrated University of North Carolina periodontal probe (UNC 15; Hu-Friedy, Chicago, IL, USA). The presence of PD was defined according to the CDC-AAP consensus for epidemiologic studies (Eke et al. 2012; Holtfreter et al. 2015). Therefore, mild PD was defined as ≥2 interproximal sites with CAL ≥3 mm and ≥2 interproximal sites with PPD ≥4 mm (not on the same tooth) or 1 site with PPD ≥5 mm. Moderate PD was defined as ≥2 interproximal sites with CAL ≥4 mm (not on the same tooth) or ≥2 YAGO LEIRA FEIJÓO 100 interproximal sites with PPD ≥5 mm, also not on the same tooth. Severe PD was defined as the presence of ≥2 interproximal sites with CAL ≥6 mm (not on the same tooth) and ≥1 interproximal site with PPD ≥5 mm. Total PD was the sum of mild, moderate, and severe PD. In addition, a recently introduced measure of PD severity and activity, the PISA was calculated (Nesse et al. 2008). PISA reflects the surface area of bleeding pocket epithelium in mm2. PISA was calculated with a Microsoft Excel spreadsheet in the following steps: (i) Mean CAL and GR for each particular tooth is calculated; (ii) Linear mean CAL and GR is translated into the periodontal epithelial surface area (PESA) for each specific tooth (Hujoel et al. 2001). The PESA for a particular tooth consists of the root surface area of that tooth measured in mm2, which is covered with pocket epithelium; (iii) The PESA for a specific tooth is then multiplied by the proportion of sites around the tooth that was affected by BoP, resulting in the PISA for that particular tooth; and (iv) The sum of all individual PISAs around individual tooth is calculated, rendering the full-mouth PISA value in mm2 of each participant. 4.4. Neurological examination 4.4.1. Neuroimaging examination A CT/MRI scan was carried out in all cases at admission. MRI images were obtained on a 1.5 T system (1.5 Magneton Symphony, Siemens, Erlangen, Germany), with echo planar capabilities of 25 mT/m gradients and 300-350 μs rise times. The MRI protocol included T1-w (TR/TE: 370/7.7 ms), T2-w (TR/TE: 6020/113 ms), DP-w (TR/TE: 6020/113 ms) and FLAIR (TR/TE: 9000/114 ms) (Rodríguez et al. 2010). One neurologist who was blinded to the clinical data carried out the evaluation of CT/MRI (MR-Y). LI was diagnosed if the patient had one of the characteristic clinical lacunar syndromes, neurological deficit lasting >24 hours, no evidence of cerebral cortical dysfunction, and a CT/MRI that was normal or showed a deep focal infarction in an appropriate location with a diameter ≤15 mm. The Clinical study 101 presence of a LI in the baseline CT in which the topography does not correspond with the present clinical syndrome was considered a SI. For the purpose of this study, the thalamus was included, along with the caudate nucleus, putamen, and globus pallidus, as “basal ganglia.” Infarct volume was not calculated because of the unreliability of the CT in measuring small infarcts. Leukoaraiosis was defined as ill-defined hyperintensities ≥5 mm on both T2 and FLAIR MRI images without prominent hypointensities on T1-w MRI scans and as ill-defined and moderately hypodense areas of ≥5 mm on CT. Leukoaraiosis was classified according to the Fazekas criteria (Fazekas et al. 1991, 1993) using the modified Fazekas scale (Pantoni et al. 2002). This method yields two separate scores for subcortical and deep white matter lesions and periventricular lesions. The four-point Fazekas scale of increasing severity was used to classify each score. For the purpose of the study, the presence of leukoaraiosis was categorized with 0 indicating a patient without leukoaraiosis and 1 with leukoaraiosis. Grade Description 0 No white matter lesions 1 Lesion limited to frontal and/or occipital ventricle horn 2 Lat eral ventricle bands 3a Disperse periventricular and/or subcortical white matter lesions 3b Confluent white matter and/or subcortical lesions 3c Homogeneous white matter and/or subcortical lesions Table 6. Modified Fazekas scale. 4.4.2. Ultrasound examination The same explorer (SA), blinded to clinical data, performed the ultrasonographic study using high-resolution B-mode ultrasound [Aplio 50 (Toshiba aplio 50, MCM1754TSA, Rome, Italy) Toshiba SSA-700 (Toshiba Medical Systems Corporation, Otawara-SHI, Japan)] with a 7.5 MHz, linear-array transducer (Linear array transducer PLT-704AT, Toshiba, Tochigi, Japan; Phased array transducer PST-20CT, Toshiba, Tochigi, Japan) (Rodríguez et al. 2010). In brief, the image was focused on the posterior (far) wall of the left carotid artery. A minimum of four measurements of the common carotid far wall was taken 10 mm YAGO LEIRA FEIJÓO 102 proximal to the bifurcation, to derive the mean carotid intima-media thickness (IMT) (Raitakari et al. 2003). The presence of an atheroma plaque was evaluated in the common and internal carotid extracraneal arteries as well as the bifurcations according to standardised scanning and reading protocols (Touboul et al. 2007). Plaque was defined as a focal structure that encroaches into the arterial lumen at least 0.5 mm or 50% of the surrounding IMT value, or demonstrates a thickness >1.5 mm as measured from the media-adventitia interface to the intimalumen interface. For the purpose of the study, the presence of carotid atheromatosis was categorized with 0 indicating a patient without carotid atheromatosis and 1 with it. 4.4.3. Outcome evaluation Functional outcome was evaluated at 3 months using the modified Rankin scale (mRS) (UK TIA Study Group, 1988; van Swieten et al. 1988), and a poor outcome was defined as a mRS score >2. Grade Description 0 No symptoms at all 1 No significant disability despite symptoms: able to carry out all usual duties and activities 2 Slight disability: unable to carry out all previous activities but able to look after own affairs without assistance 3 Moderate disability: requiring some help, but able to walk without assistance 4 Moderately severe disability: unable to walk without assistance, and unable to attend to own bodily needs without assistance 5 Severe disabilit y: bedridden, incontinent, and requiring constant nursing care and attention 6 Death Table 7. Modified Rankin Scale scores. 4.5. Serum collection and laboratory tests On admission for cases and the periodontal examination/interview day for controls, 2 mL of venous blood was collected from the antecubital fossa by venepuncture using a 20-gauge needle with a 2 mL syringe. Blood samples were allowed to clot at room temperature and Clinical study 103 after 1 hour, serum was separated from blood by centrifugation (15 min. at 3000 g) and 0.5 mL of extracted serum was immediately transferred to 1.5 mL aliquots. Each aliquot was stored at –80ºC until the time of analysis. Serum levels of all biomarkers were measured by enzymelinked immunosorbent assay (ELISA) technique following manufacturer instructions. IL-6 ELISA kit (Proteintech®, Manchester, United Kingdom) minimum assay sensitivity was 3.8 pg/ml with a intra-assay coefficient of variation (CV) of 5.0% and inter-assay CV of 6.4%; IL-10 ELISA kit (Proteintech®, Manchester, United Kingdom) minimum assay sensitivity was 0.5 pg/ml with a intra-assay coefficient of variation of 5.4% and inter-assay CV of 5.6%; PTX3 ELISA kit (Abnova™, Taipei City, Taiwan) minimum assay sensitivity was 10 pg/ml, with a intra-assay coefficient of variation of 6.9% and interassay CV of 7.0%; sTWEAK ELISA kit (Aviscera Bioscience®, Santa Clara, California, USA) minimum assay sensitivity was 10 pg/ml, with a intra-assay coefficient of variation of 5.0% and inter-assay CV of 2.3%; Aβ1-40 ELISA kit (Elabscience®, Houston, Texas, USA) minimum assay sensitivity was 9.38 pg/ml, with a intra-assay coefficient of variation of 4.8% and inter-assay CV of 6.5%; and Aβ142 ELISA kit (Elabscience®, Houston, Texas, USA) minimum assay sensitivity was 9.38 pg/ml, with a intra-assay coefficient of variation of 6.0% and inter-assay CV of 6.8%. Determinations were performed in the Clinical Neurosciences Research Laboratory. 4.6. Study variables 4.6.1. Demographic and clinical variables  Age  Gender  Education level: low (< secondary school), medium (completed secondary school), and high (university studies)  BMI: Standing body height was measured with the shoulders in a relaxed position and the arms hanging freely, using a commercial stadiometer. Thus, BMI was calculated as the weight divided by the square of height (kg/m2) YAGO LEIRA FEIJÓO 110 VARIABLES Cases (n=120) Controls (n=157) P-value FMPS (%) 54.8 ±17.5 27.0 ±12.0 <0.001 FMBS (%) 59.4 ±18.2 28.3 ±13.6 <0.001 PPD measures - Mean PP D (mm) 3.6 ±1.0 2.6 ±0.6 <0.001 - Number of sites/mouth PPD ≥4 mm 75.8 ±51.2 17.8 ±32.2 - Number of sites/mouth PPD ≥6 mm 18.1 ±25.3 1.8 ±9.4 <0.001 GR (mm) 0.6 ±0.4 0.3 ±0.3 <0.001 CAL measures - Mean CAL (mm) 4.3 ±1.4 2.9 ±0.9 <0.001 - Number of sites/mouth CAL ≥3 mm 116.3 ±39.4 109.6 ±40.0 0.169 - Number of sites/mouth CAL ≥5 mm 63.7 ±50.1 17.7 ±30.8 <0.001 Number of present teeth 20.0 ±3.7 24.6 ±2.6 <0.001 PISA (mm 2 ) 1040 .4±1145.8 193.2 ±357.1 <0.001 Last dental visit 0.152 - Within the last 12 months, n (%) 71 (59.2) 106 (67.5) - Less often, n (%) 49 (40.8) 51 (32.5) Tooth brush frequency 0.370 - <2 times/day 44 (36.7) 49 (31.2) - ≥2 times7day 76 (63.3) 108 (68.8) Use of interdental care devices, n (%) 9 (7.5) 17 (10.8) 0.409 Table 9. Clinical periodontal parameters and dental variables. Patients with LI who had PD showed significantly