A Thesis submitted to The University of Arizona College of Medicine - Phoenix in partial fulfillment of the requirements for the Degree of Doctor of Medicine. / Approximately 25% of patients with locoregional esophageal adenocarcinoma (EC) are resistant (marked by minimal tumor regression; TRG 3) to preoperative chemoradiation, including 5FU‐based and CROSS regimens. Previously, an immunohistochemistry (IHC) test that accurately identifies patients as responders (TRG 0‐2) or non‐responders (TRG 3) to neoadjuvant CTRT was developed and validated. The current study was designed to identify gene expression profile (GEP) signatures able to predict response to preoperative treatment. Methods: Formalin‐fixed, paraffin‐embedded (FFPE) tumor tissue from 24 diagnostic biopsies (14 responders, 10 non‐responders) was collected. RNA was isolated, and RT‐PCR performed to assess the expression of 96 candidate genes chosen from in silicoanalysis. Genetic signatures incorporating genes with significant expression differences in pathologically determined responders versus non‐responders were identified, and linear and non‐linear predictive modeling methods were used to assess the accuracy of the signatures for predicting treatment response. Cross validation was performed to attain corrected accuracy values. Ten‐, 18‐, and 24‐gene signatures were identified with significantly different gene expression levels in responders compared to non‐responders (p < 0.05). Functional groups represented by the signatures included DNA damage repair, extracellular matrix remodeling, and 5FU metabolism. Partial Least Squares (PLS) prediction of treatment response was compared to pathologic TRG determined by blinded pathologic reading, and resulted in an area under the curve (AUC) of 0.99 and overall accuracy of 100% for the 24‐gene signature. Corrected AUC of 0.99 and accuracy of 95% resulted from five‐fold cross validation with 20 iterations. Heatmap analysis of the 24‐gene signature separated the EC cases into two distinct clusters, the first with 93% responders and the second with 90% non‐responders. The current study identifies novel gene signatures able to accurately predict EC patient response to preoperative treatment. The GEP may allow non‐responders to avoid unnecessary toxicities associated with chemoradiation therapy.

De nombreuses études en oncoimmunologie portent sur l’échec immunitaire dans le contexte de progression tumorale, mais elles sont plus rares à porter sur un contexte de régression, lorsque le système immunitaire est efficace. De ce fait, bien souvent la littérature met en avant le rôle cytotoxique des lymphocytes T CD8+, ou bien leur anergie dans le contexte de progression tumorale, causée par des cellules myéloïdes telles que les MDSC ou les macrophages de phénotypes M2, considérés comme pro-tumoraux. J’ai pour ma part étudié la réponse immunitaire dans le cadre d’une régression tumorale. Des cellules TC1 transplantées en s.c. dans des souris C57BL6/J, donnent des tumeurs solides d’environ 6mm de diamètre 11 jours plus tard. A ce moment là (J0), les souris sont vaccinées à proximité de la tumeur (priming), par un vaccin contenant la sous-unité B non toxique de la Shiga toxine couplée au peptide E7 de l’HPV16 (exprimé par les TC1), combiné à de l’IFNα. Une semaine plus tard (J7), un « boost » est effectué. Après le boost, la croissance tumorale cesse puis la tumeur régresse. L’analyse cinétique par cytométrie révèle un infiltrat immunitaire important pendant, et précédant la régression tumorale. La nature de cette infiltrat varie avec le temps. A J5, un infiltrat myéloïde est observé, suivi d’un infiltrat lymphocytaire à partir de J8. Une déplétion des cellules T CD8+ inhibe la régression tumorale, alors que dans les souris CXCR3-/-, dans lesquelles les CD8+ ne sont pas déplétés mais leur recrutement est fortement affecté, une régréssion tumorale est possible malgré un infiltrat T CD8+ très faible. Cela laisse penser que d’autres acteurs que les LT cytotoxiques sont nécessaires à la régression tumorale, comme probablement les cellules myéloïdes qui infiltrent le tumeur avant les cellules T. L’analyse de cette population montre une activation des monocytes et macrophages (MHC II+), avec un pic d’activation autour de J9, au début de la régression. La capacité cytotoxique de ces cellules, mesurée in vitro par immunofluorescence est augmentée comparée à des myéloïdes isolées de tumeurs de souris en progression. De plus, l’ajout d’un anticorps anti-TNFα inhibe partiellement cette cytotoxicité. Cela montre qu’après vaccination, les monocytes/macrophages sont capables de tuer les cellules tumorales. Une déplétion partielle des macrophages au moment de la vaccination, à l’aide du PLX3397 (inhibiteur du CSF1R), réduit l'efficacité de la vaccination. Les cellules myéloïdes, lorsqu'elles sont présentes, contribuent fortement à la régression tumorale induit par le vaccin composite, et leur action implique probablement des interactions avec les LT CD8+. C'est ce que suggère l'observation de tumeurs vaccinées dans des souris IFNϒ-/-, dans lesquelles l'efficacité vaccinale est aussi inhibée. Cette thèse