Search

Novel therapies with chimeric antigen receptor (CAR)-transduced T cells (TCs) sparked new hope for patients with relapsed or refractory CD19-positive leukemia or lymphoma even after stem cell therapies. This review focuses on CARs recognizing the B cell antigen CD19. Both retroviral and lentiviral vectors are used, encoding various anti-CD19 CAR constructs comprising costimulatory molecules such as CD28, CD137/4-1BB, and OX40 either alone (second-generation CARs) or in combination (third-generation CARs). Current, up-to-date published studies on anti-CD19 CAR therapy for acute lymphoblastic leukemia (ALL), chronic lymphocytic leukemia (CLL), and non-Hodgkin lymphoma (NHL) with observed side effects are discussed and an outlook on 58 ongoing trials is given. Clinical responses were achieved in up to 81% of ALL, 50% of CLL, and 40% of NHL patients. Factors with potential influence on the clinical outcome might be the design of the vector, the preconditioning regimen, and the number and quality of transfused CAR TCs. The applicability of clinical CAR TC therapy might include relapse after allogeneic stem cell transplantation (alloSCT), and ineligibility for or "bridging" until alloSCT. In summary, CAR therapy represents a highly promising treatment option even in heavily pretreated patients.

International audience; Clinical information about thiotepa-based autologous stem cell transplantation (auto-SCT) outside the primary central nervous system lymphoma (PCNSL) field is sparse. In this registry-based retrospective study, we evaluated potential risks and benefits of thiotepa-based preparative regimens compared with BEAM (carmustine, etoposide, cytarabine, melphalan) in auto-SCT for diffuse large B-cell lymphoma (DLBCL, excluding PCNSL), follicular lymphoma (FL) or Hodgkin lymphoma (HL). A total of 14 544 patients (589 thiotepa and 13 955 BEAM) met the eligibility criteria, and 535 thiotepa- and 1031 BEAM-treated patients were matched in a 1:2 ratio for final comparison. No significant differences between thiotepa and BEAM groups for any survival end point were identified in the whole sample or disease entity subsets. For a more detailed analysis, 47 TEAM (thiotepa, etoposide, cytarabine, melphalan)-treated patients were compared with 75 matched BEAM patients with additional collection of toxicity data. Again, there were no significant differences between the two groups for any survival end point. In addition, the frequency of common infectious and non-infectious complications including secondary malignancies was comparable between TEAM and BEAM. These results indicate that thiotepa-based high-dose therapy might be a valuable alternative to BEAM in DLBCL, HL and FL. Further evaluation by prospective clinical trials is warranted

Chronic lymphocytic leukemia (CLL) shows a heterogeneous clinical course, which can be explained in part by prognostic factors. Most patients do not need treatment at the time of first diagnosis. The identification of prognostic factors is of major interest if strategies can be devised to treat patients according to their individual risk profile or biological subgroup. Currently, in spite of a wealth of prognostic factors, individualized treatment approaches in different genetic or risk groups are the exemption in CLL. This review summarizes the most important prognostic and predictive factors in CLL, with particular emphasis on factors affecting treatment decisions in clinical trials and routine practice.

Gene transfer into chronic myelogenous leukemia (CML) cells may become of relevance for overcoming therapy resistance. Single-stranded pseudotyped adeno-associated viruses of serotypes 2/1 to 2/6 (ssAAV2/1-ssAAV2/6) were screened on human CML cell lines and primary cells to determine gene transfer efficiency. Additionally, double-stranded self-complementary vectors (dsAAVs) were used to determine possible second-strand synthesis limitations. On human CML cell lines, ssAAV2/2 and ssAAV2/6 were most efficient. On primary cells, ssAAV2/6 proved significantly more efficient (4.1 ± 2.5% GFP(+) cells, p = 0.011) than the other vectors (

Background and aims. Because of their pluripotency, human CD34 peripheral blood progenitor cells (PBPC) are targets of interest for the treatment of many acquired and inherited disorders using gene therapeutic approaches. Unfortunately, most current vector systems lack either suffi cient transduction effi ciency or an appropriate safety profi le. Standard single-stranded recombinant adeno-associated virus 2 (AAV2)-based vectors offer an advantageous safety profi le, yet lack the required effi ciency in human PBPC. Methods. A panel of pseudotyped AAV vectors (designated AAV2/x, containing the vector genome of serotype 2 and capsid of serotype x, AAV2/1–AAV2/6) was screened on primary human granulocyte–colony-stimulating factor (G-CSF)-mobilized CD34 PBPC to determine their gene transfer effi cacy. Additionally, double-stranded selfcomplementary AAV (dsAAV) were used to determine possible second-strand synthesis limitations. Results. AAV2/6 vectors proved to be the most effi cient [12.8% (1.8–25.4%) transgene-expressing PBPC after a single transduction], being signifi cantly more effi cient (all P0.005) than the other vectors [AAV2/2, 2.0% (0.2–7.3%); AAV2/1, 1.3% (0.1–2.9%); others, 1% transgene-expressing PBPC]. In addition, the relevance of the single-to-double-strand conversion block in transduction of human PBPC could be shown using pseudotyped dsAAV vectors: for dsAAV2/2 [9.3% (8.3–20.3%); P0.001] and dsAAV2/6 [37.7% (23.6–61.0%); P0.001) signifi cantly more PBPC expressed the transgene compared with their singlestranded counterparts; for dsAAV2/1, no signifi cant increase could be observed. Conclusions. We have shown that clinically relevant transduction effi ciency levels using AAV-based vectors in human CD34 PBPC are feasible, thereby offering an effi cient alternative vector system for gene transfer into this important target cell population.