higher prevalence of leukoaraiosis and carotid atheromatosis than those without PD. Poor functional outcome at 3 months was found in 32.9% of LI patients with PD in comparison to 8.6% without PD (P=0.006) (Table 10). Clinical study 111 VARIABLES PD (n=85) No PD (n=35) P-value Age (years) 67.5±10.0 63.8 ± 9.5 0.061 Males, n (%) 60 (70.6) 22 (62.9) 0.408 BMI 26.9 ±3.7 27.1 ±2.3 0.734 Hypertension, n (%) 53 (62.4) 21 (60.0) 0.810 Diabet es mellitus, n (%) 24 (28.2) 9 (25.7) 0.779 Hypercholesterolemia, n (%) 36 (42.4) 19 (54.3) 0.233 Ischemic heart disease, n (%) 16 (18.8) 6 (17.1) 0.829 Peripheral arterial disease, n (%) 4 (4.7) 1 (2.9) 0.645 Smoking habit 0.900 - Never smoker, n (% ) 58 (68.2) 24 (68.6) - Former smoker, n (%) 12 (14.1) 4 (11.4) - Current smoker, n (%) 15 (17.6) 7 (20.0) Alcohol consumption, n (%) 18 (21.1) 3 ( 8.6) 0.201 Medication - Statins, n (%) 36 (42.4) 20 (57.1) 0.140 - Antiaggregants, n (%) 30 (35.3) 16 (45.7) 0.286 - Antihypertensives, n (%) 49 (57.6) 21 (60.0) 0.812 Education level 0.894 - High, n (%) 21 (24.7) 9 (25.7) - Medium, n (%) 36 (42.4) 16 (45.7) - Low, n (%) 28 (32.9) 10 (28.6) Leukoaraiosis, n (%) 32 (37.6) 3 (8.6) 0.001 Carotid athe romatosis, n (%) 42 (49.4) 10 (28.6) 0.036 SI s, n (%) 36 (42.4) 11 (31.4) 0.265 mRS at admission 0.5 ±0.6 0.8 ±1.0 0.132 mRS at 3 months 1.6 ±1.2 1.0 ±0.9 0.007 Poor prognosis (mRS > 2) , n (%) 28 (32.9) 3 (8.6) 0.006 LI location 0.799 - Hemispheric, n ( %) 31 (36.5) 14 (40.0) - Basal ganglia, n (%) 43 (50.6) 18 (51.4) - Brainstem, n (%) 9 (10.6) 3 (8.6) - Other locations, n (%) 2 (2.4) 0 (0.0) Table 10. Characteristics of LI patients according to periodontal status. 5.4. Association between periodontal disease and its clinical parameters and the presence of lacunar infarct After adjusting for age, gender, hypertension, diabetes mellitus, hypercholesterolemia, ischemic heart disease, smoking, and statins consumption in multiple logistic regression, among patients with PD, the odds for having LI was 3.3 (95% CI: 1.7-6.4) compared to those without PD (Figure 9). Likewise, severe PD was strongly associated YAGO LEIRA FEIJÓO 112 with the presence of LI (OR=9.8, 95% CI: 2.4-38.9; P<0.001), independently of the same confounding factors. After adjusting for the most relevant clinical periodontal parameters, PISA was mildly associated with LI (OR=1.001, 95% CI: 1.001-1.002; P<0.001) (Figure 10). Figure 9. ORs and 95% CIs of PD calculated for the presence of LI. ORs are presented as adjusted values (blue squares) for age, gender, hypertension, diabetes mellitus, hypercholesterolemia, ischemic heart disease, smoking, and statins consumption. Figure 10. ORs and 95% CIs of clinical periodontal parameters calculated for the presence of LI. ORs are presented as adjusted values (blue squares) for mean PPD, mean CAL, PISA (mm2), number of sites with PPD ≥ 6 mm, and number of sites with CAL ≥ 5 mm. Clinical study 113 5.5. Molecular analysis Patients with LI had significantly higher mean serum IL-6, PTX3, sTWEAK, and Aβ1-40 levels than control subjects. On contrary, IL-10 serum concentrations were significantly lower in patients with LI compared to the control group. No differences were observed regarding Aβ1-42 peptide (Table 11). BIOMARKERS Cases (n=120) Controls (n=120) P-value Systemic inflammation - IL - 6 (pg/mL ) 17.9±6.0 5.4 ±1.1 <0.001 - IL - 10 (pg/mL ) 5.5±1.8 13.4 ±2.4 <0.001 Endothelial dysfunction - PTX3 (pg/mL ) 1907 .2±1295.7 557.0 ±296.3 <0.001 - sTWEAK (pg/mL ) 197.9 ±93.6 27.6 ±17.3 <0.001 A β peptides - A β 1 - 40 (pg/mL ) 53.5 ±11.5 33.0 ±4.7 <0.001 - A β 1 - 42 (pg/mL ) 50.5 ±15.2 48.6 ±9.9 0. 253 Table 11. Serum levels of biomarkers in LI patients and controls. 5.5.1. IL-6 levels in sera according to periodontal status Figure 11 depicts that the presence of PD was associated with higher mean IL-6 concentrations in both cases (21.0 vs. 10.2 pg/mL, P<0.001) and controls (6.6 vs. 4.8 pg/mL, P=0.002). Figure 11. Serum IL-6 levels according to the presence or absence of PD. 0 5 10 15 20 25 Controls LI Cases S e r u m I L - 6 l e v e l s ( p g / m L ) No PD PD P=0.002 P<0.001 4.8 6.6. 21.0. 10.2 n=83 n=37 n=85 n=35 YAGO LEIRA FEIJÓO 114 5.5.2. IL-10 levels in sera according to periodontal status As shown in Figure 12, the presence of PD was associated with lower mean IL-10 levels in the control group (11.9 vs. 14.0 pg/mL, P<0.001). No statistical differences were found in the LI group (5.7 vs. 5.0 pg/mL, P=0.740). Figure 12. Serum IL-10 levels according to the presence or absence of PD. 5.5.3. PTX3 levels in sera according to periodontal status A relationship was observed between the presence of PD and elevated levels of serum PTX3 in both groups. Among control subjects, this difference was modest and not significant. Within the cases, however, there was a highly significant increase in PTX3 levels associated with the presence of PD (2205.6 vs. 1182.6 pg/mL, P<0.001) (Figure 13). 