Hepatocellular carcinoma (HCC) is one of the most devastating malignancies worldwide with increasing incidences in both developed and developing countries. Survival rates have not been significantly improved over the past decades in spite of advances in detection and therapies for this disease, suggesting that current treatments may target the wrong cells, and miss the cancer stem cells (CSCs). The cancer stem cell hypothesis presents that tumor formation, proliferation and propagation are driven by a rare subpopulation of chemoresistant CSCs that are not killed by conventional therapies and go on to cause disease relapse. The objective of this study was to identify and characterize CSCs in HCC cell lines and human liver tumor specimens using CD90 as a potential marker. The number of CD90+ cells present in HCC cell lines was found to positively correlate with tumorigenicity potentials. Injection of as few as 2,000 sorted CD90+ cells from HCC cell lines resulted in the formation of tumor nodules in nude mice, whereas no tumors formed for CD90ˉcells in the same model. The tumor xenograft generated by injection of CD90+ cells sorted from previous xenograft in a serial xenotransplantation assay exhibited recapitulation of tumor heterogeneity to original primary tumor and consistent proportion of CD90+ and CD90ˉ cells which demonstrated self-renewal and differentiation capacities of CD90+CSCs. CD45ˉCD90+ cells were detected (0.03%–6.2%) in human liver tumor specimens, but were only present in minute quantities in normal, cirrhotic and non-tumorous tissues. More importantly, CD45ˉCD90+ cells sorted from primary HCC tumor also displayed tumorigenicity, self-renewal and lineage differentiation capacities. CD90+CSCs were found to be more resistant to therapeutic drugs compared to CD90- cells, as reflected by the results of enrichment of the CD90+ CSCs and longer survival rates after chemotherapeutic treatment. The high expression of genes, such as OCT4, MRP3, ABCG2, AKT1, BirC5, BCL2, HA and CD44, in CD90+CSCs may mediate chemoresistance. The majority of CD90+ cells co-expressed CD44, another stem cell marker. Blocking CD44 activities by anti-CD44 antibody increased apoptosis of CD90+ CSCs, sensitized CD90+CSCs to chemotherapeutic drugs in vitro, and decreased tumorigenic and metastatic potentials of CD90+CSCs in vivo, indicating that a therapeutic potential of targeting CD44. However side effects may be problematic due to the endogenous expression of CD44 in healthy tissues and normal lymphocytes. To identify novel gene targets specific to liver CSCs, a sensitive RNA-sequencing (RNA-Seq) technique was used to compare the gene expression profiles between CD90+CSCs sorted from HCC primary tumors and CD90+cells from adjacent non-tumorous tissue (CD90+NTSCs). The up-regulated genes in CD90+CSCs were associated with lipid metabolism, inflammation, and drug resistance. Among the differentially expressed genes, glypican-3 (GPC3) was specifically elevated in CD90+CSCs but not in CD90+NTSCs. Therefore, GPC3 could be a promising gene candidate for HCC therapy as targeting GPC3 should not induce damage to normal liver stem cells. In summary, CD90 is a liver CSCs marker. Identification of CD90+ CSCs in HCC provides new insight into cellular basis of hepatocarcinogenesis, recurrence and metastasis, which opens new avenues for the design of future CSC-targeted therapies. / published_or_final_version / Surgery / Doctoral / Doctor of Philosophy

Chemoresistance remains a major clinical obstacle to effective management of ovarian cancer. Cancer stem cells (or tumor-initiating cells, TICs) have been discovered recently, and have played a pivotal role in changing the view of cancer development; however, the molecular mechanisms by which these cells escape conventional therapies remain elusive. In this study, TICs were isolated from ovarian cancer cells as tumor spheres with specific stem properties under TIC-selective conditions. Unlike non-TICs, TICs strongly express stem cell factor (SCF) and c-Kit. Blocking SCF-c-Kit by SCF neutralizing antibodies, c-Kit small interfering RNA (siRNA) or imatinib (Gleevec), a clinical drug that inhibits c-Kit signaling, significantly inhibited TIC proliferation. Although cisplatin and paclitaxel killed the non-TICs, they did not eliminate TICs. Importantly, the combination of cisplatin/paclitaxel with c-Kit siRNA or imatinib inhibited the growth of both non-TICs and TICs. Similar results were obtained when patient-derived TICs were used. The findings also indicate that tumor-predisposing microenvironment, such as hypoxia, may promote ovarian TICs through upregulating c-Kit expression. Furthermore, I have showed that c-Kit expression induced activation of Phosphatidylinositol 3-kinases (PI3K)/Akt, -catenin, and ATP-binding cassette G2, which could be reversed by treatment with the PI3K/Akt inhibitor or -catenin siRNA. I further studied potential gene expression in TICs using cDNA and microRNA (miRNA) microarrays. The result from these microarrays provided a general profile in gene expression of TICs compared with the bulk tumor cells. In particular, let-7a, b, and c were shown to be downregulated in TICs compared to bulk tumor cells, suggesting that their loss may contribute to ovarian cancer development. Together, this study reveals a previously undescribed therapeutic effect of SCF-c-Kit signaling blockade to prevent ovarian cancer progression by eliminating TICs and the altered genes or miRNAs may represent possible molecular targets. / published_or_final_version / Biological Sciences / Master / Master of Philosophy

