B-cell chronic lymphocytic leukaemia (CLL):

In lymph nodes, depletion of lymphoid follicles was observed. Immunohistochemically, most of the lym- phoid cells showed characteristic cell membrane CD79a positivity and negativity for CD3. The B-cell marker CD20, used in this study, does not cross-react with African lion lymphocytes. Furthermore, B-cells showed a constant co-expression of immunoglobulin M and immunoglobulin D on their surface mem- branes, which usually marks naïve B-cells. Indeed, the view that B-cell chronic lymphocytic leukaemia (B-CLL) cells are antigen-naïve and resting but ma- ture, is consistent with their appearance in tissue infiltrates, as small lymphocytes with a high nuclear- to-cytoplasmic ratio (Bennett et al. 1989). These im- munohistochemistry findings were observed in all organs, with particular evidence in the spleen, where Figure 1. Spleen of an African lion with chronic lympho- cytic leukaemia. (A) Neoplastic lymphocytes diffusely replacing most of the normal splenic cell population.

This thesis explores the possible cellular bases for these clinical responses and their heterogeneity, and addresses the adhesive properties of CLL cells, in relation their homing behaviour, as well as the cytokine regulation of CLL cell survival and apoptosis, and the role of effector cells, using both in vitro and in vivo studies. In a current clinical trial conducted in patients with early stage CLL, interferon-alpha produced partial responses, in terms of a significant reduction in lymphocyte counts, and the development of anti-interferon antibodies correlated with loss of clinical response. There was little evidence for T cell activation in patients receiving interferon, but there were significant rises in serum levels of macrophage colony-stimulating factor (M-CSF) and neopterin, suggesting that mononuclear cell activation may be involved in the generation of haem atological responses to in te rfe ro n -a lp h a .

The lower response rates observed amongst patients about to commence treatment are likely to reflect the immunosuppressive nature of CLL disease burden and reinforce the need for appropriate therapeutic vaccine regimens to be delivered at diagnosis [18]. Antibody response rates were lowest in those on active treat- ment. Bruton Tyrosine Kinase inhibitor (BTKi) therapy is known to suppress vaccine responses as a reflection of its inhibition of B cell activation [19, 20], although we found response rates to be 17% higher in patients on first line therapy compared to treatment in relapse. Antibody response rates in patients on BCL-2 inhibitor therapy were also low at 29% although the cohort size was rela- tively modest. CD20-specific antibody therapy is used widely for patients with B cell malignancies and mark- edly impairs antibody responses to vaccine challenge.

B cell chronic lymphocytic leukaemia (CLL) is the most common adult leukaemia that follows an extremely variable clinical course. Several important prognostic parameters defining pathogenic and clinical subgroups of CLL have been identified and validated recently. The biological significance of immunoglobulin (Ig) heavy chain variable region gene (IgHV) mutational status and associated ZAP-70 over-expression, CD38 and chromosomal aberrations have enabled to identify patients at high risk for early disease progression and inferior survival. Moreover, studies of the B cell antigen receptor (BCR) structure and receptor signalling have been most helpful in revealing some new aspects of the biology of this disease. In particular, the analysis of IG genes has revealed that the expressed IgHV/IgKV/IgLV gene repertoires of CLL cells differ from those of normal B cells. A further unique feature of the CLL IG repertoire is the existence of subsets of cases with “stereotyped” BCRs. Accumulating molecular and phenotypic data support the notion that CLL develop- ment and evolution is not a simple scholastic event and strongly indicates a role for antigen in driving the cell of origin for at least some subsets of CLL cases.

B cell chronic lymphocytic leukaemia (CLL) is the most common adult leukaemia that follows an extremely variable clinical course. Several important prognostic parameters defining pathogenic and clinical subgroups of CLL have been identified and validated recently. The biological significance of immunoglobulin (Ig) heavy chain variable region gene (IgHV) mutational status and associated ZAP-70 over-expression, CD38 and chromosomal aberrations have enabled to identify patients at high risk for early disease progression and inferior survival. Moreover, studies of the B cell antigen receptor (BCR) structure and receptor signalling have been most helpful in revealing some new aspects of the biology of this disease. In particular, the analysis of IG genes has revealed that the expressed IgHV/IgKV/IgLV gene repertoires of CLL cells differ from those of normal B cells. A further unique feature of the CLL IG repertoire is the existence of subsets of cases with “stereotyped” BCRs. Accumulating molecular and phenotypic data support the notion that CLL develop- ment and evolution is not a simple scholastic event and strongly indicates a role for antigen in driving the cell of origin for at least some subsets of CLL cases.

