Franz Josef Gassner

Difference in the relative distribution of CD4+ T-cell subsets in B-CLL with mutated and unmutated immunoglobulin (Ig) VH genes: implication for the course of disease.

B-cell chronic lymphocytic leukemia (B-CLL) is a clinically heterogeneous disease in which the clinical course is influenced by the presence or absence of immunoglobulin (Ig) variable heavy chain (VH) gene mutations. The poor clinical outcome of the subgroup with unmutated Ig VH genes has been linked to the persistent ability of the B-cell receptor in tumor cells from these cases to respond to antigen. As B-cell receptor signaling generally relies on T-cell help, we hypothesized that the course of B-CLL might not only be influenced by the Ig VH mutational status but also by the activation/differentiation status of T cells. We assessed the relative distribution of naive and memory T-cell subsets in peripheral blood from patients with mutated (M-CLL, n=71) and unmutated Ig VH genes (UM-CLL, n=42) and correlated it with the course of disease. We also compared the prosurvival potential of naive and memory T cells cocultured with B-CLL cells in vitro. A significant increase in relative numbers of central and effector memory T cells was observed in the CD4+ T-cell pool from UM-CLL as compared with M-CLL cases and was associated with high Rai stage, progressive disease and shorter treatment-free survival (TFS). In a multivariate analysis, the relative number of CD4+ central and effector memory T cells remained a significant prognostic parameter for TFS after correction for CD38 expression, Ig VH status, genomic aberrations, and Rai stage. The inverse correlation of memory CD4+ T cells with TFS might be explained by their potential to support survival of B-CLL cells.

Chemotherapy-induced augmentation of T cells expressing inhibitory receptors is reversed by treatment with lenalidomide in chronic lymphocytic leukemia

While T-cell dysfunction occurring alongside chronic lymphocytic leukemia (CLL) is well documented (reviewed by Pleyer et al. [1][1] and Hamblin et al. [2][2]), it was recently reported that also T-cell exhaustion is associated with CLL, reflected in the presence of PD-1+ CLL T cells and higher

Fludarabine modulates composition and function of the T cell pool in patients with chronic lymphocytic leukaemia

The combination of cytotoxic treatment with strategies for immune activation represents an attractive strategy for tumour therapy. Following reduction of high tumour burden by effective cytotoxic agents, two major immune-stimulating approaches are being pursued. First, innate immunity can be activated by monoclonal antibodies triggering antibody-dependent cellular cytotoxicity. Second, tumour-specific T cell responses can be generated by immunization of patients with peptides derived from tumour antigens and infused in soluble form or loaded onto dendritic cells. The choice of cytotoxic agents for such combinatory regimens is crucial since most substances such as fludarabine are considered immunosuppressive while others such as cyclophosphamide can have immunostimulatory activity. We tested in this study whether fludarabine and/or cyclophosphamide, which represent a very effective treatment regimen for chronic lymphocytic leukaemia, would interfere with a therapeutic strategy of T cell activation. Analysis of peripheral blood samples from patients prior and during fludarabine/cyclophosphamide therapy revealed rapid and sustained reduction of tumour cells but also of CD4+ and CD8+ T cells. This correlated with a significant cytotoxic activity of fludarabine/cyclophosphamide on T cells in vitro. Unexpectedly, T cells surviving fludarabine/cyclophosphamide treatment in vitro had a more mature phenotype, while fludarabine-treated T cells were significantly more responsive to mitogenic stimulation than their untreated counterparts and showed a shift towards TH1 cytokine secretion. In conclusion, fludarabine/cyclophosphamide therapy though inducing significant and relevant T cell depletion seems to generate a micromilieu suitable for subsequent T cell activation.