Complement-dependent cytotoxicity (CDC) is one of the effector mechanisms mediated by therapeutic anticancer monoclonal antibodies (mAbs). However, the efficacy of antibodies is limited by the resistance of malignant cells to complement attack, primarily due to the over-expression of one or more membrane complement regulatory proteins (mCRPs) CD46, CD55, and CD59. CD20-positive Burkitt lymphoma Raji cells and primary CLL cells are resistant to rituximab (RTX)-induced CDC whereas ofatumumab (OFA) proved to be more efficient in cell killing. Primary CLL cells but not CD52-positive acute lymphoblastic leukemia (ALL) REH cells were sensitive to alemtuzumab (ALM)-induced CDC. Upon combined inhibition on Raji and CLL cells by mCRPs-specific siRNAs or neutralizing antibodies, CDC induced by RTX and by OFA was augmented. Similarly, CDC of REH cells was enhanced after mCRPs were inhibited upon treatment with ALM. All mAbs induced C3 opsonization, which was significantly augmented upon blocking mCRPs. C3 opsonization led to enhanced cell-mediated cytotoxicity of leukemia cells exposed to PBLs or macrophages. Furthermore, opsonized CLL cells were efficiently phagocytized by macrophages. Our results provide conclusive evidence that inhibition of mCRPs expression sensitizes leukemic cells to complement attack thereby enhancing the therapeutic effect of mAbs targeting leukemic cells.

Allogeneic stem cell transplantation (alloSCT) is used for treating patients with T-prolymphocytic leukemia (T-PLL). However, direct evidence of GvL activity in T-PLL is lacking. We correlated minimal residual disease (MRD) kinetics with immune interventions and T-cell receptor (TCR) repertoire diversity alterations in patients after alloSCT for T-PLL. Longitudinal quantitative MRD monitoring was performed by clone-specific real-time PCR of TCR rearrangements (n=7), and TCR repertoire diversity assessment by next-generation sequencing (NGS; n=3) Although post-transplant immunomodulation (immunosuppression tapering or donor lymphocyte infusions) resulted in significant reduction (>1 log) of MRD levels in 7 of 10 occasions, durable MRD clearance was observed in only two patients. In all three patients analyzed by TCR-NGS, MRD responses were reproducibly associated with a shift from a clonal, T-PLL-driven profile to a polyclonal signature. Novel clonotypes that could explain a clonal GvL effect did not emerge. In conclusion, TCR-based MRD quantification appears to be a suitable tool for monitoring and guiding treatment interventions in T-PLL. The MRD responses to immune modulation observed here provide first molecular evidence for GvL activity in T-PLL which, however, may be often only transient and reliant on a poly-/oligoclonal rather than a monoclonal T-cell response.

Introduction The advent of T cells expressing chimeric antigen receptors (CART) for the treatment of cancer patients has been a milestone in the field of immunotherapy. For clinical application, reproducibility and safe generation of CART must be guaranteed. We investigated how different culture conditions of human peripheral blood mononuclear cells (PBMCs) influence the phenotype, i.e. naïve (N), central / effector memory (CM/EM), effector (E) cells, and the cellular composition of the resulting CART preparation, among others: the percentage of CD56+, γ/δ TCR+ and CD4+CD8+ double positive cells. Eventually, freezing/thawing of CART might also have an impact on their number, phenotype and function. Methods CART were generated by transducing human PBMCs from healthy donors with a CD19.CAR-CD28-CD137zeta vector (kindly provided by the Center for Cell and Gene Therapy, Houston) under two different stimulating culture conditions - anti-CD3/anti-CD28 antibodies with the addition of either IL-7/IL-15 or IL-2. CART generation lasted for 15 days. Cytotoxic ability of the generated CART was assessed by a standard chromium (Cr-51) release assay. CD19+ (Daudi) and CD19- (K562) cells were labeled for 2h with Cr-51 and coincubated with CART for 4h. Cr-51 release was measured in a liquid scintillation counter. In addition, multi-parametric flow cytometry was performed using a FACS LSR device. Results Overall, the transduction rate for both cytokine cocktails was around 50% of all CD3+ T cells. Freezing/thawing of CART reduced their cytotoxic activity: Lysis of the CD19+ target cell line (Daudi) reached a maximum of 80% for fresh cells and 50% for freshly thawed cells, both stimulated with IL-7/IL-15. FACS analysis revealed that freshly prepared cells showed an increased percentage of activated lymphocytes compared to cryopreserved cells. The majority of these activated lymphocytes were CD4+CD8+ double positive. Using freshly thawed cells, CART stimulated with IL-7/IL-15 showed higher cytotoxic activity than CART stimulated with IL-2: 50% vs. 38% (Daudi) and 7% vs. 5% (K562). Stimulation with IL-7/IL-15 led to an increase in the CD56+ CAR NK T cell population (20%) compared to IL-2 (8%). γ/δ TCR+ CART differed as well (3% for IL-7/IL-15, 8% for IL-2). The percentage of CD8+ and CD4+ CART was similar for both cytokine cocktails with around 60% and 30%, respectively. However, stimulation with IL-7/IL-15 showed a trend towards generation of a lower percentage of CD4+CD8+ double positive CART than stimulation with IL-2 (2% vs. 6%). The analysis of the different CART phenotypes showed significant differences: N 24% vs. 3%, CM 21% vs. 13%, EM 24% vs. 77%, E 31% vs. 7%, for IL-7/IL-15 vs. IL-2, respectively. A general phenomenon observed for both cytokine cocktails, was the difference in phenotype for CD4+ vs. CD8+ CART: CD4+ CART showed a more distinct memory phenotype, i.e. central and effector memory cells. Correspondingly, the CD8+ CART shifted more towards the naïve and effector phenotype. The percentage of highly effective stem cell memory CART (TSCM: naïve CD27+CD95+) was assessed. Using IL-7/IL-15, CAR TSCM cells accounted for 4.8%, whereas for IL-2, 2.7% had a TSCM phenotype. 75% of the CART were HLA-DR positive (activation marker) and 100% were positive for CD95 (FAS receptor, inhibition marker). Using PBMCs from healthy donor samples, the generated CART products did not contain any contamination of either B (CD19+) or stem cells (CD34+). Conclusions We have established a combination of immunophenotyping and cytotoxicity assays for CART as a preclinical potency assay. Stimulation with IL-7/IL-15 led to a balanced phenotype expression, with a similar number of N, CM, EM and E CART, whereas the majority of CART under IL-2 stimulation adopted an EM phenotype. Correlation of these data with the clinical outcome of patients receiving the corresponding cell products will allow optimization and standardization of CART therapy. Disclosures Wuchter: Sanofi-Aventis: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau; Hexal: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees. Dreger:Gilead: Consultancy; Novartis: Consultancy; Janssen: Consultancy; Gilead: Speakers Bureau; Novartis: Speakers Bureau; Roche: Consultancy.