1. Executive summary Description of proposed service Fludarabine is a relatively recently developed chemotherapeutic agent, for which an oral formulation has recently become available. It is currently licensed for use in patients with B- cell chronic lymphocytic leukaemia patients with sufficient bone marrow reserve and who have not responded to or whose disease has progressed during or after treatment with at least one standard alkylating-agent containing regimen i.e. as a second line of treatment. In 2001, NICE issued guidance for England and Wales advising the use of oral fludarabine in these patients. This report considers the effectiveness and cost-utility of fludarabine (iv and oral) used as first-line treatment of B-cell CLL as an alternative to oral chlorambucil.

Various features of An1 cells reduce their participation in immune responses. One is reduced lifespan. The half-life of follicular B-cells is 40 days, as opposed to 5 days for anergic B-cells 242,243 . The reduced lifespan is only apparent when non-anergic B-cells are also present to compete for B-cell activating factor (BAFF), an important survival factor for peripheral B-cells 244 . Anergic cells appear to have reduced BAFF receptor expression or reduced BAFF receptor signalling competence 230 . As might be predicted, over-expression of BAFF leads to autoimmunity 229 . BAFF and its related molecule APRIL may play an important role in CLL cell survival 227,245 , yet the picture is confusing as BAFF and APRIL levels may be raised in CLL, yet the BAFF-Receptor levels are reduced and reports are conflicting. Other mechanisms of decreased survival involves high expression of the proapoptotic BCL family member BIM in anergic B-cells, and anergy is lost in anti-HEL mouse B-cells lacking BIM 246 . Anergic B-cells may also undergo apoptosis initiated by FAS signalling by encounter with FASL + CD4 + T-cells 247 . In contrast, it is well known that CLL cells exhibit a defect in apoptosis, with higher expression of anti-apoptotic BCL family members 248 . A model of CLL as an equivalent of anergic B-cells has been suggested 115,116,249 . It would be consistent with the concept of CLL cells as chronically binding autoantigens in vivo. One curious feature of both anergic and CLL cells is that they generally exhibit low expression of IgM whilst preserving the expression of IgD. Why this occurs is unknown, but if the BCR in CLL is to be studied, then it must be considered in both its expressed isotypes: IgM and IgD.

ABSTRACT Chronic Lymphocytic Leukaemia (CLL) is the most common adult Leukaemia in the UK, Western Europe and America. It is a malignancy of naïve B-cells. The clinical course of patients with CLL is heterogeneous; some patients survive for years without treatment, others die of a chemotherapy resistant disease within two years of presentation. Genomic studies have found little variation in patients showing differing prognosis, suggesting that it is the same disease but with varying outcomes. At present there is no cost effective, reliable and routine clinical test which can distinguish patient prognosis and a “watch and wait” strategy is currently in clinical use.

Chemotherapy There is no doubt that recently developed agents that target key pathogenic pathways in CLL will change the way that CLL is treated. However, while these targeted therapies are being implemented, chemotherapy alone, or more usually in combination, remains the standard upfront treatment for CLL worldwide. The cytotoxicity of chemotherapy drugs affects both tumour cells and T cells, with the dramatic loss of T cells an immediate effect of chemotherapy. For example, in patients initially treated with fludarabine both CD4 and CD8 T cells are lost, but there is proportionally a greater loss of CD4 T cells (Fenchel et al, 1995; Keating et al, 1998). Although there is a recovery in T cell numbers, the counts remain low (below normal levels for CD4) even 2 years after treatment (Keating et al, 1998). Despite this persistent T cell suppression, the overall risk of major infections in fludarabine-treated patients was low (<3%) and decreased with time after treatment (Keating et al, 1998). More detailed studies of T cell subsets in fludarabine-treated patients demonstrated a reduction in Treg frequency that was not seen with non-fludarabine treatment (Beyer et al, 2005).