montre qu’après une stimulation appropriée, qui peut, comme ici, mimer une infection virale, les cellules myéloïdes peuvent participer activement à la régression tumorale. / Most oncoimmunology studies are performed in an immune failure context of progressing tumor. They rarely describe tumor regressions, when the immune response is efficient. As a result, the literature tends to highlight the cytotoxic role of CD8+ T cell or their anergy in the context of tumor progression, caused by myeloid cells such as the MDSC or M2 polarized macrophages, considered as protumoral. My PhD work has been focused on the immune response in a context of tumor regression. TC1 cells transplanted s.c. in C57 BL6 J mice, give rise to solid tumors of approximately 6 mm diameter 11 days later. At that time (day 0), mice are vaccinated peritumorally for a priming with a composite vaccine containing the subunit B of the Shiga toxin coupled to E7 peptide from HPV16 (present on TC1), combined with the IFNα. A week later (day 7), a boost is made. After the boost, tumor growth stops and the tumor regress. Kinetic cytometric analysis revealed a significant immune infiltrate during and prior to tumor regression. The nature of this infiltrate varies with time. On day 5, a myeloid infiltrate is observed, followed by a lymphocytic infiltrate which is conspicuous after day 8. Depletion of CD8+ T cells inhibits tumor regression, while in CXCR3- /- mice, in which the CD8+ are not depleted but their recruitment is severely affected, tumor regression is possible despite a very low CD8+ T cell infiltrate. This suggests that some effectors, other than cytotoxic T cells, are required for tumor regression, including probably myeloid cells that infiltrate the tumor before T cells. The analysis of this population shows an activation of monocytes and macrophages (MHC II+) with a peak of activation around day 9, early in the regression. The cytotoxic capacity of these cells was tested in vitro, by depositing F4/80+ cells from vaccinated tumors or not, on a TC1 cell monolayers in culture. Only myeloid cells from vaccinated tumors appear to kill tumor cells, and adding an anti-TNFα inhibits this cytotoxicity. This shows that after immunization, monocytes/macrophages are capable of killing tumor cells. A partial depletion of macrophages at the time of vaccination, after treatment with PLX3397 (CSF1R inhibitor), reduces the vaccine efficacy. Myeloid cells contribute significantly to the observed tumor regression, and their action involves interactions with CD8+ T cells. This hypothesis is consistent with the observation of tumors in vaccinated IFNϒ- /- mice, in which the vaccine efficacy is also inhibited. This thesis shows that after an appropriate stimulation, for instance, here, by mimicking a viral infection, myeloid cells can actively participate in tumor regression.

Novel therapies and delivery methods directed against malignancies such as melanoma, and particularly metastatic melanoma, are needed. The HIV-1 accessory protein Vpr (viral protein R) has previously been demonstrated to induce G2 cell cycle arrest as well as in vitro growth inhibition/killing of numerous tumor cell lines. In vivo electroporation has been utilized as an effective delivery method for pharmacologic agents as well as DNA plasmids that express "therapeutic" proteins and has been targeted to various tissues including malignant tumors. In this study, we assessed the ability of electroporation-mediated delivery of Vpr plasmid (pVpr) to induce growth attenuation or complete tumor regression in C57BL/6 mice with subcutaneous B16.F10 melanoma lesions. To assess the administration of intratumoral delivery of pVpr with in vivo electroporation, a range of Vpr plasmid dosages, electroporation parameters, and treatment days were evaluated in a subcutaneous B16 murine melanoma model. pVpr was injected directly into the tumors. Immediately following the injection, the subcutaneous tumors were electroporated. Treatment with 25 microgram or 100 microgram of pVpr plus electroporation on days 0 and 4 resulted in complete tumor regressions with long-term survival in 14.3% and 7.1% of the mice, respectively. In order to optimize the treatment regimen, B16 tumors were treated on days 0, 2, and 4 with 100 microgram pVpr plus electroporation which resulted in 50% of the mice with complete tumor regressions and long-term survival. Additional investigations revealed intratumoral Vpr expression and demonstrated that apoptosis was the mechanism by which Vpr caused tumor regression in vivo. This study confirmed that treatment with 100 microgram of pVpr plus electroporation led to durable complete regressions in established murine melanoma lesions. The pVpr plus electroporation treatment regimen has induced complete regressions in mice as well as resistance to tumor challenge in some of the animals. This is the first comprehensive study demonstrating the ability of Vpr, when delivered as a DNA expression plasmid with in vivo electroporation, to induce complete tumor regressions coupled with long- term survival of mice in a highly aggressive and metastatic solid tumor model.

