Vpliv označevalcev CD9 in TRIM28 na preživetje bolnikov z glioblastomom

Porčnik, Andrej (author), Velnar, Tomaž (mentor), Lah Turnšek, Tamara (comentor)

Ozadje: Glioblastom (GB) je eden izmed najmalignejših možganskih tumorjev s povprečnim preživetjem bolnikov zgolj 12–15 mesecev. Glavni razlogi za njegovo visoko malignost in slabo prognozo so invazivnost, heterogenost in odpornost glioblastomskih matičnih celic (GMC) na zdravljenje. Opredelitev bioloških označevalcev GMC in natančnejša razdelitev GB v podvrste bi lahko doprinesli k boljšemu poznavanju bolezni in učinkovitejšemu zdravljenju bolnikov. Kot možna označevalca GMC sta bila v nedavnih raziskavah prepoznana tetraspanin CD9 in transkripcijski dejavnik TRIM28. Metode: V prospektivno raziskavo smo vključili 89 bolnikov z GB, operiranih od leta 2012 do januarja 2020 na Kliničnem oddelku za nevrokirurgijo Univerzitetnega kliničnega centra Ljubljana. Opredelili smo histopatološke in molekularne značilnosti tumorjev, značilnosti zdravljenja in preživetje bolnikov. Iz vzorcev tumorja smo izolirali ribonukleinsko kislino in z verižno reakcijo s polimerazo na mikrofluidnem čipu analizirali nivoje izražanja 15 izbranih genov za opredelitev podvrst GB. S tako prilagojeno klasifikacijo po Behnanu smo vzorce GB razvrstili v klasično (CL), mezenhimsko (MES) in pronevralno (PN) podvrsto. Preostale vzorce GB, ki kriterijev za uvrstitev v omenjene podvrste niso izpolnjevali, pa smo uvrstili v mešano (MIX) podvrsto. V GB podvrstah smo določili nivo izražanja genov CD9 in TRIM28 ter preverili njun vpliv na preživetje bolnikov. Vzpostavili smo tudi celične linije tumorskih celic, ki so bile pridobljene iz biopsij tumorjev. Preverili smo vpliv ciljanja označevalcev CD9 in TRIM28 na rast človeških GMC in GB celic in vivo na modelu zarodkov rib cebric. Označevalec TRIM28 smo zavirali z nanoprotitelesom NB237, gen CD9 pa smo utišali z lentivirusnim vektorjem. Z invertnim fluorescentnim mikroskopom smo prvi in tretji dan po vsaditvi celic v možgane zarodkov rib cebric opredelili velikost in fluorescenco tumorja ter tako in vivo določili invazijo in proliferacijo GB celic. Pri preskušanju domnev smo vrednost p < 0,05 privzeli kot statistično značilno. Rezultati: Povprečna starost bolnikov ob operaciji je bila 62 ± 13 let, 65,2 % je bilo moških. Tumor se je pogosteje nahajal v desni možganski hemisferi (56,2 %), v večini primerov je bil omejen na en reženj (82,2 %). S prilagojeno klasifikacijo po Behnanu smo GB razdelili v CL (n = 21; 23,6 %), MES (n = 5; 5,6 %) in PN (n = 7; 7,9 %) podvrsto. 