APOBEC3 signature mutations in chronic lymphocytic leukemia

Cytidine deaminases of the APOBEC family (ApoB mRNA editing catalytic subunit) generate targeted damage in nucleic acids by deaminating cytosins to uracils. The catalytically active family members are APOBEC1, APOBEC3 proteins (which comprise APOBEC3A, APOBEC3B, APOBEC3C, APOBEC3D, APOBEC3F, APOBEC3G, APOBEC3H) and activation-induced deaminase (AID) (reviewed in Conticello1, 2). APOBEC1 deaminates the mRNA for Apolipoprotein B thereby generating ApoB48, which mediates the absorption of dietary lipid from the intestine.3 Whereas APOBEC3 proteins are implicated in the natural defense against retroviruses and mobile elements by mutating and thereby inactivating retroviral single stranded DNA intermediates, AID mediates somatic hypermutation and class switch recombination of immunoglobulin genes in B cells.4 As it was shown that AID can cause substantial off-target DNA damage, AID has been implicated in B-cell lymphomagenesis by mutating oncogenes and by mediating chromosomal translocations as side product of aberrant class switch recombination.5 In chronic lymphocytic leukemia (CLL), previous reports have described the presence of AID transcripts in a subset of CLL patients where AID-expressing CLL patients exhibited a significantly decreased time to treatment and overall survival.6 Paradoxically, AID levels were higher in CLL cells expressing unmutated immunoglobulin variable domains (IgV-UM) encoding the B-cell antigen receptor, which predicts for poor prognosis, whereas CLL cells with somatically hypermutated IgV (IgV-Mut) tend to have lower AID transcript abundances.7 Also, AID expression associates with complex genomic aberrations and with Richters transformation to a more aggressive disease,8 providing strong evidence that AID is implicated in clonal evolution and pathogenesis in CLL. However, recent studies have revealed that other APOBEC family members are as well implicated in cancer development by inducing genome wide C>U deaminations that turn to C>T transition mutations after DNA replication.9 It was found that APOBEC-mediated mutations can lead to mutation showers (termed kataegis) in which multiple mutations are spaced by only one to several hundred nucleotides.10 In CLL, kataegic events were recently described that exhibit a signature of AID-mediated hypermutation.11 These clustered mutations are mostly but not exclusively found within the Ig locus and are restricted to IgV-Mut CLL cases.11 To compare clustered mutations inside and outside the Ig locus, we reanalyzed CLL whole genome sequencing data from Puente and coworkers, who sequenced two IgV-UM and two IgV-Mut CLL samples (Supplementary Table 6 from Puente et al.12). We thereby found that genome-wide mutations had a mean intermutational distance from 2.8 Mb (95% confidence interval: 2.698–2.937 Mb), which was similar for IgV-Mut and IgV-UM samples (Figure 1a). By searching for clustered mutations with intermutational distances below 10 kb, we noticed that while the two IgV-UM samples did not harbor mutation foci (defined as ⩾3 mutations each spaced by ⩽10 kb), the two IgV-Mut samples exhibited 19 clusters with a total of 113 (5.4% of total) mutations where ⩾3 mutations were spaced with an intermutational distance below 10 kb (Figure 1a). From these clusters, 49 mutations (43%) were located within the Ig locus and 64 (57%) were located outside the Ig locus (Figure 1a and Supplementary Table 1). In contrast to kataegis described in breast cancer,11 the clustered mutations found in CLL do not have a particular bias toward C>T or C>G mutations, neither at Ig nor at non-Ig clusters Supplementary Table 1), but rather have a mutation feature typical for somatic hypermutation where AID-mediated cytidine deamination is processed by error prone polymerases.4, 11 In line with this, C>T transitions within clustered mutations at the Ig loci were confined to the AID hot spot motif WRCY (W=A or T; R=A or G; Y=C or T; Figure 1c). In contrast, C>T transitions at mutation clusters outside the Ig locus were primarily found at TCW motifs, which is no AID but an APOBEC3 mutation motif (Figure 1c).10 Genome-wide unclustered C>T mutations outside mutation foci did not show a certain motif bias, neither when all CpGs or CpG islands were excluded from the analysis, as methylated cytosines (which are confined to CpG dinucleotides) are more prone to spontaneous deamination than normal cytosines13 (Figure 1b). Apart from the APOBEC3 motif at C>T transitions within clustered mutations at non-Ig loci, the spectrum of mutations was quite similar to hypermutation at Ig loci, with no apparent sequence bias (Supplementary Table 1). Thus, it is tempting to speculate that in CLL, APOBEC3 might operate in combination with factors of the hypermutation machinery to induce off-target DNA damage at non-Ig loci. Figure 1 Clustered mutations