Figure 5.3. T-cell proliferation after incubation with blinatumomab. PBMCs extracted from CLL patients were cultured in the presence or absence of blinatumomab (n=7). Intracellular Ki-67 was used as a marker for proliferating CD4 + and CD8 + T-cells. The number of positive cells was measured using 8-colour flow cytometry (FACS Canto ІІ). (A) FACS analysis showing the gating strategy used to separate Ki-67+/- T-cells (Left plot: untreated control, right plot: 10ng/ml blinatumomab treated). The percentage of Ki-67 expressing CD4 + (B) and CD8 + (C) T-cells was determined after incubation of PBMCs with blinatumomab (10ng/ml) or CD3/CD28 beads for 3 or 7 days in culture. *= (P<0.05), **= (P<0.01) and ***= (P<0.001). Statistical analysis was performed using a paired t-test.

What are the signs and symptoms of CLL? About 70-80% of all new cases of CLL are chance findings on a routine blood test, and between 40% and 60% of patients with CLL are free of symptoms at the time they are diagnosed. Almost all of these cases are discovered as a result of enlarged lymph nodes being noted by a doctor at a routine check-up or, more frequently, by abnormal results from routine blood tests. The number of asymptomatic cases discovered depends on the proportion of a given population having regular physical examinations or blood counts. It is common for there to be no symptoms at the time of diagnosis even when the white blood cell count is very high (>100 x10 9 /l).

proliferate further via chemokine production. This mechanism would induce a cycle of attraction and proliferation that would allow the clone to accumulate (Ghia et al. 2002). There is then, much evidence to suggest that in B CLL the T cells express abnormal surface markers that may in fact contribute to the promulgation of the disease and explain some of the commonly found clinical phenomena. These differences result in a T cell population that do not respond to stimulation in the same way as normal T cells and which fail to interact normally with B cells. Since effective T/B cell interaction is central to immune function, this presents major problems for maintaining a functional immune response.

Anti-CD52 (Alemtuzumab) Alemtuzumab is a fully humanized IgG1 monoclonal anti- body directed against the heavily glycosylated transmembrane glycoprotein, CD52. Unlike CD20, this antigen is not restrict- ed to B cells and is expressed also by T lymphocytes, granulocytes, monocytes and macrophages as well as NK and dendritic cells [ 31 ]. Alemtuzumab exerts its cytotoxic activity primarily through CDC [ 31 , 32 ] and ADCC [ 33 ] but has also been demonstrated to induce cell death via a direct mechanism that was independent of TP53 status [ 34 ], a find- ing which was later corroborated by clinical activity in this difficult-to-treat group with 17p deletion or TP53 mutations [ 35 , 36 ]. It appears to have the greatest efficacy in those pa- tients with greater circulating disease or bone marrow infiltra- tion, with poorer responses in those patients with bulky lymph nodes (LN) (>5 cm in particular). Whether this is the result of poor penetration of this molecule into nodal tissue or the result of impaired recruitment of an immune response due to lower effector cell density in LNs is postulated but remains

apparent that active participants in the constitution of CLL microenvironment comprises more than the cells of haemopoietic origin and that the MSC offer disease-defining interactions in their molecular interaction in the instance of incipient CLL. Genes regulated by MSC interactions with an importance for CLL pathogenesis include those relating to the PI3K signalling pathway, cellular metabolic and adhesion pathways such as XIAP, MCL-1 and BCL-2 (303). Evidence for the protective effects of the microenvironment upon CLL cells originates from study of the cells in the peripheral circulation to show increased Bim, indicative of their tendency for programmed cell death. BM resident CLL cells, however, demonstrate a significant reduction in Bim levels. Also, in stromal cell co-culture, CLL cells demonstrate increased BCL-2 and increased BCL2A1, both of which offer relative resistance to BCL-2 inhibition in vitro. When MSC are cultured in vitro with CLL there is an increased proportion of CD38-positive CLL cells which may represent a predilection for these cells to home and bind to MSC cells in the in vivo setting (105). MSCs within the CLL microenvironment also play a key role in protecting the CLL cells from oxidative stress and subsequent apoptosis and this reliance upon oxidative phosphorylation CLL may be exploited therapeutically (304). In turn, CLL cells appear to exert an influence upon MSC cell signalling with an increase in levels of AKT and PI3K activity. Evidence from other diseases such as lymphoma and solid tumour- associated fibroblasts suggest an upregulated signal via the NFkB pathway as a basis for these effects.