Aminoferrocene-based prodrugs are activated under cancer-specific conditions (high concentration of reactive oxygen species, ROS) with the formation of glutathione scavengers (p-quinone methide) and ROS-generating iron complexes. Herein, we explored three structural modifications of these prodrugs in an attempt to improve their properties: (a) the attachment of a -COOH function to the ferrocene fragment leads to the improvement of water solubility and reactivity in vitro but also decreases cell-membrane permeability and biological activity, (b) the alkylation of the N-benzyl residue does not show any significant affect, and (c) the attachment of the second arylboronic acid fragment improves the toxicity (IC50) of the prodrugs toward human promyelocytic leukemia cells (HL-60) from 52 to 12 μM. Finally, we demonstrated that the prodrugs are active against primary chronic lymphocytic leukemia (CLL) cells, with the best compounds exhibiting an IC50 value of 1.5 μM. The most active compounds were found to not affect mononuclear cells and representative bacterial cells.

Sensitive identification of mutations in genes related to the pathogenesis of cancer is a prerequisite for risk-stratified therapies. Next-generation sequencing (NGS) in lymphoma has revealed genetic heterogeneity which makes clinical translation challenging. We established a 454-based targeted resequencing platform for robust high-throughput sequencing from limited material of patients with lymphoma. Hotspot mutations in the most frequently mutated cancer consensus genes were amplified in a two-step multiplex-polymerase chain reation (PCR) which was optimized for homogeneous coverage of all regions of interest. We show that targeted resequencing based on NGS technologies allows highly sensitive detection of mutations and assessment of clone size. The application of this or similar techniques will help the development of genotype-specific treatment approaches in lymphoma.

Recurrent gene mutations contribute to the pathogenesis of chronic lymphocytic leukaemia (CLL). We developed a next-generation sequencing (NGS) platform to determine the genetic profile, intratumoural heterogeneity, and clonal structure of two independent CLL cohorts. TP53, SF3B1, and NOTCH1 were most frequently mutated (16.3%, 16.9%, 10.7%). We found evidence for subclonal mutations in 67.5% of CLL cases with mutations of cancer consensus genes. We observed selection of subclones and found initial evidence for convergent mutations in CLL. Our data suggest that assessment of (sub)clonal structure may need to be integrated into analysis of the mutational profile in CLL.

T-cell prolymphocytic leukemia (T-PLL) is a rare disease with an aggressive clinical course and a median overall survival of less than three years. Although almost 75% of T-PLL patients are reported to harbor translocations causing the activation of the proto-oncogene TCL1A, T-PLL is genetically heterogenous: most T-PLL patients also have mutations or deletions in the ATM gene and the genes involved in the JAK-STAT pathway are mutated in 76% of cases. There is an urgent need for more rational based therapies, but clinical trials are difficult to perform due to the rareness of the disease. Here, we systematically explored the diversity of drug responses in T-PLL patient samples ex vivo using a drug sensitivity and resistance testing (DSRT) system including 306 oncology drugs (approved or investigational). We also aimed to determine any associations between the genetic aberrations and drug sensitivities in T-PLL patients. Primary mononuclear cells were gathered from 30 T-PLL patients for drug testing. Cells were plated in 384-well plates and subjected to the 306 substances using a 10,000-fold concentration range. After 72 hours, cell viabilities were measured, the results were depicted as dose-response curves for each compound, and differential drug sensitivity scores (sDSS), representing leukemia-specific responses, were computed by comparing patient samples to healthy donors. Drug response profiles across patients were clustered and visualized by hierarchical clustering. The subgroups resulting from the clustering were statistically compared using a two-sample t-test to understand which drug classes were driving the grouping. To delineate heterogeneous pathway dependencies, drug sensitivities were correlated with somatic genetic variants and recurrent chromosomal aberrations. Genetic characterization was performed by targeted amplicon sequencing of tumor cells to profile known recurrent genetic variants (STAT5b, IL2RG, JAK1, JAK3, ATM). Information on chromosomal aberrations (TCL1A translocations, ATM deletions) was derived from parallel clinicopathologic databases. Amplicon sequencing revealed that 70% of T-PLL patients (21/30) harbored a mutation in genes involved in the JAK-STAT pathway (JAK1, JAK3, STAT5b or IL2RG). The most prevalent mutation led to an M511I amino acid exchange in the JAK3 protein (26% of patients). Interestingly, the STAT5b mutations (5/30) did not coexist with any of the JAK mutations in our cohort. Based on DSRT analysis, all T-PLL samples were sensitive to the CDK-inhibitor SNS-032 and the anti-cancer antibiotic actinomycin D. Next, we clustered patients using sDSS values for all 306 different substances, and this showed that patient samples could be divided into 3 main groups, based on their drug responses (Figure). According to two-sample t-test, the grouping was driven by the selective sensitivities of Groups II and III to HDAC inhibitors (belinostat, panobinostat, quisinostat, CUDC-101 and vorinostat) and the selective sensitivity of Group III to PI3K/AKT/mTOR inhibitors (AZD-8055, MK-2206, apitolisib, dactolisib, PF-04691502, ZSTK474, and omipalisib), HSP90 inhibitors (BIIB021, luminespib, alvespimycin, and tanespimycin) as well as JAK inhibitors (ruxolitinib, momelotinib, tofacitinib, gandotinib). Group I samples were on the other hand relatively resistant to these classes of drugs. Surprisingly, despite the prevalence of the signature event of activation of TCL1 (the established AKT coactivator) in nearly all cases, only a subset of cases (group III) responded to PI3K/AKT/mTOR inhibitors. Strikingly, the grouping of selective responses to HDAC, JAK, PI3K/mTOR/Akt and HSP90 inhibitors did not link to the presence of JAK/STAT mutations, TCL1A translocations, or ATM deletion status. Ex vivo drug screening of primary T-PLL samples revealed heterogenous selective drug responses in specific drug classes (such as HDAC-, JAK-, HSP90- and PI3K/Akt/mTOR-inhibitors). Surprisingly, the drug response patterns did not correlate with known recurrent genetic aberrations suggesting that sequencing for recurrent genetic biomarkers cannot easily be turned into effective therapeutic strategies in T-PLL, and that further elucidation of the biological pathways driving T-PLL is needed. Figure 1. Mutation status and clustering of HDAC-, PI3K/mTOR/Akt-, HSP90-, and JAK-inhibitor responses in PLL patients based on sDSS values Figure 1. Mutation status and clustering of HDAC-, PI3K/mTOR/Akt-, HSP90-, and JAK-inhibitor responses in PLL patients based on sDSS values Disclosures Koschmieder: Novartis: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: Travel reimbursement for scientific conferences, Research Funding; Novartis Foundation: Research Funding; Baxalta/CTI: Membership on an entity's Board of Directors or advisory committees; Sanofi: Membership on an entity's Board of Directors or advisory committees; Janssen Cilag: Other: Travel reimbursement for scientific conferences ; Pfizer: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: Travel reimbursement for scientific conferences. Wennerberg:Pfizer: Honoraria, Research Funding. Ding:Merek: Research Funding. Mustjoki:Pfizer: Honoraria, Research Funding; Bristol Myers Squibb: Honoraria, Research Funding; Novartis: Honoraria, Research Funding.