are present in IgV-Mut CLL samples. (a) Analysis of whole genome sequencing data by rainfall plots shows intermutational distances (IMDs) in four individual CLL cases with unmutated (CLL1 and 2) and mutated IgV (CLL3 and 4). Mutations ... As our mutation analysis revealed that C>T transitions within small mutation clusters occur at APOBEC3 recognition motifs outside the Ig loci in CLL, we next asked whether any APOBEC members are expressed in CLL samples. We therefore measured transcript levels of catalytically active members APOBEC3A to APOBEC3H using SYBR green quantitative reverse transcription (qRT)-PCR in a set of 10 IgV-Mut and 8 IgV-UM CLL samples and compared values with that from purified B cells of 5 healthy controls. As shown in Figure 2a, we found that APOBEC3A, APOBEC3B and APOBEC3H were expressed in CLL samples with a slight upregulation compared with healthy controls (Figure 2a). In addition, we observed a slight but significant difference in APOBEC3A, APOBEC3B and APOBEC3H expression values between IgV-Mut or IgV-UM samples (APOBEC3A: median 2.87 vs 1.92, P=0.016; APOBEC3B: median 1.37 vs 0.82, P=0.034; APOBEC3H: median 4.66 vs 2.77, P=0.021; Figure 2a). We subsequently tested protein expression of APOBEC3A, APOBEC3B and APOBEC3H by immunoblotting and found bands corresponding to APOBEC3A and APOBEC3B in almost any CLL sample irrespective of IgV mutation status (Figure 2b). Although median expression values of APOBEC3A and APOBEC3B were 1.5 and 1.7 times higher in IgV-Mut compared with IgV-UM CLL samples in qRT-PCR, this difference was too small to be discerned in immunoblotting on the level of protein. APOBEC3H, albeit showing highest expression in qRT-PCR, did not show detectable protein amounts in immunoblotting. A faint band for APOBEC3H that appeared in MEC1 cell lysates served as positive control (Figure 2b). Upon stimulation of CLL cells with CpG, APOBEC3B protein was upregulated, while APOBEC3A remained quite constant or even showed reduced band intensities in immunoblotting (Figure 2b). The APOBEC member AID was only detectable on in vitro stimulation of cells with CpG before cell lysis in one IgV-Mut and one IgV-UM sample, respectively (Figure 2b). Figure 2 APOBEC3 family members are expressed in CLL. (a) SYBR green qRT-PCR data from cDNA of 10 IgV-Mut, 8 IgV-UM CLL samples and 5 healthy controls (HD) using APOBEC3A-H (abbreviated: A3A-H)-specific primer sets normalized to glyceraldehyde 3-phosphate dehydrogenase ... Overall, our data show that aside of AID, also APOBEC3 deaminases are likely implicated in mediating off-target mutations in CLL. We show that APOBEC3A and APOBEC3B are more abundant on protein level in CLL than in AID, and clustered genomic C>T mutations outside the Ig locus occur at APOBEC3 hot spot motifs in IgV-Mut CLL. As these clustered mutations were confined to IgV-Mut samples, we initially suspected that APOBEC3 members were differentially expressed according to the IgV mutation status. However, although APOBEC3A, APOBEC3B and APOBEC3H transcripts were significantly lower expressed in IgV-UM samples, it is questionable whether this small difference has biological significance, as APOBEC3A and APOBEC3B were easily detectable on the protein level in both IgV-Mut as well as IgV-UM samples. However, as it was shown that APOBEC3 levels correlate with APOBEC3-induced mutations,14 it is conceivable that even small differences in APOBEC3 expression levels could account for the observed difference in APOBEC3-induced clustered mutations in IgV-Mut vs IgV-UM CLL (Figure 2a). Alternatively, as recent studies have revealed that apart from APOBEC expression additional factors of the DNA repair machinery10 are required for inducing kataegis, our data could suggest that any of these factors are not expressed in IgV-UM samples, similar to the observation that AID expression in IgV-UM CLL samples is not sufficient to diversify IgV genes.7 The absence of any of these factors could either alter the targeting of APOBEC3 to the DNA or otherwise affect the error prone DNA repair of APOBEC3-induced DNA lesions. Hence, it is conceivable that APOBEC3 might operate non-processively in IgV-UM CLL, resulting in the generation of unclustered mutations as opposed to clustered kataegic events in IgV-Mut CLL. From our data we conclude that aside of AID, also APOBEC3 should be considered as a source for genomic mutations in CLL. As CLL has a substantial intraclonal heterogeneity,15 we suspect that ongoing APOBEC3 expression might contribute to this genetic complexity by continuously increasing the mutation load, thereby accelerating clonal evolution to a more aggressive or chemorefractory disease. We further propose that analyzing IgV-Mut vs IgV-UM CLL will likely yield a deeper insight into mechanisms that underlie the initiation of APOBEC3-mediated clustered vs unclustered DNA mutations in cancer in general.

