Peptide-drug-conjugates for Personalized, Targeted Therapy of Chronic Lymphocytic Leukemia

Overview

Using phage libraries extensively pre-absorbed on a series of normal cell types, we will isolate phage specifically internalized by B-CLL cells from newly diagnosed and untreated CLL patients. Peptide sequences are then derived by Next Generation Sequencing (NGS). NGS-based studies are contributing to an improved understanding of cancer heterogeneity in order to tailor treatment to patients based on the individual makeup of their tumor. However the use of NGS to derive phage displayed peptide sequences is so far rare (22). Traditionally, after exposure to a target and recovery by elution, the phage clones are isolated by titration on bacterial lawns. It is technically demanding and labour intensive to select and analyze more than about 15 of the sometimes thousands clones recovered. Therefore information on other potentially important sequences is missed. NGS allows sequencing of the entire recovered phage pool and provides far more detailed bioinformatic analyses of peptide sequences or motifs. RNA from the CLL cells is used for RNA-seq expression sequencing. The wide application of NGS in combination with bioinformatics tools has begun to revolutionize cancer research, diagnosis and therapy. The peptide and RNA sequencing data will afford bioinformatic testing of correlations of exome expression and clinical parameters with the pattern of peptide sequences internalized by CLL cells of different patients. This information is crucial to answering questions 1, 2 and 3 discussed on page 1 above. The results of this analysis will probably not allow identification of specific receptors targeted by the peptides. The aim at this stage of the research is to identify candidate targeting peptides. Once identified, further research will be needed to identify the receptors to which they bind. Regarding question 4, there is currently very little published information on the therapeutic potential of PDCs in leukemia. Using two peptides we have isolated that target murine A20 leukemic cells, we will prepare multi-drug PDCs (using technology we have developed) and in an animal model, test their ability to enhance the survival and quality of life of CLL bearing animals.

Methodology A: CLL specific peptides: (Specific aim 1) Whole blood samples are received from healthy volunteers and newly diagnosed and untreated CLL patients according to our Helsinki permit no. 0432-13-RMC. (Rabin Medical Center) Acquisition of samples at follow up and from an additional 80 patients will be covered by additional Helsinki submissions now underway. Lymphocytes are separated using Ficoll and the B cells isolated using a B-cell isolation kit. The B-cells from the healthy volunteers and an aliquot of those isolated cells from each patient are frozen for later use (see D: below). The remaining patient cells are divided into two fractions. One is used for RNA extraction. The other is exposed to an "Absorbed" phage library pre-absorbed on normal cells. After incubation, cells are washed and phage bound to the cell membrane are eluted and recovered. Cells are then lysed and the lysate retained. Phage in the eluate and lysate (membrane-bound and internalized respectively) are amplified separately in E.coli. The amplified phage are isolated and DNA prepared for NGS. Sequencing is carried out in an external core facility (The Technion Genome Center). . B: Bioinformatic analyses of peptide sequences (Specific aim 1) I) Sequencing of peptides: Sequence reads are obtained from phage DNA libraries generated and analyzed with high-throughput sequencing (Illumina HiSeq system). Quality control checks are performed using FastQC tool. The reads are processed to trim adaptors and remove the consensus sequences from either side using the Cutadapt software (23). The remaining fragment tags contain the peptides' DNA sequence and are expected to be 21 nucleotides in length. Next, the BioString package from the Bioconductor framework are used to translate these nucleotide sequences, and the number of occurrences of each peptide sequence is calculated to find the most frequent peptides. Following normalization to library size, peptide sequence counts from different patient libraries are compared to find common sequences. Sequences from the membrane-bound pool fraction are subtracted from the internalized pool fraction so as to remove membrane-bound clones that may not have been successfully eluted off the CLL cell. The sequences remaining in the internalized pool serve as candidates for peptide-conjugate therapy. A univariate logistic regression approach will then be used to test possible