Introduction The heterogeneous molecular landscape of cytogenetically normal acute myeloid leukaemia (CN-AML) renders it an ongoing therapeutic challenge worldwide. The latest European LeukaemiaNet (ELN) 2017 guidelines attempt to address this by guiding post-remission therapy according to six prognostically informative mutations. However, its applicability in a South African setting remains elusive due to limited local data. This retrospective study aimed to describe a South African CN-AML cohort according to clinicopathological, molecular and treatment outcomes and consequently investigate the local applicability of a triple mutation testing approach for nucleophosmin (NPM1), fms-like tyrosine kinase internal tandem duplication (FLT3-ITD) and CCAAT/enhancer binding protein alpha (CEBPα) mutations in accordance with the ELN 2017 guidelines. Methods A review of cytogenetic results for all adult de novo AML cases diagnosed at Groote Schuur Hospital between 2005 and 2018 was performed. CN-AML cases were further characterized via molecular testing and review of clinical and laboratory data. Results In total, 218 patients with AML were identified of which fifty-six (33%) were cytogenetically normal. NPM1, FLT3-ITD and CEBPα mutations were found in 39%, 34% and 9% of CN-AML cases respectively, and allowed for definitive prognostication of 50% of cases. The 2-year overall survival rate for the entire CN-AML cohort was 16%. Conclusion Local rates of CN-AML and associated NPM1 and FLT3-ITD mutations were comparable to European cohorts. In contrast, local survival outcomes were notably inferior. Triple testing proved a resource effective prognostication approach for CN-AML. High throughput sequencing for adverse risk mutations should be considered for CN-AML patients inconclusively stratified via triple testing.

Background: Human immunodeficiency virus (HIV) is associated with an increased risk of developing Hodgkin lymphoma (HL). South Africa (SA) has the highest HIV prevalence rate in the world. There is currently no 5-year overall survival (OS) outcome based data for HIV-associated HL from SA. Methods: A bone marrow database was compiled of all bone marrow biopsies (BMB) reported at National Health Laboratory Service (NHLS) Groote Schuur Hospital (GSH) between January 2005 and December 2012. Patients who had a BMB performed for staging of HL or where HL was diagnosed on the BMB were included for further analysis. Clinical and laboratory data was extracted from medical and laboratory records. Primary outcome measures included histological subtype, bone marrow infiltration (BMI) by HL, CD4 count, HIV-viral load (HIV-VL), tuberculosis (TB) data, treatment with chemotherapy and 5-year overall survival (OS). Results: The database included 6569 BMB and 219 patients of these had HL and were included for analysis. The median age at presentation (32 years) was similar in the HIV+ and HIV-populations. While males predominated in the HIV-group, females predominated in the HIV+ group (male:female ratio of 1.5:1 vs 0.7:1, respectively). The majority of patients (71%) were HIV negative (HIV-) and 29% were HIV positive (HIV+). The diagnosis of HL was made on BMB in 17% of cases. BMI was seen in 37%(82/219) overall, and was found in more HIV+ patients (61%; 39/64) than HIV-patients (28%; 43/155; p= 0.03). The histological subtype varied according to HIV status with nodular sclerosis classical Hodgkin lymphoma (NSCHL) being most frequent in the HIV-group and classical Hodgkin lymphoma (CHL)-unclassifiable the most frequent in the HIV+ group. HIV+ patients had a median CD4 count of 149 x106/L and 39% were anti-retroviral therapy (cART) naive at HL diagnosis. HIV+ patients had received anti-TB therapy more frequently than HIV-patients (72% vs 17%; p= 0.007). More HIV+ patients did not receive chemotherapy than HIV-patients (31% vs 3%; p= 0.001). The 5-year OS was 56%. HIV+ patients with BMI had a 5-year OS of 18%. BMI, HIV status, low CD4 count, histological subtype and TB therapy had a statistical significant impact on 5-year OS (p< 0.01). Conclusion: BMB provided the diagnosis of HL in 17% of cases, confirming its diagnostic utility in our setting. BMI by HL was more common in HIV+ patients and was associated with significantly worse survival. Our cohort showed similar survival outcomes to other countries in Africa, Asia and Central America with comparable socio-economic constraints to SA.