0 2 4 6 8 10 12 14 16 Controls LI Cases Serum IL-10 levels (pg/mL) N o P D PD P<0.001 P=0.740 14.0 11.9. 5.7. 5.0 n=83 n=37 n=85 n=35 Clinical study 115 Figure 13. Serum PTX3 levels according to the presence or absence of PD. 5.5.4. sTWEAK levels in sera according to periodontal status Similarly to IL-6, PD was associated with higher mean sTWEAK levels in both LI cases (240.7 vs. 40.7 pg/mL, P<0.001) and controls (45.6 vs. 19.5 pg/mL, P=0.029) (Figure 14). Figure 14. Serum sTWEAK levels according to the presence or absence of PD. 0 500 1000 1500 2000 2500 Controls LI Cases Serum PTX3 levels (pg/mL) No PD PD P=1.000 P<0.001 610.4 2205.6. 1182.6 n=83 n=37 n=85 n=35 533.2 0 50 100 150 200 250 300 C o n t r o l s L I C a s e s Serum sTWEAK levels (pg/mL) N o P D PD P=0.029 P<0.001 19.5 45.6. 240.7. 90.7 n=83 n=37 n=85 n=35 YAGO LEIRA FEIJÓO 116 5.5.5. Aβ1-40 levels in sera according to periodontal status Figure 15 depicts that PD was associated with raised mean levels of Aβ1-40 in both cases (58.7 vs. 41.0 pg/mL, P<0.001) and controls (36.6 vs. 31.4 pg/mL, P=0.001). Figure 15. Serum Aβ1-40 levels according to the presence or absence of PD. 5.5.6. Aβ1-42 levels in sera according to periodontal status No statistical differences were found in terms of Aβ1-42 serum levels neither in the case group nor in the control group (Figure 16). Figure 16. Serum Aβ1-42 levels according to the presence or absence of PD. 0 10 20 30 40 50 60 70 C o n t r o l s L I C a s e s Serum Aβ1-40 levels (pg/mL) No PD PD P=0.001 P<0.001 31.4 36.6. 58.7. 41.0 n=83 n=37 n=85 n=35 0 10 20 30 40 50 60 70 C o n t r o l s L I C a s e s Serum Aβ1-42 levels (pg/mL) N o P D PD P=1.000 P=1.000 48.7 48.3. 50.0. 51.8 n=83 n=37 n=85 n=35 Clinical study 117 5.6. Association between periodontal disease and periodontal inflamed surface area and elevated serum levels of biomarkers in lacunar infarct patients 5.6.1. IL-6 Mean serum IL-6 levels were significantly and positively associated with greater mean PISA (Model I: R2=0.624, P<0.001). When PD exposure was examined as a categorical variable, multivariable linear regression analysis again indicated that PD in patients with LI was associated with significantly higher serum IL-6 concentrations (Model II: R2=0.656, P<0.001) (Table 12). B SE PC P - value Model I (R 2 =0.624) Age 0.080 0.043 0.177 0.065 Gender - 1.097 0.800 - 0.131 0.173 Hypertension - 0.745 0.830 - 0.086 0.372 Diab etes mellitus 0.483 0.872 0.053 0.581 Hypercholesterolemia 1.367 1.117 0.117 0.224 Ischemic heart disease - 0.580 0.997 - 0.056 0.562 Peripheral arterial disease 0.944 1.833 0.050 0.608 Smoking habit 0.222 0.529 0.040 0.676 Statins - 0.652 1.179 - 0.053 0 .581 Leukoaraiosis - 0.163 0.831 - 0.019 0.845 Carotid atheromatosis 0.006 0.771 0.001 0.994 PISA (mm 2 ) 0.004 0.000 0.791 <0.001 Model II (R 2 =0.656) Age - 0.007 0.041 - 0.017 0.862 Gender - 0.008 0.770 - 0.001 0.992 Hypertension - 0.744 0.795 - 0.090 0 .351 Diabetes mellitus 0.561 0.835 0.065 0.503 Hypercholesterolemia 0.614 1.068 0.055 0.567 Ischemic heart disease 0.764 0.950 0.077 0.423 Peripheral arterial disease - 2.539 1.744 - 0.139 0.148 Smoking habit 0.425 0.507 0.081 0.404 Statins 0.868 1.137 0.074 0.447 Leukoaraiosis 0.091 0.789 0.011 0.908 Carotid atheromatosis - 0.031 0.738 - 0.004 0.966 PD 11.055 0.773 0.810 <0.001 Table 12. Multivariable linear regression models for IL-6 serum levels in LI patients. YAGO LEIRA FEIJÓO 118 5.6.2. PTX3 When examined as a continuous variable (i.e., PISA), multivariable regression analysis indicated that the presence of PD in patients with LI was associated with significantly higher mean PTX3 serum levels (Model I: R2=0.260, P<0.001). In a model that included PD as a categorical variable and other confounding factors, PD was strongly associated with increased serum PTX3 levels (Model II: R2=0.115, P<0.001) (Table 13). B SE PC P - value Model I (R 2 =0.260) Age 21.073 12.856 0.129 0.104 Gender - 255.460 240.594 - 0.084 0.291 Hypertension - 164.096 249.745 - 0.052 0.513 Diabetes mellitus 216.444 262.332 0.065 0.411 Hypercholesterolemia 203.250 335.876 0.048 0.546 Ischemic heart disease 103.360 299.976 0.027 0.731 Peripheral arterial disease - 643.183 551.188 - 0.092 0.246 Smok ing habit 115.599 159.245 0.057 0.469 Statins 82.851 354.539 0.018 0.816 Leukoaraiosis - 457.492 249.888 - 0.144 0.070 Carotid atheromatosis 142.866 231.941 0.049 0.539 PISA (mm 2 ) 0.609 0.096 0.502 <0.001 Model II (R 2 =0.115) Age 10.830 14.076 0.07 4 0.443 Gender - 154.114 264.705 - 