associations of peptide counts with clinical and genetic findings. When clinical parameters are treated as continuous variables, linear regression will be implemented. Candidate peptides will be queried against the PepBank database (24) in order to check if they are known to be related to cancer or other diseases and conditions such as apoptosis or angiogenesis. Finally, a bioinformatic pipeline script will be developed automating the above bioinformatics analysis steps. II) RNA-seq to detect gene expression patterns and fusion genes: The main aim of the RNA-seq analysis is to detect specific gene expression patterns which can be correlated with peptide sequences and clinical parameters. These expression patterns will give clues to cell surface receptor expression and intracellular pathway activity. Patients will be divided based on the peptide sequences internalized by B-CLL cells, and a statistical test will be performed in order to detect differentially expressed genes. Pathway analysis will be performed in order to infer functional changes which may shade a light on the possible role of the peptide receptors in the cell signaling cascade. Trim galore software will be used for adapter trimming, and for removing low quality bases from the ends of reads. Trimmed reads will be mapped to the human genome (hg38) using TopHat2 software. The number of reads overlaping each of the annotated genes will be counted using the HT-seq python package. DESeq within the Bioconductor framework will be used for normalization and differential expression analysis using variants of Fisher's exact test. The Generally Applicable Gene-set Enrichment (GAGE) method will be applied to detect up and down-regulated pathways. RNA-seq is also used to detect fusion gene events in the patient's cancer cells and can be used as an independent validation method to test the sensitivity of the peptide-dye conjugate used as for disease monitoring. For this aim, three different tools TopHat-fusion, defuse and ChimeraScan will be used to identify fusion transcripts. Several filtration steps will be applied in order to remove low-quality candidates as described before, and only those candidates detected by all the three tools will be chosen for RT-PCR validation. C: Syntheses of Peptides and conjugates (Specific aims 1,2,3) Peptides will be synthesized using solid-phase chemistry. In the first stage of the project, the two peptides specific for A20 mouse leukemic cells already identified will be synthesized, namely HIS SER THR PRO SER SER PRO (Peptide 1) and ASP SER SER LEU PHE ALA LEU (Peptide 2). In later stages (see below), selected human CLL specific peptides will be synthesized as their sequences become available. Only peptides will multiple repeat reads (\>5) (see B above) will be selected (maximum of 3 per patient) as these represent phage clones with greater propensity to induce uptake into cells. Peptides will be synthesized both as fluorescent dye- and drug-conjugates based on our previous work with targeted multi-drug conjugates (21). We will employ drugs with differing mechanisms of action, such as the nitrogen mustard Chlorambucil used to treat elderly CLL patients, and the microtubule inhibitor Combretastatin 4A, known to induce apoptosis in CLL cells (25). Purity and composition of the products will be verified by HPLC and LC/MS. Syntheses will be carried out by an external supplier. D: In vitro validation of peptide specificity and PDC activity: (Specific aims 1, 2, 3) Specificity of peptides: The specificity of targeting peptides will be demonstrated by incubating peptide-dye conjugates with target and controls cells. For the mouse CLL system, A20 cells will be used as target cells and normal mouse splenic lymphocytes and primary cultures of mouse skin and epithelial cells will be used as controls cells. For the human CLL system, patient CLL cells and normal human lymphocytes and HUVEC cells will be used as target and controls cells respectively. Cells will be incubated with FITC-labeled peptides at both 4oC and 37oC, washed and analyzed for binding and internalization respectively. Binding will be evaluated by flow cytometry and internalization by confocal microscopy. Cytotoxicity of PDCs: Target and control cells will be incubated for 24 and 48 hrs with increasing concentrations (0-50uM drug equivalents) of free drug or PDCs. Percent Growth Inhibition will be assessed by the XTT assay. Cytotoxic A20 specific PDCs will then be evaluated in vivo (see E: below). Cell culture systems: The A20 cell line is grown and maintained in FBS supplemented RPMI culture medium. Normal primary cultures are maintained with selective media. On the other hand, the survival