The cancer stem cell (CSC) concept is still very controversial; therefore identification and isolation of this specific population remain challenging. A variety of putative markers have been described and measurement of high aldehyde dehydrogenase (ALDH) activity has been defined as a characteristic of stem cells (SCs). In this study, a library of novel small molecules (1,4-disubstituted acetalanthraquinones, AAQs), containing an acetal group as protected aldehyde functionality, was designed with the aim of probing affinity for ALDH metabolism and demonstrating their potential as molecular fluorescent probes to identify CSCs. The AAQs were shown to be subjective to acidic hydrolysis using 2M HCl at 37ºC; however compounds containing secondary or tertiary amine functionalities in their sidechain were only partly hydrolysed at 70 ºC. Metabolism studies were conducted using cytosolic fractions from rat liver enriched in ALDHs, yeast ALDH and human recombinant ALDH1A1. Some evidence was demonstrated which linked ALDH metabolism with aldehyde functionalities of hydrolysed AAQs (HAAQs). The AAQs were shown to emit far-red fluorescence (600-750 nm). A close relationship between structure modifications and alteration of cellular localisation, with gained specificity for selected sub-cellular compartments were achieved when assessed in A549 and U-2 OS cell lines. Thermal DNA denaturation and chemosensitivity assays were used to obtain information about DNA binding properties and cytotoxicity of AAQs and HAAQ congeners. All compounds were shown to be weak-to-moderately binding to DNA, and symmetrical 1,4-disubstituted compounds were shown to be non-toxic (IC50 = 100 :M) with nonsymmetrical analogues generating IC50 values in the 1-100 :M range. No fundamental variation in the biological activity was observed when comparing AAQs with HAAQs in the A549 (+ALDH) and MCF7 (-ALDH) cell lines. A pilot investigation revealed that aberrant gene methylation was cell-type dependent for three ALDH isoforms (1A1, 2, 3A1). Decitabine treatment led to enhanced protein expression for ALDH1A1 (A549), ALDH2 (MCF7) and ALDH3A1 (A549). In contrast, the protein level was reduced for ALDH1A1 in HT29 cells after decitabine treatment. ALDH1A1, ALDH2 and ALDH3A1 were highly expressed in prostate cell lines, with expression linked to promoter methylation. In contrast, low levels of DNA methylation were found in primary prostate cancer cells and benign prostatic hyperplasia. Interestingly, ALDH1A1, considered a SC marker, was found to be expressed at low levels in CD133⁺/ α₂β₁ hi stem cell fraction and upregulated in CD133⁻/ α₂β₁ lo differentiated prostate cancer cells. In summary, the results in this thesis demonstrate the complexity and tumour type specificity of ALDH expression. This creates challenges for the development of selective probes for CSC isolation, such as the AAQs discussed in this thesis. Although inconclusive results were obtained in regard to AAQs and their potential in targeting ALDHs, selected AAQs were shown to reveal interesting biological features highlighting them as potential non-invasive cytometric probes for tracking molecular interactions in live cells.

The cancer stem cell (CSC) concept is still very controversial; therefore identification and isolation of this specific population remain challenging. A variety of putative markers have been described and measurement of high aldehyde dehydrogenase (ALDH) activity has been defined as a characteristic of stem cells (SCs). In this study, a library of novel small molecules (1,4-di-substituted acetalanthraquinones, AAQs), containing an acetal group as protected aldehyde functionality, was designed with the aim of probing affinity for ALDH metabolism and demonstrating their potential as molecular fluorescent probes to identify CSCs. The AAQs were shown to be subjective to acidic hydrolysis using 2M HCl at 37ºC; however compounds containing secondary or tertiary amine functionalities in their sidechain were only partly hydrolysed at 70 ºC. Metabolism studies were conducted using cytosolic fractions from rat liver enriched in ALDHs, yeast ALDH and human recombinant ALDH1A1. Some evidence was demonstrated which linked ALDH metabolism with aldehyde functionalities of hydrolysed AAQs (HAAQs). The AAQs were shown to emit far-red fluorescence (600-750 nm). A close relationship between structure modifications and alteration of cellular localisation, with gained specificity for selected sub-cellular compartments were achieved when assessed in A549 and U-2 OS cell lines. Thermal DNA denaturation and chemosensitivity assays were used to obtain information about DNA binding properties and cytotoxicity of AAQs and HAAQ congeners. All compounds were shown to be weak*to*moderately binding to DNA, and symmetrical 1,4-di-substituted compounds were shown to be non*toxic (IC50 = 100 :/! with non-symmetrical analogues generating IC50 values in the 1-100 :/ range. No fundamental variation in the biological activity was observed when comparing AAQs with HAAQs in the A549 (+ALDH) and MCF7 (-ALDH) cell lines. A pilot investigation revealed that aberrant gene methylation was cell-type dependent for three ALDH isoforms (1A1, 2, 3A1). Decitabine treatment led to enhanced protein expression for ALDH1A1 (A549), ALDH2 (MCF7) and ALDH3A1 (A549). In contrast, the protein level was reduced for ALDH1A1 in HT29 cells after decitabine treatment. ALDH1A1, ALDH2 and ALDH3A1 were highly expressed in prostate cell lines, with expression linked to promoter methylation. In contrast, low levels of DNA methylation were found in primary prostate cancer cells and benign prostatic hyperplasia. Interestingly, ALDH1A1, considered a SC marker, was found to be expressed at low levels in CD133+/ α2β1hi stem cell fraction and upregulated in CD133-= α2β1lo differentiated prostate cancer cells. In summary, the results in this thesis demonstrate the complexity and tumour type specificity of ALDH expression. This creates challenges for the development of selective probes for CSC isolation, such as the AAQs discussed in this thesis. Although inconclusive results were obtained in regard to AAQs and their potential in targeting ALDHs, selected AAQs were shown to reveal interesting biological features highlighting them as potential non-invasive cytometric probes for tracking molecular interactions in live cells. / EPSRC, Biostatus / The full text was made available at the end of the re-embargo period, 1st September 2017.