INTRODUCTION: T-prolymphocytic leukemia (T-PLL) is a rare T cell malignancy with an aggressive clinical course. Preliminary clinical data suggest that allogeneic stem cell transplantation (alloSCT) may provide long-term disease control in a proportion of patients. However, direct evidence that graft-versus-leukemia (GVL) activity is indeed effective in T-PLL is lacking. We sought to investigate GVL in T-PLL by correlating minimal residual disease (MRD) kinetics with immune modulatory interventions (immunosuppression tapering, donor lymphocyte infusions (DLI), chronic graft-versus-host disease (cGvHD)), and T cell receptor (TCR) repertoire diversity alterations after alloSCT. METHODS: The study sample consisted of 10 consecutive patients who received alloSCT for T-PLL at the University of Heidelberg between 2007 and 2015. Quantitative MRD monitoring was performed using clone-specific real-time quantitative PCR (RQ-PCR) of clonal TCR beta (TRB) and/or gamma (TRG) gene rearrangements. Data interpretation followed EuroMRD guidelines. In selected patients, TCR repertoire diversity was analyzed longitudinally by next-generation sequencing (NGS). TRBV-TRBD-TRBJ gene rearrangements were amplified according to BIOMED2 protocol on genomic DNA. PCR products were sequenced on Illumina's MiSeq platform. NGS data were analyzed through a purpose-built bioinformatics immunoprofiler (EuroClonality-NGS consortium). Rearrangements with very similar junctional amino acid sequences and identical TRBV and TRBJ gene usage were defined as clonotypes. RESULTS: Patients underwent alloSCT in remission after first-line (8) or salvage (2) alemtuzumab-based therapy. 5 patients were allografted with an unrelated donor, 4 with a related donor, and one received haploidentical alloSCT. Conditioning was fludarabine with cyclophosphamide and/or total body irradiation-based. All patients had a cytological complete response (CR) after alloSCT. 2 patients died early because of acute GvHD, and one had no MRD marker, leaving 7 patients for MRD monitoring. Of these, 3 were MRD- at alloSCT, whereas 5 patients remained or became MRD+ early after alloSCT. In all of these 5 patients, immunosuppression tapering (3) or DLI (2) resulted in significant reduction of MRD levels (range 1-3 log) and was accompanied by cGvHD in 3 patients. However, durable MRD- was obtained in only 2 patients (alive 81+ and 12+ months post transplant), whilst MRD re-increased in 3 patients after 5-28 months despite ongoing cGvHD in one of them. NGS of the TCR repertoire was performed longitudinally in 3 patients with the longest follow-up. A total of 104 samples (blood (n=91) and BM (n=10) plus 3 donor blood samples) were sequenced. The sample at diagnosis showed one or two major clonotypes in all 3 sequenced T-PLLs, predominantly reflecting a mono- or biallelic leukemic TRB gene rearrangement. Kinetics of this leukemic clonotype followed kinetics of RQ-PCR MRD measurement, demonstrating that NGS can be used to quantify MRD in T cell lymphoma. Immediately after transplantation, the TRB repertoire was heavily skewed in all 3 patients, but recovered over time. Also in all 3 patients, MRD responses were reproducibly associated with a shift from a clonal, T-PLL-driven profile to a polyclonal signature. This corresponded to the donor clonotype repertoire and disappeared with increasing MRD levels. In each of the samples after alloSCT, several expanded non-leukemic clonotypes were observed at a maximum frequency from 3% to 54% of total TRB sequences. However, during MRD response, novel dominant clonotypes that could explain a clonal GVL effect did not emerge. Figure 1 shows an example of a patient repeatedly showing MRD response to immune interventions. Currently, 5 patients are alive and in cytological CR (4-81 months after alloSCT), translating into a median relapse-free survival of 40 months. CONCLUSIONS: The MRD responses to immune interventions observed here provide the first direct evidence for the efficacy of GVL in T-PLL. However, the GVL effect in T-PLL appears to be often only limited or transient. It also does not seem to be caused by the emergence of novel dominant T cell clones but is rather relying on a poly-/oligoclonal T cell response. Figure 1. Example for MRD kinetics in relation to immune interventions. Figure 1. Example for MRD kinetics in relation to immune interventions. Disclosures No relevant conflicts of interest to declare.