0.056 0.562 Hypertension - 149.189 273.188 - 0.053 0.586 Diabetes mellitus 246.879 286.878 0.083 0.391 Hypercholesterolemia 104.292 367.181 0.027 0.777 Ischemic heart disease 293.242 326.614 0.086 0.371 Peripheral arter ial disease - 1105.987 599.569 - 0.176 0.068 Smoking habit 132.839 174.300 0.073 0.448 Statins 200.735 390.968 0.050 0.609 Leukoaraiosis - 288.986 271.232 - 0.102 0.289 Carotid atheromatosis 190.620 253.742 0.072 0.454 PD 1072.218 253.774 0.363 <0.001 Table 13. Multivariable linear regression models for PTX3 serum levels in LI patients. Clinical study 119 5.6.3. sTWEAK Mean sTWEAK levels in serum were significantly and positively associated with greater mean PISA (Model I: R2=0.697, P<0.001). To be older, male as well as a previous history of hypercholesterolemia were also associated with increased serum levels of sTWEAK in LI patients. When PD exposure was examined as a categorical variable, multivariable linear regression analysis again indicated that PD in patients with LI was associated with significantly higher serum sTWEAK concentrations (Model II: R2=0.527, P<0.001) (Table 14). B SE PC P - value Model I (R 2 =0.697) Age 1.263 0.594 0.202 0.036 Gender - 25.424 11.178 - 0.216 0.025 Hypertension 2.868 11.552 0.02 4 0.804 Diabetes mellitus 2.420 12.182 0.019 0.843 Hypercholesterolemia 32.471 15.521 0.199 0.039 Ischemic heart disease - 10.296 13.884 - 0.072 0.460 Peripheral arterial disease 20.233 25.470 0.077 0.429 Smoking habit 8.280 7.359 0.109 0.263 Statins - 18.432 16.421 - 0.108 0.264 Leukoaraiosis 10.351 11.554 0.087 0.372 Carotid atheromatosis - 18.675 10.717 - 0.167 0.084 PISA (mm 2 ) 0.068 0.004 0.830 <0.001 Model II (R 2 =0.527) Age 0.007 0.744 0.001 0.993 Gender - 11.654 14.044 - 0.053 0.409 Hyperten sion 3.348 14.445 0.015 0.817 Diabetes mellitus 5.426 15.227 0.023 0.722 Hypercholesterolemia 20.759 19.396 0.068 0.287 Ischemic heart disease 11.515 17.286 0.042 0.507 Peripheral arterial disease - 33.204 31.666 - 0.066 0.297 Smoking habit 10.965 9.207 0.075 0.236 Statins - 0.198 20.685 - 0.001 0.992 Leukoaraiosis 21.767 14.330 0.096 0.132 Carotid atheromatosis - 16.336 13.401 - 0.077 0.226 PD 149.082 14.099 0.670 <0.001 Table 14. Multivariable linear regression models for sTWEAK serum levels in LI patients. YAGO LEIRA FEIJÓO 126 5.7.5. Correlation between periodontal inflamed surface area and significant biomarkers in patients with poor outcome A positive and strong correlation was found between PISA and IL6 (r=0.738, P<0.001; Figure 20), PTX3 (r=0.468, P=0.008; Figure 21), sTWEAK (r=0.771, P<0.001; Figure 22), and Aβ1-40 (r=0.745, P<0.001; Figure 23) in patients with poor prognosis. Figure 20. Correlation between PISA and IL-6 in patients with poor functional outcome. Figure 21. Correlation between PISA and PTX3 in patients with poor functional outcome. Clinical study 127 Figure 22. Correlation between PISA and sTWEAK in patients with poor functional outcome. Figure 23. Correlation between PISA and Aβ1-40 in patients with poor functional outcome. EXPERIMENTAL STUDY: complementary investigation Experimental study 131 EXPERIMENTAL STUDY: complementary investigation 1. JUSTIFICATION Results from the clinical study showed that PD is prevalent in LI patients and when present it contributes to elevated serum levels of proinflammatory mediators and markers of endothelial dysfunction. Although this association was independent of other contributing factors, some conditions (i.e., ageing, hypertension or diabetes) as well as asymptomatic CSVD subtypes (i.e., leukoaraiosis) are also of great importance in the pathophysiology of LI. In order to investigate whether PD could be one of the main contributors to a low-grade chronic systemic inflammation state leading to endothelial dysfunction in LI, it makes reasonable to carry out a complementary experimental study using a PD-induced model in rats. 2. HYPOTHESIS Based on the results obtained in the clinical study, we hypothesized that experimental PD in systemically healthy rats may evoke a mild acute inflammation effect (i.e., significant increase of IL-6 and PTX3 within the first week after PD induction) followed by dysfunction of the endothelium (i.e., significant increase of PTX3, sTWEAK and Aβ1-40), thus, demonstrating that PD may be a significant contributor to an increased systemic pro-inflammatory state and endothelial dysfunction leading to a higher risk for developing LI. 3. OBJECTIVE  To investigate in systemically healthy rats whether PD is associated with an enhanced inflammatory response and YAGO LEIRA FEIJÓO 132 vascular endothelial impairment measured by the same biomarkers analysed in the clinical study that were linked with an increased risk of LI. 4. MATERIAL