published_or_final_version / abstract / toc / Zoology / Master / Master of Philosophy

published_or_final_version / Anatomy / Master / Master of Philosophy

No description available.

No description available.

Recent studies have shown that the proliferation of various human and murine tumour lines can be inhibited by the addition of gamma-linolenic acid (GLA) to the culture medium. These findings are consistent with the proposal put forward by Horrobin (1980) that malignant cells lack the enzyme/ A 6 desaturase; which is responsible for the conversion of linoleic acid (LA) to GLA. Since GLA is a prostaglandin (PG) precursor/ inadequate conversion of LA to GLA would result in decreased production of PGs/ particularly PGEi/ which has been shown to have an inhibitory effect on cell growth. Provision of GLA to enzyme deficient malignant cells should therefore bypass this blockade/ increase PGET synthesis and thus "normalise malignant cells". This study was performed to examine further the effects of exogenous GLA on growth of malignant cells in vitro and in vivo. Cells of the continuous murine sarcoma (M52B) line and primary cultures of non malignant fibroblasts were used to investigate effects of GLA in vitro. Cultures were exposed to either single or multiple doses of a range of concentrations of GLA. Radioimmunoassay (RIA) was performed to compare the amounts of PGE and PGF released into the medium by GLA treated and control M52B cultures and thus determine whether the addition of GLA in vitro significantly affected production of these PGs. Athymic BALB/c mice and immunocompetent BALB/c and Biozze mice as well as mice of the "Onderstepoort Strain" were used in various in vivo studies. Tumours were induced by the subcutaneous inoculation of approximately 1 x 106 cells of either the M52B line (into immunocompetent and athymic mice) or human breast carcinoma (NUB 1) line (into athymic mice). Take rates and latent periods were recorded. GLA treatment was initiated after tumours were established. In one study the fatty acid in hydrogenated coconut oil (HCO), which contains no PG precursors/ was administered parenterally (100 ug/ml/day) to Biozze mice. Control mice were either untreated or injected with HCO only. In another study, BALB/c mice and mice of the "Onderstepoort Strain" had their diet supplemented with GLA (in the form of EPO) to an extent of 3.5%. Control mice consumed either standard laboratory chow only or, chow supplemented with either 35% sunflower seed oil (SSO) or 35% HCO/ neither of which contain GLA. All diets were supplied ad libitum. Tumour sizes were measured every 48 hours and at the end of each experiment at which time tumours were excised and examined histologically. GLA was found to produce inhibitory and toxic effects on growth of both M52B cells and non malignant fibroblasts in vitro/ although the effect in the latter was observed only with high concentrations of the fatty acid. The inhibition of malignant cell growth was concentration dependant and was positively related to the duration of exposure to the fatty acid. Prior to death/ cells treated with GLA accumulated vii paranuclear granules which were shown histochemically to be lipid in nature. Electron microscopy confirmed the presence of large lipid deposits. Cultured M52B cells treated with GLA also released more PGE and PGF into the medium than did cells not exposed to the fatty acid. However, analysis of results using the Mann Whitney U test showed these differences to be statistically non significant for both PGE and PGF on two tailed tests. In contrast to the inhibition of M52B cell growth observed in vitro, growth of solid M52B sarcomas and NUB 1 carcinoma xenografts in athymic mice was apparently unaffected by administration of dietary GLA. Analysis of data using an unpaired student's t-test showed that the differences in tumour volumes between control and treated groups were not statistically significant either before or at the end of the experiment. While the inhibition of malignant cell growth caused by GLA in vitro was consistent with Horrobin's proposal that malignant cells may lack this PG precursor, whether or not these actions are mediated by the PGs remains obscure. Although an increase in PGE production by M52B cells was observed following GLA treatment, besides this increase being statistically non significant, it was not possible to determine whether this was due to PGE, (as suggested by Horrobin) or PGE2. It is possible that the effect produced in vitro was due to some factor other than raised PGE production, for example a non-specific fat-overload effect or a change in cell membrane fluidity. The lack of effect of GLA on tumour growth in vivo may have been due to inadequate delivery of the fatty acid to the tumour site. However, whatever the mechanism of action of GLA in vitro/ since oral GLA was supplemented to the maximum tolerated extent and produced no effect in immunodeficient mice inyiyo, it would seem that in a similar clinical situation oral doses which would be practical may be ineffective. / Thesis (M.Sc.)-University of Natal, Durban, 1985.