of human CLL cells in culture is dependent on several unique microenvironmental factors, including antigen stimulation of the B-cell BCR, T cell help through the CD40-CD40L interaction, stimulation from the bone marrow stromal cell-derived CXCL12 chemokine and stimulation of Toll-Like Receptor 9. These conditions will be mimicked in vitro (26) by addition to the culture medium of anti-IgM (to crosslink the BCRs), co-culture with fibroblasts expressing CD40L, addition of recombinant CXCL12 and CpG dinucleotides respectively. These conditions will be used to test the specificity and efficacy of human PDCs against cryopreserved samples of patient CLL cells and normal human lymphocytes (controls). E: Animal studies using A20 mouse model of CLL (Specific aim 3) While several antibody-drug conjugates for the treatment of CLL have been used or are being tested (16), there is little published information as to the efficacy of PDCs for this disease. Therefore we aim to study the behavior of our A20 specific PDCs in an animal model of CLL, according to protocols approved by the Ariel University Institutional Animal Ethics Committee, permit no. IL-47-08-13. To validate the calibration of CLL development, A20 cells will be injected intravenously into Balb/c mice and the development of CLL symptoms monitored. Usually this takes between 40-50 days. Study A): At this point, we know little about the most effective treatment schedule with PDCs. For this reason, and based on published literature, we will first test the systemic tolerance of the animals to the drugs in the PDCs by injecting groups of non-tumor bearing mice intraperitoneally with 1, 7.5 or 15 mg/Kg of one Chlorambucil and Combretastatin 4A containing PDC every third days for 3 weeks. Animals will be monitored for systemic morbidity and survival. On the basis of these results, an initial treatment schedule will be selected to compare two concentrations for each of the two PDCs. Animals showing initial signs of CLL will be treated and followed for survival and development of CLL symptoms for at least and additional 70 days. Study B: Using the Study A results, the most effective PDC will be studied further. The PK of the drug conjugate will be analyzed by administering it in 5 drug concentrations, including that used for Study A. Blood samples will be drawn from the tail vein at 0.25, 0.5, 1, 2, 5, 10 and 24 hrs. PDC will be extracted and analyzed by LC/MS for content and integrity. Drug concentrations will be analyzed by HPLC and used to calculate standard PK parameters \[Cmax; TCmax; T1/2\]. Based on these experiments, an additional treatment protocol will be devised aimed at maintaining a maximum blood PDC concentration over the treatment period. The protocol will be tested on an additional group of animals. After the treatment period, animals will be followed for survival and development of CLL symptoms for at least and additional 70 days. F: Detection of CLL cells in animal and patient whole blood. (Specific aims 1 and 2) The aim of these experiments is to assess the potential of CLL-specific peptide-dye conjugates to detect CLL cells in blood samples. Calibration experiments will be carried out on three levels. 1. Cell doping: Cryopreserved samples of patient CLL cells will be added to either RPMI culture medium or normal human blood (in three concentrations). B-cells isolated as described in A above will be incubated with the appropriate Peptide-dye conjugate for 30mins at 37OC. Cells will be washed and analyzed for peptide binding by flow cytometry. Similarly, A20 cells will be added to normal mouse blood and tested with A20 specific peptide-dye conjugates. 2. Animal model: Using the animal model described above, blood samples will be taken once/week during the tumor induction phase. After RBC lysis, the remaining cells will be incubated with peptide-dye conjugates, washed and analysed by flow cytometry. Furthermore, samples will be taken once/fortnight from animals surviving Study B and analysed for Minimal residual Disease and relapse. 3. Patient follow up: Aliquots of blood samples taken during routine follow up of the CLL patients will be used. Cells will be incubated with the appropriate Peptide-dye conjugate for 30mins at 37OC. Cells will be washed and analyzed for peptide binding by flow cytometry. G: Statistics Cytotoxicity of PDCs and free drug will be assessed by calculating % Growth Inhibition of treated versus untreated cells. Experiments will be performed in triplicate and repeated thrice. In animal studies, survival will be assessed using Kaplan-Meier curves. Additional statistics related to the bioinformatic analyses are described in detail above in section 1.7-B.