136 and studies of CD5 knockout mice showed that CD5 -/- T cells display increased proliferation, intracellular Ca 2+ release and tyrosine phosphorylation of PLCγ-1 and LAT all in response to TCR stimulation 220-221 . Therefore, it seemed logical to examine whether CD5 was a target of Lck in CLL cells. Our data show that BCR crosslinking on CLL cells results in an immediate phosphorylation of tyrosine residues within CD5 as detected by Western blot analysis of immunoprecipitated CD5 with anti-phosphotyrosine 4G10 and PY20 antibodies. We then examined whether this tyrosine phosphorylation was mediated by Lck by pre-treating the CLL cells with Lck-i. We found that such Lck inhibition did not affect BCR-induced phosphorylation of CD5, whereas the presence of the pan-SFK inhibitor dasatinib did. Thus, other SFKs such as Lyn may be responsible for the induction of CD5 phosphorylation in BCR-stimulated CLL cells. This seems to be consistent with what has been shown in CD5 + B1a cells 20 . In this subset of B cells BCR ligation results in induction of apoptosis, a response that is CD5 dependent because BCR crosslinking in CD5 knockout mice induced proliferation of these B cells rather than apoptosis. This study concluded that CD5 is a negative regulator of BCR signalling in B1 cells 218 , and is likely dependent on Lyn because BCR stimulation of Lyn -/- B1a cells also results in proliferation 219 . Whether CD5 operates in a similar manner in CLL cells is not fully understood. Studies by Perez-Chacon et al have shown that crosslinking CD5 can induce pro-survival signalling in CLL cells, but does so independently of the BCR 223 370 . However, another study has indicated that Lyn constitutively phosphorylates CD5 in CLL cells to allow association of SHP-1 where it functions as a negative regulator of BCR signalling 224 . Our results agree with a potential role of Lyn in the phosphorylation of CD5, but this only takes place when BCR is crosslinked in CLL cells.

The treatment options for chronic lymphocytic leukaemia patients with relapsed or unfavourable prognostic markers have greatly improved over the past few years. The introduction of two novel targeted therapeutics, the kinase inhibitors, Ibrutinib and Idelalisib, targeting the BCR signalling pathway has transformed treatment of patients with relapsed CLL. Prior to the development of these targeted agents, patients with relapsed disease, following chemoimmunotherapy or single agent immunotherapy, had a very poor prognosis and outcome. Despite the efficacy of these novel therapies, some patients fail to respond and others have inferior responses due to aggressive disease as a result of presence of poor prognostic indicators such as TP53 abnormalities (Farooqui et al. (2015), and others do not tolerate these treatments. In addition, resistance to Ibrutinib is already emerging leading to relapse (Woyach et al., 2017, Ahn et al., 2017). Resistance to Idelalisib in CLL has not yet been described in patients, although resistance in B-cell lymphoma cell lines has been shown (Iyengar et al., 2013, Yahiaoui et al., 2017). These inhibitors also carry a high economic burden in cost alone and amplified by the fact they are not a cure and most patients require the inhibitors over long periods. Therefore it is important to find alternative therapies for poor prognostic patient groups and non-responders that in addition may reduce the economic burden associated with these treatments.

7. Transplant in CLL 7.1 Autologous transplant is not recommended as part of standard care in CLL. 7.2 Allogeneic stem cell transplant may be offered as a consolidation therapy for all fit patients with high risk disease (fludarabine refractory/T53 abnormality) who have achieved remission. Where possible planning should allow this to be considered before their disease becomes treatment resistant.