Splenectomy has been practiced for the treatment of ITP for the past few decades. Currently it is utilised when a patient is either dependent or resistant to steroid treatment and the platelet count remains less than 30×109/L. Recently new agents have been added to the armamentarium used to treat ITP, including immune-suppressants such as rituximab and the new thrombopoetin-receptor agonists. This has brought into question the role of surgery for the treatment of ITP, and the need to compare the response and complication rates of splenectomy to these newer agents. Historic studies done on splenectomy for the treatment of ITP have been performed in the setting of low HIV prevalence. There is a relative paucity of data on the response rate in HIV-associated thrombocytopenia to splenectomy and the durability of response to splenectomy is unclear in this patient population. We retrospectively analysed 73 consecutive patients who underwent splenectomy for ITP from 2001 to 2011. The primary objective was to determine the rate of complete response, this was defined as a platelet count greater than 100×109/L at one year post splenectomy. Results were compared between HIV positive and HIV negative patients. The secondary objectives were: to evaluate the intra-operative and post†operative complications and mortality in the HIV positive and HIV negative groups, and to investigate for associations between co-morbidities, pre-operative treatment and response to splenectomy.

This is a retrospective review of the results of consecutive bone marrow biopsies performed at our institution over a three year period on HIV positive patients for the investigation of fever and cytopenias. Clinical data, haematological parameters, morphology of bone marrow biopsy, Ziehl-Neelsen staining and microbiological culture results were analyzed. The aim of the study was to determine the diagnostic yield of this investigation.

Recovery following bone marrow transplantation is associated with the reduction in the clonogenic potential of stroma-adherent CD34⁺ progenitors. The bone marrow stroma is also affected, resulting in poor support for the growth of stroma-adherent multipotent progenitors that form blastic colonies (CFU-bl). To determine the possible mechanisms of marrow damage in patients treated with peripheral blood stem cells (PBSCs) and bone marrow transplantation (BMT), we studied the effects of bFGF, a cytokine known to stimulate the survival and proliferation of fibroblasts, on the propagation of clonogenic progenitors and the marrow stroma. METHODS: In order to obtain the optimal bFGF concentrations to be used on patients' haematopoietic progenitors and bone marrow rnicroenvironment, haematologically normal individuals were studied in dose response studies. Control mononuclear cells from the Ficoll-Histopaque interface layer were divided into two aliquots: one to establish a stromal layer and to culture fibroblastic progenitors (CFU-Fs) in the presence of 0, 2 and 20 ng/ml bFGF with and without 20 ng/ml heparan sulphate (HS). The second aliquot was for the selection of the progenitor population. Stroma was quantitated by placing culture dishes on a grid containing 1mm squares and scoring the number of squares occupied by stromal layers as the percentage area covered. After 3 weeks of culture, a single cell suspension was prepared by incubation of stroma with 5% trypsin solution, and number of cells in the dishes enumerated with a particle counter. CFU-Fs were terminated on the 9th day of culture, stained with May-Grilnwald-Giemsa and scored using an inverted microscope. From the second aliquot, CD34⁺ cells were incubated with paramagnetic beads and target cells isolated with a magnet. Selected l x l 04/ml cells were cultured on prefom1ed controls' stroma that has been treated with 0, 2 and 20 ng/ml bFGF. Stroma-adhered cells were covered with 0.3% agar and cultured for 6 days. Aggregates of more than 20 cells were counted as CFU-bl. RESULTS: In normal individuals, the median surface area of the petri dish covered by stroma at 3 weeks of culture was 55% (range 30-65) and was significantly improved upon the addition of 2 ng/ml (median 70%; range 50-95 ; p= &lt;0.05) and 20 ng/ml bFGF (median 80%; range 65-99; p= 0.004). Stromal cell numbers were 0.61 x 10⁶/2ml (range 0. 15-1.66), and they increased significantly with the addition of 2 and 20 ng/ml bFGF (p=0.03). The median colony forming unit-blasts (CFU-bl) scores were 121.8 x 10⁴/ml (range 43-271), and they expanded significantly with the addition of 20 ng/ml bFGF with and without heparan sulphate (p=0.03 and 0.01). It was then concluded that the 2 and 20 ng/ml bFGF with heparan sulphate be used on patients' cells in vitro. The median surface area covered by stromal layers in patients' samples at 3 weeks was 40% (range 0-55 vs. normal 50%), and it was in1proved significantly upon the addition of bFGF (median 78 vs. 50%; p= 0.001). Supplementation of stromal layers with bFGF accounted for a significant increase in patients stromal cell numbers (2.11 vs. normal 0.61 x 106/2ml; p&lt; 0.05). Patients' CD34⁺ cells panned on normal stromal layers resulted in significantly fewer CFU-bl (median 40 vs. normal 90 x 10⁴/ml CFU-bl; p=0.009), but CFU-bl numbers were corrected following the addition of bFGF, matching the scores achieved by normal individuals (85 vs. normal 90 x l 04/ml). Normal CD34⁺ cells proliferated poorly on patients' stromal layers in the absence of bFGF (3 9 vs. normal 90 x 10⁴/ml), but colony numbers increased significantly upon addition of bFGF (91 vs. normal 90 x 10⁴/ml CFU-bl). Thirteen patients receiving peripheral blood stem cells (PBSCs) and nine patients undergoing bone marrow transplantation (BMT) were compared. These two stem cell sources were then compared to the normal population. Stromal layers in patients receiving PBSC grafts covered a greater surface area than the area covered in patients undergoing BMT (median 65 vs. 30%). They also had higher blastic colony than scores from BMT recipients (47 vs. 12 x 10⁴/ml CFU-bl; p= 0.2). Results can be summarized that bone marrow stroma from patients receiving mobilized progenitor cells proliferates better than those receiving bone marrow grafts. It can be concluded that poor stromal layer and CD34⁺ cell proliferation following peripheral blood stem cell transplantation can be corrected by addition of basic Fibroblastic Growth Factor.