Chronic lymphocytic leukaemia induces an exhausted T cell phenotype in the TCL1 transgenic mouse model

Although chronic lymphocytic leukaemia (CLL) is a B cell malignancy, earlier studies have indicated a role of T cells in tumour growth and disease progression. In particular, the functional silencing of antigen‐experienced T cells, called T cell exhaustion, has become implicated in immune evasion in CLL. In this study, we tested whether T cell exhaustion is recapitulated in the TCL1tg mouse model for CLL. We show that T cells express high levels of the inhibitory exhaustion markers programmed cell death 1 (PDCD1, also termed PD‐1) and lymphocyte‐activation gene 3 (LAG3), whereas CLL cells express high levels of CD274 (also termed PD‐ligand 1). In addition, the fraction of exhausted T cells increases with CLL progression. Finally, we demonstrate that exhausted T cells are reinvigorated towards CLL cytotoxicity by inhibition of PDCD1/CD274 interaction in vivo.

AID induces intraclonal diversity and genomic damage in CD86+ chronic lymphocytic leukemia cells

The activation‐induced cytidine deaminase (AID) mediates somatic hypermutation and class switch recombination of the Ig genes by directly deaminating cytosines to uracils. As AID causes a substantial amount of off‐target mutations, its activity has been associated with lymphomagenesis and clonal evolution of B‐cell malignancies. Although it has been shown that AID is expressed in B‐cell chronic lymphocytic leukemia (CLL), a clear analysis of in vivo AID activity in this B‐cell malignancy remained elusive. In this study performed on primary human CLL samples, we report that, despite the presence of a dominant VDJ heavy chain region, a substantial intraclonal diversity was observed at VDJ as well as at IgM switch regions (Sμ), showing ongoing AID activity in vivo during disease progression. This AID‐mediated heterogeneity was higher in CLL subclones expressing CD86, which we identified as the proliferative CLL fraction. Finally, CD86 expression correlated with shortened time to first treatment and increased γ‐H2AX focus formation. Our data demonstrate that AID is active in CLL in vivo and thus, AID likely contributes to clonal evolution of CLL.