Objective Currently standard recombinant adeno-associated virus serotype 2(rAAV2)–based vectors lack the efficiency for gene transfer into primary human CD34 + peripheral blood progenitor cells (PBPC). Materials and Methods An advancement in vector development now allows the generation of rAAV capsid mutants that offer higher target cell efficiency and specificity. To increase the gene transfer into hematopoietic progenitor cells, we applied this method for the first time on primary human CD34 + PBPC cells. Results On a panel of leukemia cell lines (CML/AML), significantly higher gene transfer efficiency of the rAAV capsid mutants (up to 100% gene transfer) was observed compared to standard rAAV2 vectors. A higher transduction efficiency in the imatinib-resistant cell line LAMA84-R than in their sensitive counterpart LAMA84-S and a pronounced difference in susceptibility for the capsid mutants vs rAAV2 in LAMA84-S were particularly striking. On solid tumor cell lines, on the other hand, rAAV2 was more efficient than the capsid mutants, suggesting an increased specificity of our capsid mutants for hematopoietic progenitor cells. On primary human CD34 + PBPC significantly higher (up to eightfold; 16% green fluorescent protein–positive) gene transfer could be obtained with the newly generated vectors compared to standard rAAV2 vectors. Conclusion These novel vectors may enable efficient gene transfer using rAAV-based vectors into primary human blood progenitor cells for a future clinical application.

Introduction: The impact of mutations and pathway deregulation for drug sensitivity is only partly understood. Here we systematically investigated heterogeneity of drug response and genetic lesions in leukemia and lymphoma using primary tumor cells in order to identify pathway dependencies which can be exploited by rational treatment approaches. Methods: Primary leukemia / lymphoma cells obtained from peripheral blood were characterized with an ex vivo high-throughput drug screening platform in two steps: In an initial screen 2221 different compounds were tested for cytotoxicity on a limited cohort (n=20). 67 promising compounds targeting different pathways were selected for further validation on a larger cohort (n=111; 97 chronic lymphocytic leukemia (CLL), 5 T-prolymphocytic leukemia (T-PLL), 6 non-CLL B-Non-Hodgkin-Lymphoma (B-NHL), mononuclear cells of 3 healthy donors). Heat inactivated human serum was supplemented to mimic micro-environmental conditions. Cell viability was assessed by quantification of ATP (CellTiterGlo®) 48 and 72 hours after drug application. To understand heterogeneous pathway dependencies, drug sensitivity was regressed on genetic profiles. Genetic characterization was performed by FISH, targeted sequencing of recurrent aberrations (BRAF, MYD88, NOTCH1, SF3B1, TP53) as well as whole exome sequencing. Results: We initially focused on the impact of compounds interfering with or dependent on the p53 pathway. Nutlin-3 and fludarabine induced cell death more efficiently in cells with wild-type (n=80) than in CLL with mutated p53 (n=17; Nutlin-3 10µM: 51±18 vs 80±18; fludarabine 10µM: 40±36 vs 67±23 [% viability of untreated control], both p Citation Format: Leopold Sellner, Malgorzata Oles, Mikolaj Slabicki, Carolin Blume, Jennifer Hüllein, Tatjana Stolz, Marina Lukas, Martin Sill, Christopher C. Oakes, Sascha Dietrich, Olaf Merkel, Anna Jauch, Manfred Hensel, Davide Rossi, Katja Zirlik, Jan Dürig, Ingo Ringshausen, Marco Herling, Martina Seiffert, Peter Dreger, Christof von Kalle, Anthony D. Ho, Hanno Glimm, Wolfgang Huber, Thorsten Zenz. Systematic mapping of drug sensitivity in hematological malignancies identifies vulnerability of chronic lymphocytic leukemia with mutant p53. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 5557. doi:10.1158/1538-7445.AM2014-5557

Introduction Critical pathway alterations contribute to heterogeneous biology and clinical course in chronic lymphocytic leukemia (CLL). The role of particular mutations for drug sensitivity has been demonstrated for CLL and other entities. Here we systematically investigated heterogeneity of drug response and genetic lesions in primary CLL. The comprehensive identification of pathway dependencies will lead to a better understanding of hematologic malignancies and this may be exploited by development of rational-driven treatment approaches. Methods Primary CLL cells were characterized through an ex vivo high-throughput drug screening platform in 384-well format. The drug library consisted of 67 compounds (inhibitors targeting critical pathways in CLL and other cancers as well as compounds with known activity against CLL cells (including ABT-263, CAL-101, dasatinib, fludarabine, ibrutinib)) at different concentrations. Heat inactivated human serum was supplemented to mimic micro-environmental conditions. CLL samples represented untreated CLL, sensitive or refractory disease as well as major genetic subgroups. Cell viability was assessed by quantification of ATP (CellTiterGlo®) 48 and 72 hours after drug application. To understand heterogeneous pathway dependencies, drug sensitivity was regressed on genetic profiles. Genetic characterization was performed by FISH, targeted sequencing of recurrent aberrations (BRAF, MYD88, NOTCH1, SF3B1, TP53) as well as whole exome sequencing. Results We report results from 97 primary CLL samples screened ex vivo. Validity of the approach was demonstrated by the response profiles of p53-interfering compounds: Nutlin-3a and fludarabine induced cell death in cells with wild-type (n=80) but considerably lessin CLL with mutated p53(n=17; Nutlin-3a 10µM: 51±18vs80±18;fludarabine 10µM: 40±36vs 67±23 [% viability of untreated control], both p Conclusion Our data provide a first comprehensive map of genetically determined drug sensitivity in CLL. The work offers a novel functional classification of CLL based on drug sensitivity and identifies synthetic lethal interaction for TP53 mutant CLL. Disclosures: No relevant conflicts of interest to declare.