Il est généralement admis que le cancer du sein représente un ensemble de plusieurs maladies, définies comme des sous-types ayant des caractéristiques moléculaires et cliniques qui leurs sont propres. Une meilleure compréhension des mécanismes qui sous-tendent l'hétérogénéité du cancer du sein est essentielle au développement de thérapies mieux ajustées. Le concept de cellules souches cancéreuses (CSC) pourrait être un des clés de cette compréhension. A ce jour, un certain nombre de marqueurs ont été proposés pour isoler et caractériser les cellules souches dans le cancer du sein, mais aucun ne semble totalement satisfaisant.Le but de mon travail était de déterminer un marqueur ou une combinaison de marqueurs avec lesquels les fractions enrichies en CSC pourraient être isolées de manière reproductible dans le cancer du sein de sous-types basal (BLBC). En effet, les tumeurs basales représentent 15% de toutes les tumeurs mammaires, mais constituent le sous-type le plus agressif. À cet effet, j'ai analysé un certain nombre de marqueurs par analyse FACS et tri cellulaire et utilisé la capacité de formation de mammosphères (MS) comme critère de validation pour la présence de CSC. Les lignées cellulaires utilisées comme modèles étaient les SUM 159, MDA-MB-231, MDA-MB-436, HCC1143, MDA-MB-468, Hs578T et BT-549 correspondant aux modèles basal-A et B. J'ai également testé trois lignées luminales les MCF7, T47D et BT474.De tous les marqueurs testés, seules, la combinaison des protéines de surface cellulaire CD44/CD24/EpCAM et l’activité enzymatique ALDH élevée ont permis d’obtenir un enrichissement significatif en CSC. Toutefois, le niveau de l'activité ALDH est apparu inconstant d’une lignée cellulaire à une autre et selon le type de tumeurs. D'autres marqueurs membranaires ont donné des résultats mitigés dans le cancer du sein ER-. En effet, la plupart des lignées basales ont montré des profils FACS assez homogènes avec des proportions élevées de cellules CD44+. Cependant, l'association de la positivité de CD44 avec l'EMT et la souchitude, ainsi que la bonne corrélation observée dans les modèles luminaux de la population de cellules CD44+/CD24- avec l’enrichissement en CSC, nous a incité à déterminer si le niveau d'expression en CD44 faisait une différence dans les tumeurs basales. Sur cette base, j’ai montré que les cellules CD44 high présentent une forte capacité à former des MS dans toutes les lignées cellulaires testées. Cette constatation nous a incités à utiliser CD44high vs. CD44low comme critère de tri cellulaire et à utiliser ces fractions pour effectuer une analyse du transcriptome afin d'identifier d'autres marqueurs non encore déterminés, pouvant isoler des fractions cellulaires plus faibles avec un enrichissement plus élevé en CSC. / It is now accepted that breast cancer is a compendium of several diseases defined as subtypesthat are associated with different clinical outcomes and molecular characteristics. A betterunderstanding of the mechanisms underlying breast cancer heterogeneity is critical to the development of better adjusted therapies. One of the keys to breast cancer heterogeneity may be explained by cancer stem cells (CSC). A number of markers have been proposed to isolate and characterize breast cancer stem cells, but none appears totally satisfactory.The purpose of my work was determine a marker or combination of markers with which CSC enriched fractions could be reproducibly isolated in basal like breast cancer (BLBC). BLBC represent 15% of all breast tumors, but are the most aggressive subtype. To this aim, I have analyzed a number of markers by FACS analysis and cell sorting and used the capacity to form mammospheres (MS) as a validation criterion for the presence of CSCs. The cell lines used as models were SUM 159, MDA-MB-231, MDA-MB-436, HCC1143, MDA-MB-468, Hs578T and BT-549 comprising both Basal A and Basal B models. I also tested three luminal models MCF7, T47D and BT474.Of all the markers tested those that most consistently allowed enrichment of CSCs were the combination of cell surface proteins CD44/CD24/EpCAM and elevated ALDH enzyme activity. However, ALDH activity appeared irregular, ranging from good to inconsistent according to the cell line. Other cell surface markers gave mixed results in ER- breast cancer because the elevated fraction of CD44+ cells found in most of basal breast cancer cell lines and their propensity to show rather homogenous FACS labeling patterns. However, the association of CD44 positivity with EMT and stemness, as well as the good correlation, we observed in luminal models, of CD44+/CD24- cell population with CSC enrichment incited us to determine whether the level of expression of CD44 could make a difference in basal like models. I show that CD44high cells present higher capacity to form MS in all cell line models tested. This prompted us to use CD44high vs. CD44low as a cell sorting criterion and use these fractions to perform transcriptome analysis in order to identify other markers yet not determined, that may point to smaller cell fractions with a higher CSC enrichment.