京都大学 / 0048 / 新制・課程博士 / 博士(医学) / 甲第19403号 / 医博第4054号 / 新制||医||1012(附属図書館) / 32428 / 京都大学大学院医学研究科医学専攻 / (主査)教授 前川 平, 教授 平岡 眞寛, 教授 羽賀 博典 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

Abstract Lymphomas are a heterogeneous group of malignancies that arise from lymphatic tissues. Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoma sub-type. It is an aggressive malignancy with an increasing incidence. The prognosis of DLBCL has improved significantly, but problems also remain. The clinical significance of central nervous system (CNS) relapses has become increasingly important. As secondary CNSL (sCNSL) and primary CNS lymphoma (PCNSL) are known to have poor prognoses; the prevention of sCNSL is of crucial importance. Peripheral T-cell lymphomas (PTCL) are rare neoplasms and include several lymphoma subtypes that possess complex and also overlapping morphological and immunophenotypic characteristics. The identification of different entities has improved, but the biological knowledge remains scarce when compared to DLBCL. The optimal treatment schemas for PTCLs are still lacking and they have long been treated with the same therapies as B-cell lymphomas, mainly with suboptimal treatment results. The aim of this study was to identify poor prognostic markers in DLBCL and PTCLs and potential biological markers for the prediction of DLBCL CNS relapse. The study material included patients with systemic DLBCL without CNS affision (sDLBCL), sCNSL, PCNSL and PTCLs. The expression of epithelial-mesenchymal transition (EMT) transcription factors (TFs), chemokines and their receptors and adhesion-, migration- and inflammatory responses-associated molecules were studied by means of immunohistochemistry. IEM was used to verify the specific subcellular location of the studied molecules. GEP was performed on 12 PTCL samples in order to compare the poor prognosis group with the good prognosis group and on one sDLBCL and one sCNSL sample from the time of primary diagnosis. The EMT TFs were found to be expressed in both DLBCL and PTCLs, where they ultimately proved to have prognostic relevance as well. In PTCLs, these TFs were able to delineate a disease group with a specific gene-expression profile. CXCR4, CXCR5, ITGA10, PTEN and CD44 were found to be differently expressed between DLBCL cases with CNS affision when compared to those without CNS disease. These molecules seem to play a role in the development of CNS relapse and hopefully, if further verified, will lead towards the identification of biological markers for CNS relapse prediction. / Tiivistelmä Lymfoomat ovat heterogeeninen ryhmä imukudossyöpiä, joista diffuusi suurisoluinen B-solulymfooma (DLBCL) on yleisin alatyyppi. Se on aggressiivinen maligniteetti, jonka insidenssi on noussut viime vuosina. DLBCL potilaiden ennuste on parantunut merkittävästi, mutta yhä osa potilaista menehtyy tautiinsa. DLBCL:n keskushermostorelapsin kliininen merkitys on tänä päivänä aiempaa suurempi. Sekundaarisen keskushermostolymfooman (sCNSL) ja primaarin aivolymfooman (PCNSL) ennusteet ovat nykyhoidoilla huonoja, joten keskushermostorelapsin ennaltaehkäiseminen on tärkeää. Perifeeriset T-solulymfoomat (PTCLs) ovat ryhmä harvinaisia neoplasioita, joka sisältää useita eri alatyyppejä, joiden morfologiset ja immunofenotyyppiset ominaisuudet ovat monimuotoisia ja osin päällekkäisiä. Eri entiteettien indentifiointi on parantunut, mutta PTCL:ien biologinen tietämys on yhä DLBCL:aa heikompaa. PTCL:ien optimaalinen hoito ei ole selvillä ja tätä tautiryhmää on pitkään hoidettu samoilla hoidoilla kuin DLBCL:aa, mutta huonommilla hoitotuloksilla. Tutkimuksen tavoitteena oli löytää huonon ennusteen markkereita, joilla myös pystyttäisiin ennustamaan DLBCL:n keskushermostorelapsia. Aineisto koostui DLBCL, sCNSL, PCNSL ja PTCL näytteistä. Immunohistokemiallisilla värjäyksillä tutkittiin epiteliaalis-mesenkymaalisen transition (EMT) transkriptiotekijöitä (TF), kemokiinireseptoreita sekä adheesioon-, migraatioon ja inflammaatioon assosioituja molekyylejä. Immunoelektronimikroskopialla varmennettiin molekyylien lokalisaatio soluissa. Geeniekspressioprofiloinnilla (GEP) verrattiin kahdentoista hyvän ja huonon ennusteen ryhmään kuuluvan PTCL näytteen välisiä geeniekspressioeroja sekä kahden DLBCL potilaan näytteitä, joista toiselle kehittyi keskushermostorelapsi. EMT TF:ien ekspressiota nähtiin DLBCL ja PTCL näytteissä, joissa niillä myös todettiin olevan ennusteellista merkitystä. PTCL:ssa TF:t pystyivät erottelemaan tautiryhmän, jolla oli oma spesifinen geeniekspressioprofiilinsa. CXCR4, CXCR5, ITGA10, PTEN ja CD44 ekspressio oli erilaista systeemisissä DLBCL tapauksissa verrattuna sCNSL tapauksiin. Edellä mainituilla molekyyleillä näyttää olevan oma roolinsa keskushermostotaudin kehittymisessä ja jos nämä tulokset pystytään vahvistamaan tulevissa tutkimuksissa, johtavat ne toivottavasti kohti keskushermostorelapsiriskin tarkempaa tunnistamista.

High grade aggressive lymphomas have high mortality. By their nature, more than 40% of patients die from these diseases even with the improved treatment strategies currently available for oncology patients. The characteristic feature is that they are functionally heterogeneous and therefore have different biological and molecular signatures which make it difficult for all groups to respond to same line of treatment. Based on the above, I set out to look at the impact of viral and cellular gene products on these groups of diseases: In chapter 3 I developed monoclonal antibodies against HERV‐K10. I subsequently investigated their expressions in aggressive lymphomas including Diffuse Large B‐cell lymphoma, Hodgkin’s lymphoma and Primary CNS lymphomas. I showed HERV‐K10 is expressed in cell lines of aggressive lymphomas, but not in paraffin‐embedded tissues. In chapter 4 I showed that the expression of ATM using immune‐histochemistry techniques in aggressive lymphomas does offer a guide to prognosis and treatment. Nearly 30% of Diffuse Large B‐cell lymphomas express ATM, 55% of Hodgkin’s lymphomas and more than 80% of Primary CNS lymphomas. I also showed there is a correlation of ATM expression and EBV‐driven aggressive lymphomas and that this has a poor prognostic significance. Chapter 5 analysed the results obtained by generating, validating and evaluating data base of DLBCL and PCNSL from a retrospective cohort over a 17‐year period. The results confirmed that prognostic indicators including ATM, S1PR2, Autotaxin and EBV using immuno‐histochemistry techniques help with categorising aggressive lymphomas into different prognostic groups and does influence future management. In summary, my results showed there is a critical place for immuno‐histochemistry techniques in convincingly helping understand the expressions of viral and cellular gene products in aggressive lymphomas and in contributing positively to their management.