Introduction CD20 is a cell surface glycoprotein, encoded by the MS4A1 gene, which is highly expressed on most B-cells. Although CD20 is suggested to be involved in calcium signaling, its exact function is unknown. The monoclonal antibody rituximab, which targets CD20, has revolutionized the treatment of B-cell lymphoma. Despite the widespread use of CD20 antibodies across lymphoma subtypes, little is known about the endogenous regulation of CD20. The tight correlation between CD20 expression and responsiveness to CD20 antibody suggests that increasing CD20 expression levels could improve rituximab sensitivity. This could be particularly useful for B-cell lymphoma with low CD20 surface expression (e.g. chronic lymphocytic leukemia). The main goal of the project is to dissect and understand the CD20 regulation using an unbiased RNAi approach and to exploit this knowledge to improve the efficacy of anti-CD20 therapy. Methods In order to discover novel regulators of CD20 expression, we performed a genome-wide RNAi screen. The shRNA library targets over 5000 mRNAs, with 5 to 6 shRNAs per target. The CD20 regulatome was studied in a Burkitt lymphoma cell line (Raji) characterized by medium expression of CD20. Results MS4A1 was the most significant screen hit that decreased CD20 surface levels with three shRNAs and served as positive control. Based on stringent selection criteria, we identified hits involved in diverse molecular mechanisms such as signaling, trafficking, chromatin remodeling and cell membrane composition. In total, 21 genes targeted with two non-overlapping shRNAs were selected for validation: 12 decreasing and 9 increasing CD20 levels. For every gene, two shRNAs were synthesized and cloned into shRNA expression vectors. Every shRNA was investigated in separate experiments for its effect on CD20 expression levels on Raji cells. We were able to validate and confirm the influence of those genes on CD20 phenotype on Raji cells, as well as on other B-cell lymphoma cell lines (SuDHL5, BL2 and BL60). The group of genes decreasing CD20 expression can provide an insight into physiological regulation and the machinery involved in CD20 trafficking and processing. More relevant from a therapeutic viewpoint are genes where silencing increases the expression of CD20. In order to investigate a potential therapeutic synergy of our hits with Rituximab, we investigated complement-mediated lysis triggered by Rituximab in vitro after silencing of several hits. In this assay we observed direct correlation between CD20 surface levels with the sensitivity towards Rituximab for our hits: increased CD20 levels led to more Rituximab induced lysis (up to two fold increase), and vice versa. One of the main aims of the screen was the discovery of druggable targets that can be potentially used in combination therapy with Rituximab. Indeed, a chemical inhibitor targeting one of our screen hits was able to increase CD20 expression levels on Raji cells. Conclusion With an unbiased shRNA screen we identified potent regulators of CD20 expression levels. Our approach led to the discovery of novel druggable targets, which modulate CD20 levels. This may provide the rationale for novel combinatory therapies, which might improve the efficiency of anti-CD20 therapy in B-cell lymphoma. Disclosures: No relevant conflicts of interest to declare.

The surface glycoprotein CD20 is expressed on healthy and malignant B-cells and can be therapeutically exploited by monoclonal antibodies (e.g. Rituximab, Ofatumumab, GA-101). Stem cells and pro-B-cells are lacking CD20 surface protein, which allows the reestablishment of B-cell populations after treatment. Other cell lineages do not express CD20. Rituximab has revolutionized clinical practice in the treatment of malignant B-cell lymphoma. However, refractory lymphoma cells can lose the expression of CD20 after Rituximab exposure and B-cell lymphoma with low CD20 surface expression may be less responsive to treatment. Despite the widespread use of CD20 antibodies across lymphoma subtypes, little is known about the endogenous regulation of CD20. It was shown that several chemical compounds may modulate the surface density of CD20 (e.g. histone deacetylase inhibitors), but the regulatory network controlling CD20 levels is not understood. We aim to dissect the CD20 regulatome and to exploit the findings for novel therapeutic strategies. The CD20 regulatome is studied in a Burkitt lymphoma cell line (Raji) characterized by medium expression of CD20 in comparison to other lymphoma models. As a proof of principle, we were able to down-regulate the surface expression of CD20 assayed by FACS with four independent silencing triggers targeting CD20 mRNA. To discover novel regulators of CD20 expression, we performed an unbiased genome-wide RNAi screen. The screen hits are involved in diverse molecular mechanisms such as signaling, trafficking, chromatin remodeling and membrane composition. The group of genes decreasing the expression of CD20 expression can provide an insight into the physiological and pathological machinery involved in CD20 trafficking and processing. The strongest hit in this group, next to CD20 knockdown, is potentially involved in the direct transport of CD20 to the cell surface. Clinically most relevant is the group of genes where silencing increases the expression of CD20. The two strongest hits in this category are a transcription factor and a gene involved in signaling. Detailed molecular characterization of several chosen hits is currently ongoing. Their inhibition may lead to increased sensitivity to anti-CD20 antibodies and thus provide starting points for rational drug design. Citation Format: Mikolaj Slabicki, Leopold Sellner, Alexander Jethwa, Michael Boettcher, Thorsten Zenz. Modulating therapeutic targets in B-cell lymphoma: Loss-of-function RNAi screen for CD20 regulators. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 3159. doi:10.1158/1538-7445.AM2013-3159