Includes bibliographical references. / The risk of pneumococcal infection post splenectomy is life long so all patients undergoing splenectomy are given polyvalent (23-valent) unconjugated pneumococcal polysaccharide vaccine, preoperatively. The aim of this study was 1. to measure the success of splenectomy at a tertiary institution at 5 years and beyond. 2. To review the incidence of complications peri-operatively and long term. 3. Review the need for possible re-vaccination (as recommended in American and British guidelines and 4. perform a re-appraisal of patients found to be refractory to treatment.

Background: Since its discovery about three decades ago, the Human Immunodeficiency Virus (HIV) has claimed over millions of lives globally. Although our understanding of the mode of transmission and action of this causative agent for the Acquired Immune Deficiency Syndrome (AIDS) has increased through research, and treatment regimens developed and improved, in certain parts of the world the pandemic continues to expand. Sub-Saharan Africa, which is the epicentre of this global health concern, accounts for approximately 66% of the total number of individuals affected, with South Africa enduring the heaviest burden. South Africa has the world's largest antiretroviral therapy (ART) programme and as such, HIV infected people are living longer, and consequently the incidence of HIV co-morbidities has increased dramatically. HIV/AIDS defining cancers are such co-morbidities with Non- Hodgkin's lymphomas (NHL) being the second most common HIV-associated cancer. Diffuse Large B-cell lymphoma (DLBCL) and Burkitt's lymphoma (BL) are the main subtypes and both present aggressively in HIV positive patients with rapid progression. The use of highly active antiretroviral therapy (HAART) has decreased the incidence of DLBCL in HIV positive patients, however the prevalence of these cancers still remain high in some settings. It has been suggested that the pathogenesis of these cancers in HIV infected individuals is complex and different to that in HIV uninfected individuals, with the possibility that the virus may have an oncogenic role. This has already been demonstrated in the case of the HIV/AIDSdefining cancer Kaposi Sarcoma. However, the same has not been unequivocally demonstrated in HIV-associated NHL. In light of this, the mechanisms through which viruses and viral components promote cellular transformation is an area of active research. One of these mechanisms manipulated by viruses is through the dysregulation of cellular microRNAs (miRNAs) which are small non-coding RNA molecules that are key regulators of gene expression. While they are essential for normal cellular functioning, their expression has been found to be deregulated in diseases including cancer. Several studies have described specific miRNA signatures for NHLs including for DLBCL and BL but none have been described for the HIV-association of these cancers. Aim: The aim of this project was to identify and characterise miRNAs involved in the pathogenesis of HIV-associated NHLs. This thesis reports on the changes in expression of miRNAs in B-cells exposed to an attenuated form (structurally intact but non-infectious) of HIV. Methods: We designed a custom miRNA microarray to identify deregulated miRNAs in the BL cell line Ramos that were exposed to HIV compared to microvesicle treated cells. It was initially planned to use both normal B-cells (L1439A) and BL cells for analysis but Ramos was selected due to technical reasons for this step. Thereafter we validated selected miRNAs by quantitative real-time PCR (qPCR) using single-tube TaqMan® Assays which was predominantly performed in the lymphoblastoid cell line L1439A, which is derived from a healthy donor. We then focused on further characterising the role of one miRNA in the development of HIV-associated NHL by using prediction programmes to predict its putative gene targets and then confirmed its target by using qPCR and western blot analyses. Results: Extensive and comprehensive analysis of the array data led to the identification of a large number of miRNAs which were differentially expressed, with 32 being selected for further studies. These 32 miRNAs include 16 upregulated and 16 downregulated miRNAs, and were selected because they displayed changes in expression by two or more folds. Thereafter, four miRNAs, namely miR-363-3p, miR-222-3p, miR-200c-3p and miR-575, were chosen for validation based on their reported involvement in cancer for validation. The results of two miRNAs (miR-575 (upregulated) (p<0.05) and miR-200c-3p (downregulated) (p<0.05)) were found to be consistent with the results obtained from the miRNA microarray whilst the other two were opposite to that result (both downregulated) (p<0.05). Using online tools as well as the published literature, several potential target genes of miR-575 were identified, namely DENND5A, CDK1, CSTA and ATAD5. One particular target, the BH3- like motif containing inducer of cell death (BLID), which is involved in apoptosis, has previously been confirmed as a gene target in non small cell lung cancer. Using qPCR, we found that BLID messenger RNA (mRNA) was downregulated in normal B-cells when exposed to HIV-1 AT-2. Unfortunately, the BLID protein could not be detected using western blot analysis despite several attempts at detecting varying concentrations of the protein and using two different positive control cell lines. Conclusion: The reverse correlation, between miR-575 and BLID mRNA expression in the same cell line and under the same treatment conditions, supports the notion that the downregulation of miR-575 may be physiologically relevant. However, this could not be further verified as the BLID protein could not be detected in the L1439A cells, even in the microvesicle treated control cells. Future studies should look at further characterisation of miR- 575 in the pathogenesis of HIV-associated NHLs by investigating other predicted gene targets of the miRNA. This will then be followed by loss and gain of function assays to confirm the miRNA:mRNA relationship. Furthermore, functional analyses, such as measure of apoptosis, expression of key regulators of the cell cycle, and other cellular events characteristic of cancer should be carried out to define the role of the miR-575 in the development of HIV-associated lymphoma.