AND METHODS 4.1. Experimental periodontitis model 4.1.1. Porphyromonas gingivalis lipopolysaccharide-induced periodontitis Pg has been identified as a major aetiological factor in the pathogenesis of PD (van Winkelhoff et al. 2002). The LPS component of the cell wall of this bacterium and other gram-negative microorganisms is a significant inflammatory stimulus that triggers an innate immune response. Indeed, is responsible for the recruitment of PMNs, oedema, and vascular dilation in inflammatory periodontal tissue (Page and Schroeder 1981) as well as cytokines secretion, lytic enzyme production, and osteoclast activation (Jiang et al. 2002). Therefore, this model produces a histopathological aspect similar to that observed in established human PD, characterized by increased infiltration of leukocytes, higher levels of pro-inflammatory cytokines, collagen degradation, and alveolar bone loss (Graves et al. 2012). There are other models of PD induction such as ligature and oral gavage models (Graves et al. 2008). Ligature model is indicated for situations in which an acute development of PD is required, as the thread around the tooth promotes rapid and severe bone destruction (de Molon et al. 2013). On the other hand, the oral gavage model is considered to be a chronic model of PD owing to the longer period of time required to evoke bone loss in comparison to other models (de Molon et al. 2013). The LPS model is considered to be a direct and easy method for the induction of controlled experimental PD because bone loss is localized, the stimulus is constant, and the alveolar bone destruction usually occurs within 7 days after the start of the bacterial LPS injections (Dumitrescu et al. 2004). This model is useful in Experimental study 133 investigating the host-bacteria interaction and the activation of signalling pathways, in the studying of several pro-inflammatory and vascular function mediators, and in testing the performance of a specific cell or molecule in the pathogenesis process. However, the main limitation of the LPS model is the lack of bacterial colonization and the need for constant injections throughout the experiment (Table 20). Hence, according to the purpose of the experimental study, it seems reasonable to use the Pg-LPS model for PD induction. Model Advantages Disadvantages Ligature  Acute model for PD  Rapid and severe bone destruction  Bone loss occurs predictably over a period of 7 days  Periodontal destruction may be aggravated by mechanical lesion  Host response decreases periodontal destruction  No significant gingival inflammation or bone loss in germ-free rats LPS injection  Accurate time of infection  More direct and controlled PD model  Greater experimental control over the pathogenic stimulus  Lack of bacterial colonization  Need for constant injections throughout the experiment Oral gavage  Chronic mo del for PD  Easier and simple induction of PD  Establish a relationship between induction of PD and systemic conditions  Examines the impact of various components of the host responses  Requirement of longer periods of time to produce bone loss  Host response decreases periodontal destruction Table 20. Summary of advantages and disadvantages of the different rodent models of PD induction. Adapted from Graves et al. (2013). YAGO LEIRA FEIJÓO 134 4.1.2. Experimental design An experimental study was carried out in the Clinical Neurosciences Research Laboratory of the University Clinical Hospital of Santiago de Compostela. The experimental protocol was approved by the Research Commission of the University Clinical Hospital of Santiago de Compostela. All experimental procedures were performed according to the Animal Care Committee European Union rules and the Spanish regulation (86/609/CEE, 2003/65/CE, 2010/63/EU, RD1201/2005 and RD53/2013). The Animal Research Reporting of In Vivo Experiments (ARRIVE) guidelines were followed in this experiment (Kilkenny et al. 2010). Prior to PD induction, MRI and micro-CT (μCT) examination along with blood extraction (i.e., baseline) were carried out in all animals. In order to induce PD, Pg-LPS injections were performed on Mondays, Wednesdays, and Fridays during two weeks. Another MRI analysis together with a visual gingival analysis with a microscope was performed immediately before the second round of injections. After the last couple of injections, MRI and μCT scan were done to confirm periodontal tissue inflammation and alveolar bone loss. A set of blood extractions was carried out 1 day, 1, 2, and 3 weeks after the last couple of injections. Once all serum samples were stored, determinations of several biomarkers were done (Figure 24). Experimental study 135 Figure 24. Experiment outline. 