Abstract 3086 Background: Therapeutic options for patients with relapsed multiple myeloma (MM) after autologous stem cell transplantation (ASCT) include novel agents, conventional chemotherapy or salvage ASCT, with no standard of care. Design and Methods: We retrospectively reviewed all myeloma patients who relapsed after upfront ASCT and received an autologous re-transplantation as salvage therapy at our center over a period of 15 years (n=200). The aim of our study was to evaluate the role of salvage ASCT in terms of efficacy and to define patients who may benefit most from this treatment approach. Results: The median follow up after salvage ASCT was 57.1 months. Median progression-free survival (PFS) and overall survival (OS) after salvage ASCT were 15.2 and 42.3 months, respectively. Overall response rate (≥partial response) was 80.4% at day 100, excluding 6 patients who died before assessment. Factors associated with improved PFS and OS after salvage ASCT included remission duration of >18 months after upfront ASCT, bortezomib- or lenalidomide containing therapies for reinduction, response to reinduction, ISS stage I prior to salvage ASCT and year of salvage ASCT 2005 or thereafter. The remission duration after upfront ASCT and the ISS prior to salvage ASCT were combined to create a risk-stratification model with three distinct prognostic groups. Conclusions: Salvage ASCT causes low morbidity and is capable of achieving sustained disease control in patients with myeloma. The use of lenalidomide and bortezomib for reinduction has improved the results after salvage ASCT, suggesting that novel agents and salvage ASCT are complementary rather than alternative treatment approaches. Stratification by duration of remission after upfront ASCT and ISS allows prediction of survival after salvage ASCT. To circumvent the problem of stem cell harvest in intensively pretreated patients, we recommend collecting and cryopreserving sufficient stem cells for both upfront and salvage ASCT early in the course of myeloma treatment. Disclosures: No relevant conflicts of interest to declare.

Abstract 3568 Introduction: High dose chemotherapy with melphalan followed by autologous peripheral blood stem cell transplantation (PBSCT) is a standard treatment regimen for young patients with multiple myeloma. However, there are few studies, mainly with low patient counts, testing the benefit of the reapplication of high dose chemo therapy in relapsed or refractory myeloma. Methods: Here we retrospectively analyzed 178 patients (56% male, 44% female, median age 60 years) with relapsed or refractory myeloma who were treated by reapplying high dose chemotherapy with melphalan followed by autologous PBSCT in our institution over the last 18 years. The median follow up of this study was 54 months. Result: Median progression free survival (PFS) and overall survival (OS) after relapse autologous transplantation were 16 and 35 months, respectively. 66% of the patients received newer antimyeloma agents for reinduction therapy (39% thalidomide, 6% bortezomib, 21% lenalidomide). In univariate analysis, time between first transplantation and progression of disease had a significant impact on PFS and OS (p=0.001 and p Conclusion: Reapplication of high dose chemotherapy can be an effective treatment option for relapsed or refractory myeloma, in particular in patients with a time to progression after first autologous transplantation of more than one year. Disclosures: No relevant conflicts of interest to declare.

Abstract 4506 Introduction Gene transfer into malignant leukemia cells can be of relevance to overcome conventional therapy-resistance. Either suicide or immune stimulating gene therapy vectors may be a tool for second line treatment of imatinib-resistant chronic myelogenous leukemia (CML). Unfortunately, for gene transfer into CML cells, most current vector systems either lack sufficient transduction efficiency or an acceptable safety profile. Conventional adeno-associated virus (AAV) based vectors have an advantageous safety profile, yet lack the required efficiency. Methods Pseudotyped recombinant adeno-associated viruses of the serotypes 2/1 to 2/6 (rAAV2/1 to rAAV2/6) were screened on a panel of human CML cell lines and primary CML cells to determine their gene transfer efficacy. Additionally, double-stranded self complementary rAAV (scAAV) were used to determine possible second strand synthesis limitations. Results On CML cell lines, generally rAAV2/2 and rAAV2/6 were most efficient (Fig. 1). For both, an interesting difference in transduction efficiency between the imatinib-resistant LAMA84-R and imatinib-sensitive LAMA84-S cells were observed. On primary human CML cells, rAAV2/6 proved to be significantly more efficient than the other tested vectors (4.6% ± 5.3% GFP-positive cells, p=0.011). Additionally, the transduction efficiency could be significantly increased by using scAAV vectors (scAAV2/6: 43.1% ± 25.9% GFP-positive cells, p Conclusions Overall, pseudotyped rAAV and scAAV vectors offer efficient clinically relevant gene transfer into human CML cells with a potential for future clinical application. Disclosures: No relevant conflicts of interest to declare.

Due to their pluripotency, human CD34+ hematopoietic stem cells are targets of interest for the treatment of many acquired and inherited disorders of the hematopoietic system using gene therapeutic approaches. Human CD34+ peripheral blood progenitor cells (PBPC) can be readily mobilized into the blood and harvested by leukapheresis, providing an easily accessible source of hematopoietic progenitor cells. Unfortunately, for gene transfer into CD34+ PBPC, most current vector systems either lack sufficient transduction efficiency or an acceptable safety profile. Standard adeno-associated virus-based vectors have an advantageous safety profile, yet lack the required efficiency. Therefore a panel of pseudotyped recombinant adeno-associated viral (rAAV2/1 - rAAV2/6) vectors expressing the eGFP gene was screened on human G-CSF-mobilized CD34+ PBPC to determine their efficacy. In addition, the difference in transgene expression between conventional single-stranded rAAV and self complementary rAAV (scAAV) vectors was determined. For each vector n≥6 was performed and data are shown as mean ± SD. Of all screened conventional rAAV vectors, rAAV2/6 proved to be the most efficient (13.5% ± 9.8% GFP+ and CD34+ PBPC; p10%) with expression rates up to 60% in human CD34+ PBPC using an AAV-based vector system, thereby providing an efficient alternative vector system for gene transfer into this clinically important target cell.