62,9 % vseh GB (n = 56) ni bilo moč uvrstiti v omenjene tri podvrste, zato smo jih uvrstili v MIX podvrsto. Izražanje označevalca CD9 je bilo statistično značilno povečano v CL podvrsti v primerjavi z MES (p = 0,003) in MIX podvrsto (p < 0,001), izražanje v PN podvrsti pa v primerjavi z vsemi preostalimi ni bilo pomembno različno (vsi p > 0,05). Primerjava izražanja označevalca TRIM28 v različnih podvrstah ni pokazala statistično značilnih razlik (vsi p > 0,05). Pri vseh bolnikih je bila opravljena vsaj delna resekcija tumorja, pri 42,7 % bolnikov je bila resekcija popolna. Celokupno preživetje bolnikov z GB je bilo 14,1 mesecev (95 % interval zaupanja 11,1–17,0 mesecev). Kot neodvisna napovedna dejavnika preživetja sta se izkazala obseg resekcije tumorja (nepopolna vs. popolna odstranitev: razmerje ogroženosti 2,42; 95 % interval zaupanja 1,40–4,20) in pooperativna ocena zmogljivosti bolnika po Karnofskem (? 70 vs. < 70: razmerje ogroženosti 0,50; 95 % interval zaupanja 0,27–0,95). Stopnja izražanja označevalcev CD9 in TRIM28 ni imela vpliva na preživetje bolnikov z GB (vsi p > 0,05). Preživetje bolnikov s CL podvrsto je bilo krajše kot preživetje bolnikov z MIX podvrsto (p = 0,013), med preostalimi podvrstami pa v preživetju ni bilo značilnih razlik. In vivo v možganih zarodkov rib cebric se je po aplikaciji nanoprotitelesa TRIM28 invazija GB celic NB237 zmanjšala za 13–15 %, invazija GMC pa za 25 % (vsi p < 0,05). Invazija GMC se je v skupini s z utišanim genov CD9 zmanjšala za 20 % (p < 0,001), invazija GB celic U87 pa za 24 % (p < 0,05). Sklepi: Prilagojena klasifikacija po Behnanu, ki smo jo v raziskavi uporabili za razvrstitev GB v različne podvrste, se morda ni izkazala za optimalno, saj smo velik del tumorjev uvrstili v MIX podvrsto, kar lahko kaže na heterogenost GB. Izražanje označevalca CD9 je bilo statistično značilno povečano v CL podvrsti v primerjavi z MES in MIX podvrsto, izražanje označevalca TRIM28 pa se med podvrstami ni pomembno razlikovalo. Stopnja izražanja označevalcev CD9 in TRIM28 ni imela vpliva na preživetje bolnikov z GB. Ugotovili pa smo, da imata tako CD9 kot TRIM28 pomembno vlogo pri invaziji GB celic in GMC, ki smo jo z utišanjem gena ali ciljanjem s selektivnimi nanotelesi zmanjšali. Z raziskavo smo validirali zaviralno nanoprotitelo proti TRIM28, kar odpira vrata novim možnostim zdravljenja. Ugotovili smo, da je s kirurškega vidika predvsem pomembno stremeti k popolni resekciji GB, saj se je le-ta izkazala kot neodvisen napovedni dejavnik boljšega preživetja. Kljub vsem trenutnim možnostim zdravljenja pa ostaja GB bolezen z zelo slabo prognozo. Njegova nizka pojavnost ter velika inter- in intratumorska heterogenost, ki smo jo prikazali tudi z našimi rezultati, otežujeta njegovo raziskovanje. Med raziskavo smo v slovenskem prostoru vzpostavili biobanko GB poimenovano Gliobanka, kar nam bo omogočilo izvedbo raziskav na večjem številu vzorcev in učinkovitejše proučevanje bolezni.