4.1.3. Animals and Anaesthesia Six male Sprague-Dawley rats of 7 weeks of age and weighted between 300 and 350 g were used. Animals were housed individually, in stable environmental conditions (environmental temperature of 23ºC), relative humidity of 40% and a light-dark cycle of 12h, as well as free access to food and water. Each animal was initially placed into an induction chamber attached to a sevoflurane anaesthetic vaporizer and anaesthesia was induced with 6% sevoflurane in a gas of 70% NO2 and 30% O2, followed by the application of a nose cone with 4% sevoflurane in the same proportion of the aforementioned gases to maintain anaesthesia during the experimental procedures. During surgery, all animals were subjected to temperature control, maintaining temperature at 37±0.5ºC by a thermostat-controlled electric pad (NeoBiotect, Spain) (Figure 25). YAGO LEIRA FEIJÓO 142 interproximal contact point of the first and second molars crown, the distance between de cemento-enamel junction (CEJ) and the alveolar bone crest were measured for the distal surface of the first molar and the mesial surface of the second molar just below the contact point and 0.2 mm palatal to the contact point (Hiyari et al. 2015) (Figure 32). All μCT analyses were performed by a single biologist (NG-L). Figure 32. The distance from the CEJ to the alveolar crest was measured at the sagittal plane intersecting the interproximal molars. Red lines depict the measurement that was taken for distal of first molar and mesial of second molar. 4.4. Serum collection and laboratory tests Prior to PD induction (baseline) and 24 h, 7, 14 and 21 days after the last couple of injections, 1800 μL of venous blood were collected from the tail of each animal by venepuncture using a 22-gauge needle with a 1 mL syringe (Figure 33). Blood samples were allowed to clot at room temperature and after 1 hour, serum was separated from blood by centrifugation (7 min. at 3000 g) and 700 μL of extracted serum was immediately transferred to 1.5 mL aliquots. Each aliquot was stored at –80ºC until the time of analysis. Serum levels of all biomarkers were measured by ELISA technique following manufacturer instructions. IL6 ELISA kit (Picokine™, Boster Biological Technology, Pleasanton, California, USA) minimum assay sensitivity was 5.0 pg/ml with a intraassay CV of 1.7%; IL-10 ELISA kit (Picokine™, Boster Biological Technology, Pleasanton, California, USA) minimum assay sensitivity was 4.0 pg/ml with a intra-assay CV of 7.4%; PTX3 ELISA kit (Fine Experimental study 143 Test, Wuhan Fine Biotech, Wuhan, China) minimum assay sensitivity was 0.094 ng/ml, with a intra-assay CV of 1.5%; sTWEAK ELISA kit (Fine Test, Wuhan Fine Biotech, Wuhan, China) minimum assay sensitivity was 9.375 pg/ml, with a intra-assay CV of 4.4%; Aβ1-40 ELISA kit (Fine Test, Wuhan Fine Biotech, Wuhan, China) minimum assay sensitivity was 46.875 pg/ml, with a intra-assay CV of 3.2%; and Aβ1-42 ELISA kit (Fine Test, Wuhan Fine Biotech, Wuhan, China) minimum assay sensitivity was 9.375 pg/ml, with a intra-assay CV of 2.5%. Determinations were performed in the Clinical Neurosciences Research Laboratory. Figure 33. Blood extraction from vein tail. 4.5. Euthanasia After the last blood extraction and with the animal fully anaesthetised, sacrifice was performed by 2 mL intracardiac injection of potassium chloride. YAGO LEIRA FEIJÓO 144 4.6. Statistical analysis All data analyses were performed with IBM SPSS Statistics 20.0 software for Mac (SPSS Inc., Chicago, IL, USA). Mean and standard deviation was calculated for continuous variables, after the method of Shapiro-Wilk was applied to confirm that the data were sampled from a normal distribution. Paired t test and analysis of variance for repeated measures were used to compare differences over time. Additionally, post hoc comparisons were carried out using Bonferroni corrections. All tests were performed at a significance level of α = 0.05. 5. RESULTS 5.1. Periodontal inflammation Gingival inflammation was evident at day 7 of the experiment affecting the palatal side of the first and second molar region in both sides of the upper jaw (Figure 34). Figure 34. a: Gingival inflammation at 7 days (first upper left molar). b: Gingival inflammation at 7 days (first upper right molar). Experimental study 145 Periodontal inflammation was confirmed by means of MRI analysis. Differences in relative T1 and T2 signal intensities were found at 7 and 14 days after periodontal induction onset compared to baseline (Figure 35 and 36). When relative T2-w signal intensity was analysed in the first and second upper right molar, statistical differences were observed between 7 and 14 days (P<0.05). Figure 35. Relative T1-w signal intensity. 