Introduction: G-CSF mobilized peripheral blood progenitor cells (PBPC) are an excellent target for the cure of acquired and inherited disorders of the hematopoietic system. AAV-based vectors have an advantageous safety profile, due to their low immunogenicity, the lack of any associated human disease and its episomal localization. Unfortunately, currently standard rAAV2-based vectors lack the required efficiency for gene transfer into human blood progenitor cells. To increase the gene transfer efficiency on primary hematopoietic cells we applied an AAV random peptide library on primary human CD34+ PBPC. Materials and Methods: The AAV random peptide library was applied on primary human CD34+ PBPC samples (3 selection rounds). Several mutant blood progenitor-targeted rAAV2 capsid clones were obtained and sequenced. Five new rAAV vectors with the most efficient clones were produced. Titers were verified using real time PCR-based titration. To determine efficiency and specificity, a panel of leukemia (CML and AML) and non-leukemia cell lines, as well as primary human CD34+ PBPC were transduced. GFP expression was determined against uninfected control cells. Results: On a panel of leukemia cell lines at least one capsid mutant was significantly more efficient (%GFP+; p Conclusion: These novel vectors may finally enable efficient (>20%) gene transfer using rAAV-based vectors into primary human PBPC and may offer a safe alternative to other vector systems which is applicable in vivo.

Resistance of chronic myelogenous leukemia (CML) to imatinib is a serious clinical challenge. In vivo targeted gene therapy may be a strategy to overcome resistance. For gene therapy of CML standard rAAV2-based vectors lack the required gene transfer efficiency as well as the in vivo selectivity. An advancement in vector development (Muller et al., Nat. Biotech., 2003) now allows the generation of rAAV capsid mutants that offer higher target cell specificity and efficiency. We used the “AAV random peptide library” method to generate capsid mutants efficient and specific for CML cell lines, primary CML cells and primary human peripheral blood progenitor cells PBPC. After previously applying this method successfully on CML cell lines, an effective vector for CML cell lines could be obtained. On primary human CML and PBPC the new vector was more efficient than rAAV2 but still at an unsatisfying level. To increase the gene transfer on primary cells we applied the “AAV library” on primary human CML and PBPC. A total of 44 different mutant capsid clones were obtained, most of them at three or less copies. Four clones were observed >10x to 39x and were selected for further testing. To determine efficiency and specificity, a panel of leukemic (CML, imatinib-sensitive and -resistant, and AML) and other solid tumor cell lines were transduced (moi 100). On a panel of leukemic cell lines similar or higher gene transfer efficiency of the rAAV capsid mutants was observed compared to standard rAAV2 vectors. The higher transduction efficiency in imatinib-resistant cell line LAMA84-R than in their sensitive counterpart LAMA84-S was particularly striking, as the difference in susceptibility for the capsid mutants vs rAAV2 in LAMA84-S (table 1). On solid tumor cell lines rAAV2 was more efficient, confirming the higher specificity of our capsid mutants. Compared to our previous work ( Gene transfer efficiency (%GFP+ cells) Cell line Capsid mutant Standard rAAV2 Data are given as mean ± SD LAMA84-S 78.2 ± 7.9 9.2 ± 2.5 LAMA84-R 97.8 ± 0.3 93.0 ± 0.2 HL-60 54.0 ± 5.4 57.6 ± 8.2 BV173 48.4 ± 7.4 30.6 ± 1.8 EM3 3.9 ± 0.7 0.2 ± 0.2 KG1a 24.7 ± 0.7 12.8 ± 5.0

Motivation The study of immunoglobulins and T cell receptors using next-generation sequencing has finally allowed exploring immune repertoires and responses in their immense variability and complexity. Unsurprisingly, their analysis and interpretation is a highly convoluted task. Results We thus implemented ARResT/Interrogate, a web-based, interactive application. It can organize and filter large amounts of immunogenetic data by numerous criteria, calculate several relevant statistics, and present results in the form of multiple interconnected visualizations. Availability and Implementation ARResT/Interrogate is implemented primarily in R, and is freely available at http://bat.infspire.org/arrest/interrogate/ Supplementary information Supplementary data are available at Bioinformatics online.

Background For many promising target cells (e.g.: haematopoeitic progenitors), the susceptibility to standard adeno-associated viral (AAV) vectors is low. Advancements in vector development now allows the generation of target cell-selected AAV capsid mutants. Methods To determine its suitability, the method was applied on a chronic myelogenous leukaemia (CML) cell line (K562) to obtain a CML-targeted vector and the resulting vectors tested on leukaemia, non-leukaemia, primary human CML and CD34+ peripheral blood progenitor cells (PBPC); standard AAV2 and a random capsid mutant vector served as controls. Results Transduction of CML (BV173, EM3, K562 and Lama84) and AML (HL60 and KG1a) cell lines with the capsid mutants resulted in an up to 36-fold increase in CML transduction efficiency (K562: 2-fold, 60% ± 2% green fluorescent protein (GFP)+ cells; BV173: 9-fold, 37% ± 2% GFP+ cells; Lama84: 36-fold, 29% ± 2% GFP+ cells) compared to controls. For AML (KG1a, HL60) and one CML cell line (EM3), no significant transduction (+ cells) was observed for any vector. Although the capsid mutant clone was established on a cell line, proof-of-principle experiments using primary human cells were performed. For CML (3.2-fold, mutant: 1.75% ± 0.45% GFP+ cells, p = 0.03) and PBPC (3.5-fold, mutant: 4.21% ± 3.40% GFP+ cells) a moderate increase in gene transfer of the capsid mutant compared to control vectors was observed. Conclusion Using an AAV random peptide library on a CML cell line, we were able to generate a capsid mutant, which transduced CML cell lines and primary human haematopoietic progenitor cells with higher efficiency than standard recombinant AAV vectors.