Diffuse large B-cell lymphoma (DLBCL) is a highly heterogeneous and aggressive disease and is the most common subtype of non-Hodgkin lymphoma (NHL) in adults. Additionally, this subtype of lymphoma is also the most common in people infected with the Human Immunodeficiency virus (HIV), and the incidence has remained high despite the advent of Antiretroviral therapy (ART). About 30-40% of DLBCL patients' relapse, or are refractory to standard first-line therapy, and this is attributed to the high genetic and clinicopathological heterogeneity of the disease. Reports indicate that, among HIV-positive DLBCL patients, the response rate is even poorer. In low resourced settings, this is further aggravated by multiple factors including access to health facilities and gaps in communication. Recent genetic studies of DLBCL tumours allowed for the further subclassification of the ABC- and GCB- subtypes into at least 5 new groups, each with distinct genetic, molecular and clinicopathologic features, revealing potentially novel drivers of the disease. There is therefore a need to further understand the molecular pathology of these distinct subtypes, including within the context of HIV. The latter formed the basis of this study and used multiple approaches and methodologies, including immunophenotyping and mutational analysis of samples from local patient cohorts, as well as gene set enrichment analyses of publicly available DLBCL datasets. An analysis and comparison of immune cell populations in peripheral blood mononuclear cells, of HIV negative and HIV positive DLBCL patients was performed using flow cytometry (Chapter 3). The participants were newly diagnosed, chemotherapy naïve DLBCL patients. Some of the key observations were as follows: HIV-positive patients were diagnosed with DLBCL at significantly younger ages (75% under the age of 50 years), compared to the HIV negative DLBCL group. Furthermore, more extranodal disease and EBV infection were observed in the HIV-infected group, and both these factors are known to be indicative of more aggressive, advanced-stage disease. In general, cytopenias were observed in the DLBCL patient cohort, regardless of HIV status. Since most of the HIV infected patients were not adequately receiving antiretroviral therapy at the time of DLBCL diagnosis, the CD4+ helper T-cell population within this group was significantly reduced, in comparison to the HIVuninfected group. Interestingly, there was a significant difference in monocyte count between the two groups, with lower counts observed among the HIV-infected DLBCL patients. Additionally, increased activation of cytotoxic T-cells (CD8+CD38+), as well as lower mature cytolytic CD56dimCD16+ Natural Killer (NK) cells were observed in the HIV-positive DLBCL group. The prevalence of Myeloid differentiation primary response factor 88 (MYD88) L265P and Cluster of Differentiation 79B (CD79B) Y196 activating mutations were analysed in a cohort of archived ABC-DLBCL tumours (consisting of both HIV positives and negatives) (Chapter 4). Genomic DNA was isolated, the relevant genomic regions amplified by PCR, subjected to Sanger sequencing and then analysed and confirmed. The co-occurrence of both these mutations is characteristic of a newly described subset of DLBCL shown to have an inferior outcome. The prevalence of these mutations within an African population has as yet not been reported. The MYD88L265P mutation was detected in 26% of the amplified ABC-DLBCL tumours, while mutations at CD79BY196 were observed in 12% of the tumours. Co-occurrence of both these MYD88 and CD79B mutations were present in only 3 tumours, all of which were HIVnegative. Analysis of patient survival in relation to the mutations highlighted a trend showing that patients harbouring both mutations had worse overall survival. Interestingly, this pathogenic effect was more prominent for CD79B mutations, and this was comparable to the survival probability observed for HIV-positive patients. For the MYD88L265P mutation, an HIV positive status further decreased the survival probability. The final study presented in this thesis focused on investigating the expression and regulation of the Suppressor of cytokine signalling 1 (SOCS1) gene. SOCS1 has been recently reported to be a frequently altered gene in DLBCL, but molecular pathological studies on the role of SOCS1 in DLBCL are scarce. Using cell line models, basal SOCS1 gene and protein expression levels were assessed. In two DLBCL cell lines (SUDHL-4/GCB-DLBCL subtype and HBL-1/ABC-DLBCL subtype), SOCS1 expression was reduced at both the mRNA and protein levels when compared to control lymphoblastoid cell lines (LCLs), while in a third ABC-DLBCL cell line (U2932), results were varied. In silico analysis using the Cancer Genome Atlas (TCGA) database confirmed that SOCS1 is in the top 20 frequently mutated genes in DLBCL. Furthermore, these frequent mutations, which lead to reduced or low SOCS1 expression, are associated with DLBCL disease in its early stages (stages I and II) and with better survival estimates. In contrast, high expression of SOCS1 was associated with poor overall survival. This was further corroborated by gene set and pathway enrichment analyses, which highlighted factors involved in the enhancement of cancer-promoting processes such as proliferation, migration, invasion, and metastasis. Additionally, a previously reported association between the expression of methylation-related genes and SOCS1 expression was confirmed. Overall, these studies have uncovered novel insights into the pathology of DLBCL, including features unique to HIV-associated DLBCL. The implementation of a differential approach to the management of the disease, based on specific genetic and molecular features, should be explored. These findings support the importance of more studies, incorporating comprehensive genomic and molecular technologies, to continue to unravel the complex disease that is DLBCL.