Background: Early efficacy outcome measures in rectal cancer after total neoadjuvant treatment are increasingly investigated. We examined the prognostic role of pathological complete response (pCR), tumor regression grading (TRG) and neoadjuvant rectal (NAR) score for disease-free survival (DFS) in patients with rectal carcinoma treated within the CAO/ARO/AIO-12 randomized phase 2 trial. Methods: Distribution of pCR, TRG and NAR score was analyzed using the Pearson’s chi-squared test. Univariable analyses were performed using the log-rank test, stratified by treatment arm. Discrimination ability of non-pCR for DFS was assessed by analyzing the ROC curve as a function of time. Results: Of the 311 patients enrolled, 306 patients were evaluable (Arm A:156, ArmB:150). After a median follow-up of 43 months, the 3-year DFS was 73% in both groups (HR, 0.95, 95% CI, 0.63–1.45, p = 0.82). pCR tended to be higher in Arm B (17% vs. 25%, p = 0.086). In both treatment arms, pCR, TRG and NAR were significant prognostic factors for DFS, whereas survival in subgroups defined by pCR, TRG or NAR did not significantly differ between the treatment arms. The discrimination ability of non-pCR for DFS remained constant over time (C-Index 0.58) but was slightly better in Arm B (0.61 vs. 0.56). Conclusion: Although pCR, TRG and NAR were strong prognostic factors for DFS in the CAO/ARO/AIO-12 trial, their value in selecting one TNT approach over another could not be confirmed. Hence, the conclusion of a long-term survival benefit of one treatment arm based on early surrogate endpoints should be stated with caution.

Les cancers colorectaux et les cancers du sein sont deux des affections malignes les plus fréquentes dans les pays industrialisés. Lorsqu’elles sont diagnostiquées à un stade précoce, ces tumeurs sont traitées efficacement par la chirurgie associée ou non à la radiothérapie. Pour des tumeurs localisées ayant certains facteurs de pronostic péjoratifs, pour les tumeurs localement avancées, ou pour les tumeurs métastasiques, une chimiothérapie est instaurée. Cependant, la chimiothérapie à elle seule ne permet pas de guérir les patients aux stades avancés de ces 2 types de cancers. C’est pourquoi d’autres alternatives comme l’immunothérapie ou des traitement plus ciblés sont en cours d’étude ou viennent récemment d’être validés.Notre travail a porté sur deux modèles d’immunothérapie dans deux formes de cancers . 1/Dans un modèle expérimental chez le rat, un traitement par un analogue de lipide A permet de guérir tous les animaux porteurs de carcinomatose péritonéale macroscopique d’origine colique. 2/ Chez la femme porteuse de carcinome mammaire localement avancé et surexprimant l’oncoprotéine HER2, un traitement préopératoire comportant du trastuzumab permet d’obtenir la disparition complète des tumeurs et des métastases axillaires chez la moitié des patientes. A partir de prélèvements tissulaires issus de ces 2 modèles, nous avons pu montrer que des analyses histologiques et surtout immunohistochimiques pouvaient permettre d’appréhender les mécanismes de régression tumorale.Dans les régressions des carcinomatoses chez le rat, le rôle du monoxyde d’azote, de l’apoptose des cellules tumorales et de l’infiltration des tumeurs par les cellules dendritiques puis des macrophages a été proposé. Dans la régression des tumeurs du sein, le rôle des cellules NK, via un mécanisme d’ADCC a été suggéré. Ces travaux ont permis de montrer que, malgré ses limites, et en particuliers l’impossibilité de démonstration mécanistique, l’immunohistochimie peut permettre de proposer des hypothèses intéressantes, qui doivent être secondairement confirmées par des expérimentations complémentaires. / In the developed countries, colorectal and breast cancers are two of the most common malignancies. When the diagnosis is made at the beginning of the disease, surgery, associated or not with radiotherapy may cure the patients. For tumors that present some pejorative prognostic factors,for locally advanced tumors or for metastasic disease, chemotherapy has to be prescribed. However, in these 2 types of cancer, chemotherapy is not able to cure patients suffering from metastasic stage and alternative treatments as immunotherapy or targeted therapies are studied or are still validated.Our work was based on 2 different models of immunotherapy in 2 types of cancer. In an experimental model of carcinomatoses of colorectal cancer in rats, Lipid A injections are able to cure all the rats, even at a macroscopic stage and in half of the women affected by locally advance HER2-overexpressing breast cancer, trastuzumab based preoperative treatment are able to obtain the total disappearance of the tumor. When we look on tumor samples obtained in these two models, we can show that histology and even more immunohistochemistry are able to propose some mechanisms of regression of the tumors. In the rat model, we can show the role of nitric oxide, apotosis, dendritic cells and macrophages and in the breast cancer regression, we show the role of NK cells and ADCC. Theses works show that immunohistochemistry, even if insufficient, may propose some interesting hypotheses that have to be confirm by other experimentations.