• Language: Slovenian
• Keywords: biološka označevalca TRIM28 in CD9, glioblastom in njegove podvrste, glioblastomske matične celice, invazija tumorksih celic, ribe cebrice
• Work type: Doctoral dissertation
• Organization: MF - Faculty of Medicine
• Year: 2024
• PID: 20.500.12556/RUL-154203
• Publication date in RUL: 01.02.2024

Secondary language

• Language: English
• Title: The impact of biomarkers CD9 and TRIM28 on overall survival of patients with glioblastoma

Abstract

Introduction: Glioblastoma (GB) is one of the most malignant brain tumors with an average overall patient survival of 12–15 months. The main reasons for its high malignancy and poor prognosis are the invasiveness, heterogeneity and resistance of glioblastoma stem cells (GSC) to treatment. The definition of GSC biomarkers and a more accurate classification of GB into subtypes could contribute to a better understanding of the disease and more efficient treatment of patients. The tetraspanin CD9 and the transcription factor TRIM28 have been suggested as possible markers of GSC in recent studies. Methods: In a prospective study, we included 89 patients with GB who were operated from 2012 to January 2020 at the Clinical department of neurosurgery at the University medical centre Ljubljana. We determined the histopathological and molecular characteristics of the tumors and patient survival. To define of GB subtypes, we isolated ribonucleic acid from tumor samples and analyzed the expression levels of 15 selected genes using polymerase chain reaction on a microfluidic chip. With such application of adapted of Behnan classification, GB samples were classified into classical (CL), mesenchymal (MES) and proneural (PN) subtypes. The remaining GB samples, which did not meet the criteria for inclusion in the aforementioned substypes, were classified as the mixed (MIX) subtype. We determined the expression level of CD9 and TRIM28 genes in GB subtypes and analysed their effect on patient survival. We also established cell lines from tumor cells obtained from tumor biopsies. We examined the effect of CD9 and TRIM28 markers targeting on the growth of human GSC and GB cells in vivo in the zebrafish embryo model. The TRIM28 protein was inhibited with its selective nanobody NB237, and the CD9 gene was silenced with a lentiviral vector. On the first and third day after the implantation of cells into the brain of zebrafish embryos, we determined the size and fluorescence of the tumor using the fluorescence microsopy, thus in vivo determining the invasion and proliferation of GB cells. When testing hypotheses, we assumed a value of p < 0.05 as statistically significant. Results: The average age of the patients was 62 ± 13 years, 65.2% were men. The tumor was more often located in the right cerebral hemisphere (56.2%) and in most cases it was limited to one lobe (82.2%). Using the adapted Behnan classification, GB were divided into CL (n = 21; 23.6%), MES (n = 5; 5.6%) and PN (n = 7; 7.9%) subtypes. 62.9% of all tumors (n = 56) could not be classified into the mentioned GB subtypes, so we classified them as the MIX subtype. CD9 marker expression was statistically significantly increased in CL subtype compared to MES (p = 0.003) and to MIX subtype (p < 0.001); expression in PN subtype was not significantly different compared to the other subtypes (all p > 0.05). Comparison of TRIM28 marker expression in different subtypes showed no statistically significant differences (all p > 0.05). The tumor resection was complete In 42.7% of patients. The average survival of patients with GB was 14.1 months (95% confidence interval 11.1–17.0 months). The extent of tumor resection (incomplete vs. complete resection: hazard ratio 2.42; 95% confidence interval 1.40–4.20) and postoperative Karnofsky performance score (⡥ 70 vs. < 70: hazard ratio 0.50; 95% confidence interval 0.27–0.95) were found to be independent predictors of patient survival. The expression level of CD9 and TRIM28 markers had no effect on the survival of patients with GB (all p > 0.05). The survival of patients with CL subtype was shorter than that of patients with MIX subtype (p = 0.013); there were no significant differences in survival between the remaining subtypes. In vivo in the brain of zebrafish embryos, GB cell invasion was reduced by 13–15% and GSC invasion by 25% (all p < 0.05) after addition of TRIM28 nanobody NB237. GSC invasion was reduced by 20% (p < 0.001) and U87 GB cell invasion by 24% (p < 0.05) in the group of cells silenced for CD9 gene. Conclusion: The adapted Behnan classification, which we used in the study to classify GB into different subtypes, did not prove to be optimal, since we had to classify a large part of the tumors as the MIX subtype indicating GB heterogeneity. CD9 marker expression was statistically significantly increased in the CL subtype compared to MES and MIX subtypes, while TRIM28 marker expression was not significantly different among subtypes. The expression level of CD9 and TRIM28 markers had no influence on the survival of patients with GB. However, we found that both CD9 and TRIM28 play an important role in the invasion of GB cells and GSC, which we managed to reduce by lowering their expression by gene silecing and using selective nanobodies respectively. We have validated the selective nanobody against TRIM28, which opens the door to new treatment options. Complete GB resection should be achieved, as we have proved that it was an independent predictor of better patient survival. Despite all current treatment options, GB remains a disease with a very poor prognosis. Its low incidence and great inter- and intratumoral heterogeneity, which we also demonstrated with our results, make its research difficult. During the study, we established the GB biobank in Slovenia, which will enable us to conduct GB research on a larger number of samples and to study the disease more efficiently.

• Keywords: biomarkers TRIM28 and CD9, glioblastoma and subtypes, glioblastoma stem cells, invasion of tumour cells, zebra fish