No description available.

The purpose of this research was to determine the fatty acid composition of lipids as compared to those of normal tissue. The altered molecular structure of these lipids could cause changes in the plasma membrane fluidity and membrane integrity producing the membrane characteristics associated with carcinomas.Plasma membranes-were isolated by differential centrifugation from both tumor and normal tissue from Strong A female mice. The lipids were extracted, separated by thin-layer chromatography, saponified and the fatty acid methyl esters identified and quantified by gas chromatography. The lipids were separated into four classes, cholesterol esters, triglycerides, free fatty acids and phospholipids. The carcinoma samples exhibited a reduction in the fatty acids of 10 to 14, carbons in chain length by as much as 32 percent. The level of palmitic acid increased in the tumor fractions as much as double the normal amount. The level of palmitoleic acid also increased in the tumor fractions nearly proportional to the palmitic acid. The stearic acid content in the tumor fractions increased as did the oleic acid in three of the four lipid classes, again nearly proportional to the stearic acid. Levels as high as 29 percent of henicosanoic acid in some normal tissue samples were found, but the carcinoma samples exhibited no higher than 10 percent. The increased production of heptadecadienoic acid in the carcinoma fractions along with the appearance of eicosatrienoic acid when only low scattered amounts were found in the normal fractions indicated that a shift in fatty acid biosynthesis has occurred, it would appear that in the carcinoma, the biosynthetic pathway for synthesis of fatty acids in the plasma membranes shifted from a de novo to a chain elongation pathway as the principal route.

Progress in understanding cancer progression has been hampered over the years by the different types of mutations present and irregular changes of gene regulation associated with any given cancer. In this work, molecular interactions between TALI, LMO1, and decorin were investigated. Numerous studies have shown that ectopic expression of decorin protein induced growth suppression in neoplastic cells of various histogenic origins. Furthermore, ectopic expression of TAL1 and LMOI oncoproteins has been shown to occur in approximately 50% of the cases of T-cell acute lymphoblastic leukemia (T-ALL). It was of interest then to determine the preventive or interactive role decorin played with the oncogenic activity of TAL I and LMO1. In this investigation, decorin was introduced into a murine T-cell line (AKR-DP-603) through the use of the mammalian expression vector pcDNA3.1 (-). This particular cell line was previously engineered to express TALI and LMO1. Protein expression patterns in all cell populations were analyzed using the Western blot technique and a proteoglycan with a molecular weight of 100 kDa before chondroitinase ABC treatment and a core protein of55 kDa after treatment with chondroitinase ABC was seen. This finding is significant since it implies that the pcDNA3. 1(-) vector containing decorin cDNA was present, and the corresponding decorin peptides were expressed in both cytoplasmic and nuclear extracts. Furthermore, Northern blot analysis was performed on total RNA extracts to determine the transcriptional state of endogenous decorin rRNA, as well as exogenously introduced decorin. Northern blot analysis revealed no decorin-specific mRNA transcripts from the various cell populations. This result did not imply a lack of possible regulatory effect on protein and mRNA levels of TALL and LMOI by decorin. Finally, cell growth assays were performed on all cell populations and cell counts were used to assess the growth pattern of each population after serum withdrawal. The results show possible growth suppressive effects of decorin on TAL1 and LMOI expressing cells. Results obtained from this study shed further light on the molecular interactions influencing T-ALL and may also help in the design of potentially beneficial cancer treatments using decorin. / Department of Biology