Alternative splice variants of AID are not stoichiometrically present at the protein level in chronic lymphocytic leukemia

Activation‐induced deaminase (AID) is a DNA‐mutating enzyme that mediates class‐switch recombination as well as somatic hypermutation of antibody genes in B cells. Due to off‐target activity, AID is implicated in lymphoma development by introducing genome‐wide DNA damage and initiating chromosomal translocations such as c‐myc/IgH. Several alternative splice transcripts of AID have been reported in activated B cells as well as malignant B cells such as chronic lymphocytic leukemia (CLL). As most commercially available antibodies fail to recognize alternative splice variants, their abundance in vivo, and hence their biological significance, has not been determined. In this study, we assessed the protein levels of AID splice isoforms by introducing an AID splice reporter construct into cell lines and primary CLL cells from patients as well as from WT and TCL1tg C57BL/6 mice (where TCL1 is T‐cell leukemia/lymphoma 1). The splice construct is 5′‐fused to a GFP‐tag, which is preserved in all splice isoforms and allows detection of translated protein. Summarizing, we show a thorough quantification of alternatively spliced AID transcripts and demonstrate that the corresponding protein abundances, especially those of splice variants AID‐ivs3 and AID‐ΔE4, are not stoichiometrically equivalent. Our data suggest that enhanced proteasomal degradation of low‐abundance proteins might be causative for this discrepancy.

CD1d expression on chronic lymphocytic leukemia B cells affects disease progression and induces T cell skewing in CD8 positive and CD4CD8 double negative T cells

Chronic lymphocytic leukemia develops within a complex network driven by genetic mutations and microenvironmental interactions. Among the latter a complex interplay with the immune system is established by the clone. Next to a proposed recruitment of support from T and myeloid cells, potential anti-CLL immune reactions need to be subverted. By using TCL1 mice as a CLL model, we show that TCR-Vβ7+ NK1.1+ T cells are overrepresented in this disease model and constitute a main subset of peripheral CD3+ cells with biased TCR usage, showing that these cells account for a major part for T cell skewing in TCL1 mice. Moreover, we show that overrepresentation is dependent on CD1d expression in TCL1 mice, implicating that these cells belong to a NKT-like cell fraction which are restricted to antigen presented by the MHC-like surface marker CD1d. Accordingly, we observed a high fraction of CD161+ cells within overrepresented T cells in CLL patients and we found downregulation of CD1d on the surface of CLL cells, both in TCL1 mice and patients. Finally, we show that in TCL1 mice, CD1d deficiency resulted in shortened overall survival. Our results point to an interaction between CLL and CD161+ T cells that may represent a novel therapeutic target for immune modulation.

B cell receptor usage correlates with the sensitivity to CD40 stimulation and the occurrence of CD4+ T cell clonality in chronic lymphocytic leukemia

The mutation status of the B-cell receptor (BCR) is a strong prognostic factor in chronic lymphocytic leukemia (CLL), dividing patients into BCR mutated and unmutated CLL (CLL-Mut, CLL-UM), the latter predicting for worse prognosis.[1][1]–[3][2] Furthermore, the occurrence of highly similar BCRs

Lysine residue at position 22 of the AID protein regulates its class switch activity.

Background Activation induced deaminase (AID) mediates class switch recombination and somatic hypermutation of immunoglobulin (Ig) genes in germinal centre B cells. In order to regulate its specific activity and as a means to keep off-target mutations low, several mechanisms have evolved, including binding to specific cofactors, phosphorylation and destabilization of nuclear AID protein. Although ubiquitination at lysine residues of AID is recognized as an essential step in initiating degradation of nuclear AID, any functional relevance of lysine modifications has remained elusive. Methodology/Principal Findings Here, we report functional implications of lysine modifications of the human AID protein by generating a panel of lysine to arginine mutants of AID and assessment of their catalytic class switch activity. We found that only mutation of Lys22 to Arg resulted in a significant reduction of class switching to IgG1 in transfected primary mouse B cells. This decrease in activity was neither reflected in reduced hypermutation of Ig genes in AID-mutant transfected DT40 B cell lines nor recapitulated in bacterial deamination assays, pointing to involvement of post-translational modification of Lys22 for AID activity in B cells. Conclusions/Significance Our results imply that lysine modification may represent a novel level of AID regulation and that Lys22 is important for effective AID activity.