The observations that patients with MRD after allogeneic transplantation may subsequently become MRD negative (Esteve et al, 2002), and the clinical benefit of donor lympho- cyte infusions, strongly suggests that the long-term disease-free survival achievable following allogeneic transplantation is immunologically mediated (Ritgen et al, 2002). These data, together with the high treatment-related mortality associated with standard allogeneic transplantation, have provided the impetus for studies using low intensity conditioning regimens. Treatment-related mortality is reduced but the non-relapse morbidity and mortality remains high in older patients with advanced disease. Disabling GVHD can result in considerable reduction in quality of life. The optimal conditioning regimen and approach to GVHD control is currently uncertain. In vivo T-cell depletion using Campath IH significantly reduces GVHD at the expense of a high incidence of CMV reactivation. In a study of 129 patients with lymphoproliferative disorders receiving a sibling non-myelo-ablative transplant and condi- tioning with fludarabine and melphalan, there was no differ- ence in event-free or overall survival between patients receiving either Campath 1H and ciclosporin A or methotrexate and ciclosporin A for GVHD prophylaxis (Perez-Simon et al, 2002). A recent overview of 77 low intensity transplants for CLL in Europe has shown a cumulative treatment-related mortality of 18% (95% CI 9–27) with event-free and overall survival at 24 months of 56% (CI 43–69) and 72% (CI 61–83) respectively (Dreger et al, 2003).

Some of these 20 proteins were linked to an important biological role in cancer. For example, pleckstrin homology domain-containing family A member 2 ( TAPP-2) was reported to participate in phosphatidylinositol 3- kinase (PI3K) signalling, an important pathway in B-lymphocyte activation and proliferation, following BCR stimulation (Marshall et al., 2002). Interestingly, TAPP-2 protein was reported to associate with poor prognosis CLL (ZAP-70 + CLL and U-CLL) (Costantini et al., 2009). In addition, zinc finger protein Aiolos was linked to apoptosis, where it induces BCL2 expression and prevents cell death in T-cells (Romero et al., 1999). Furthermore, Stathmin was shown to be involved in cell differentiation and proliferation and is highly expressed in a number of different types of cancer such as ovary cancer and breast cancer (Sherbet and Cajone, 2005, Price et al., 2000, Curmi et al., 2000) . This may indicate that this strategy is useful to highlight proteins with potential involvement in CLL using proteomics. Of these 20 proteins, 2 were selected on the basis of the quality of their mass spectrometry data for further study on normal B-cells and CLL cells (see chapter six).

Macrophage-mediated therapeutic antibody resistance in CLL. Similar to other cancers, inherent and/or acquired resistance to therapeutic antibodies is common in CLL, and in the relapsed setting, is a significant challenge in treating CLL patients [84, 85]. Recently, we and others have made significant progress in understanding the molecular mechanisms that drive resistance to therapeutic antibodies in CLL. Therapeutic monoclonal IgG antibodies bind to their cognate target on CLL cells to form an immune complex that is recognised by, and binds to, type I Fc receptors (FcγR) on macrophages. This interaction with a tumour-specific immune complex is the basis for the anti-tumour activity of macrophages. Thus, it is important to understand FcγR biology in the context of a major immune effector cell type, the macrophage, in order to understand resistance mechanisms. To date, six different classes of FcγRs (FcγRI, FcγRIIA, FcγRIIB, FcγRIIC, FcγRIIIA and FcγRIIIB) have been described. Activating FcγRs comprise FcγRI, IIa, IIc, IIIa and IIIb, and induce ADCC/ADP via a cascade of signalling events involving Lyn, Syk, PI3K and Btk activation/phosphorylation [67, 75, 86, 87]. In contrast, FcγRIIb inhibits ADCC/ADP tumour cell killing via activation of phosphatases such as Ship1/2, which catalyse the conversion of PIP3 to PIP2 thus blocking FcγR-dependent activation [88, 89] (Figure 3). Hence, FcγR-dependent events (e.g. ADCC or ADP) are controlled by the balance between activating (referred to collectively as ITAM) and inhibitory (referred to as ITIM) signalling pathways within the effector cells. Most importantly, ADCC and ADP responses to therapeutic antibodies are determined by the ratio of ITAM/ITIM signalling pathways [42, 77, 90]. For example, a low ITAM/ITIM ratio would reflect a poor response and a high ITAM/ITIM ratio would favour a good