1M (right) 1M (left) 2M (right) 2M (left) Time of assessment (days) Time of assessment (days) Time of assessment (days) Time of assessment (days) Relative T1 signal intesity (%) Relative T1 signal intesity (%) Relative T1 signal intesity (%) Relative T1 signal intesity (%) YAGO LEIRA FEIJÓO 146 Figure 36. Relative T2-w signal intensity. *7 vs. 14 days, P<0.05. In terms of palatal thickness, which we considered as a surrogate measure of oedema, an increase in the relative percentage was found due to periodontal inflammation. In fact, statistically significant differences were observed for this measure in the second upper right molar between 7 days of periodontal induction and baseline (Figure 37). 1M (right) 1M (left) 2M (right) 2M (left) Time of assessment (days) Time of assessment (days) Time of assessment (days) Time of assessment (days) Relative T2 signal intesity (%) Relative T2 signal intesity (%) Relative T2 signal intesity (%) Relative T2 signal intesity (%) * * Experimental study 147 Figure 37. Relative palatal thickness (%) from T1-w. *0 vs. 7 days, P<0.01. 5.2. Alveolar bone loss The μCT analysis revealed statistical significant alveolar bone loss at the interproximal space after periodontal induction (i.e., 14 days) between the first and second maxillary molars bilaterally at the LPSinjected sites compared to baseline (Figure 38). Indeed, the distance between the CEJ and the bone crest was significantly greater at 14 days in both sides of the upper jaw compared to baseline measurements (Figure 39 and 40). 1M (right) 1M (left) 2M (right) 2M (left) Time of assessment (days) Time of assessment (days) Time of assessment (days) Time of assessment (days) Relative palatal thickness (%) from T1-w * Relative palatal thickness (%) from T1-w Relative palatal thickness (%) from T1-w Relative palatal thickness (%) from T1-w YAGO LEIRA FEIJÓO 148 Figure 38. a: μCT image of upper molars at baseline. b: μCT image of upper molars at 14 days. Figure 39. Alveolar bone loss in the Figure 40. Alveolar bone loss in the upper right side. upper left side. a b 0 1 2 3 D i s t a n c e C E J - b o n e c r e s t ( m m ) B a s e l i n e 14 days P=0.001 5.7 2.0. 1.5 0 1 2 Distance CEJ-bone crest (mm) B a s e l i n e 14 days P<0.001 5.7 1.8. 1.3 Experimental study 149 5.3. Biomarkers 5.3.1. Systemic inflammation A sharp increase was observed for IL-6 24 h after periodontal induction, which confirms its nature as an early inflammatory mediator. The levels of IL-6 decreased during the following 3 weeks but still were significantly higher compared to baseline (Figure 41). Figure 41. Changes in serum levels of IL-6 (pg/mL) at 24h, 7, 14, 21 days after periodontal induction. #P<0.001 compared to baseline, *P<0.05 compared to baseline. On contrary, PD evoked a significant decrease in IL-10 serum levels at 24 h after the last LPS-injection, and the levels of this anti-inflammatory mediator continued to reduce up to 21 days (Figure 42). Figure 42. Changes in serum levels of IL-10 (pg/mL) at 24h, 7, 14, 21 days after periodontal induction. #P<0.001 compared to baseline, *P<0.05 compared to baseline. YAGO LEIRA FEIJÓO 150 5.3.2. Endothelial dysfunction A mild acute significant increase in PTX3 levels at 24h following periodontal induction was observed, which was more pronounced as time was drawn on (Figure 43). Levels of sTWEAK were significantly elevated 1 week after the last LPS-injection compared to baseline and continued to increase in the following weeks (Figure 44). Figure 43. Changes in serum levels of PTX3 (ng/mL) at 24h, 7, 14, 21 days after periodontal induction. #P<0.001 compared to baseline, *P<0.05 compared to baseline. Figure 44. Changes in serum levels of sTWEAK (pg/mL) at 24h, 7, 14, 21 days after periodontal induction. #P<0.001 compared to baseline, *P<0.05 compared to baseline. Experimental study 151 5.3.3. Aβ peptides Similarly to PTX3, PD evoked a slight acute elevation of Aβ1-40 serum levels that reached statistical significance compared to baseline and was maintained during the following two weeks. However, at day 21, a reduction in the levels of this biomarker was observed (Figure 45). Following experimental PD, Aβ1-42 peptide levels peak much latter than the other endothelial dysfunction biomarkers (i.e., 21 days), confirming its nature of cognitive decline molecule (Figure 46). Figure 45. Changes in serum levels of Aβ1-40 (pg/mL) at 24h, 7, 14, 21 days after periodontal induction. #P<0.001 compared to baseline, *P<0.05 compared to baseline. Figure 46. Changes in serum levels of Aβ1-42 (pg/mL) at 24h, 7, 14, 21 days after periodontal induction. #P<0.001 compared to baseline, *P<0.05 compared to baseline.