Background: The recognised variation observed in normal haematological parameters in different populations and geographic locations, emphasizes the need to establish locally derived reference intervals (RIs) with appropriate representation of the various ethnic groups. Accurate RIs are essential to distinguish between health and disease. Objective: To establish locally derived RIs for full blood count (FBC) and white blood cell (WBC) differential count parameters in healthy adults in the Western Cape region of South Africa. Methods: A prospective, descriptive study was performed, utilizing blood samples of healthy first-time blood donors, presenting voluntarily for blood donation to the Western Cape Blood Service (WCBS) between November 2016 and October 2017. African, Coloured and Caucasian participants aged between 18 and 60 years of age were included based on convenience sampling. Participants testing positive for human immunodeficiency virus (HIV), hepatitis B and C viruses (HBV, HCV) and syphilis, and those with serum ferritin levels outside the reference range were excluded. Donors with an elevated serum ferritin were also excluded. Reference intervals were derived using non-parametric statistical methods and expressed to include the central 95% of the sample population (2.5th to 97.5th percentiles). Outliers for individual parameters were identified and excluded from the analysis. Results: A total of 376 females and 244 males were included for analysis; 31.61% were African, 39.68% Coloured and 28.71% Caucasian. For all race groups combined, gender-based differences were found in most FBC parameters, including the haemoglobin (Hb), WBC count, neutrophil count and platelet count. When comparing RIs for males and females in the three ethnic groups, statistically significant differences were found for parameters including the Hb, WBC count and red cell indices. There were no significant differences in the absolute eosinophil counts and mean cell volume (MCV) in females, and platelet counts in males. The ranges differed for a number of FBC variables compared to the National Health Laboratory Service (NHLS) coastal reference ranges in current use. Conclusion: Locally established and population-specific RIs are essential for accurate interpretation of blood counts. Implementation of separate RIs for the main ethnic groups in the Western Cape should be considered and would have implications for the diagnosis of anaemia and other blood count abnormalities as well as decision rules on haemoglobin levels for blood donor deferral.

No description available.