Background
Mantle cell lymphoma (MCL) is an aggressive B-cell lymphoma, molecularly defined by the translocation of
CCND1 [
1]. The malignant cells harbor a number of molecular abbreviations such as overexpression of SOX11 [
2] and constitutive activation of the nuclear factor-κB (NF-ĸB) pathway [
3]. The NF-ĸB pathway regulates a number of genes involved in apoptosis, cell adhesion, proliferation and tissue remodeling. Especially, relapsed MCL has increased activity of the pathway which most likely has a key role in maintaining tumour cell viability and drug resistance, through overexpression of several anti-apoptotic proteins [
4,
5].
Traditionally, MCL was characterized by initial sensitivity to standard chemotherapy followed by relapse, and unfavorable outcome [
6,
7]. However, addition of high-dose cytarabine treatment as part of the induction therapy has resulted in great improvement in survival in subgroups of MCL patients [
8]. Cytarabine (ara-C, cytosine arabinoside) is a deoxycytidine nucleoside analogue, an S-phase specific anti-metabolite, which is used in modern MCL combinatorial treatment protocols [
9]. High-dose cytarabine is effective due to the improved retention of ara- CTP by target cells [
10], but likewise toxic, causing mainly hematological side effects. Thus, understanding the molecular mechanism(s) responsible for resistance, identifying predictive markers for resistance and/or sensitizing agents, would be of great clinical value.
Cytarabine is a prodrug, which first needs to be transported across the plasma membrane, and secondly become activated through phosphorylation. Transportation of nucleosides and nucleoside analogues across the plasma membrane is mediated by transporter proteins belonging to the solute carrier families 28 and 29 (
SLC28 and
SLC29).
SLC28 genes encode the three members of the concentrative nucleoside transporter (CNT) family, while the four members of equilibrative nucleoside transporter (ENT) proteins are encoded by
SLC29 genes [
11]. Both ENT and CNT recognise most of the nucleoside analogues used for cancer therapy and as such they are interesting targets for further studies. For most of the nucleoside analogues commonly used for anti-cancer therapy, the first phosphorylation step is catalysed by deoxycytidine kinase (dCK). Both de novo resistance and acquired resistance to cytarabine, including cross-resistance to other nucleoside analogues, have been linked to down-regulation of dCK on gene and protein level [
12‐
14].
Today, there are many treatment alternatives available for relapsed or recurrent MCL patients but only little information available on which patients that would benefit from each alternative. Thus, the aims of the present study were to (i) characterize the mechanisms of cytarabine resistance in MCL, (ii) identify drugs suitable for treatment of relapsed/recurrent MCL patients treated with Ara-C and (iii) to suggest preventive measurements based on in vitro-model data.
To do so we have established a unique MCL resistant model in which cytarabine resistance repetitively and molecularly reproducibly can be induced in a highly controlled manner. Using molecular profiling, we show that down-regulation of the dCK protein is key to development of resistance. The cellular model, representing three stages of resistance development (naïve sensitive, exposed sensitive and resistant) was further characterized using gene expression analysis and functional analysis. Key gene changes, including upregulation of the transcription factor SPIB was identified. We further show that similar to relapsed/recurrent MCLs, the resistant cells are not only defined by the lack of dCK and increased SPIB, but also high levels of NF-κB.
Functional screens using (i) chemotherapeutics or (ii) epigenetic regulators were used to identify drugs with potential cross-resistance and/or sensitivity, and to select individual epigenetic candidate drugs for sensitisation of cytarabine-resistant cells. Co-treatment with bortezomib and cytarabine prevented resistant development, but could not overcome resistance once dCK was abolished.
The importance of dCK for response to therapy was confirmed by the analysis of primary MCLs where 97% of the patients have high or intermediate dCK at diagnosis.
Methods
Cell culture
The MCL cell line Z138 was purchased from ATCC (Manassas, VA, USA). Both untreated and resistant cell lines were cultured in R10 medium (RPMI 1640 (HyClone Laboratories, Logan, UT, USA) supplemented with 10% fetal bovine serum (Invitrogen, Carlsbad, CA, USA) and 1% 2 mM L-glutamine (Invitrogen)).
Reagents
Cytarabine (147–94-4, Pfizer, New York, NY, USA) was aliquoted and stored at 4 °C with bulk concentrations of 411 mM. SCREEN-WELL® Epigenetics library (BML-2836, Enzo LifeSciences Inc., Farmingdale, NY, USA) and Chemotherapeutic Agent Library (L1500, Selleck, Munich, Germany) were stored at − 80 °C until use. Substances used for proliferation studies including bortezomib (2204S, Cell Signaling Technologies), lenalidomide (PCID-216326 Santa Cruz Biotechnology, USA), apicidin (A8851, Sigma Aldrich), belinostat (PXD101), M-344 (M5820, Sigma Aldrich), oxamflatin (O3139, Sigma Aldrich, St. Louis, MO, USA), scriptaid (S7817, Sigma Aldrich), trichostatin A (T1952, Sigma Aldrich) and vorinostat (SAHA MK0683, Selleck) were dissolved in DMSO (Sigma Aldrich), aliquoted and stored at − 80 °C until use.
Establishment of resistant sub-clones
The first resistant sub-clone defined as Z138 Cytarabine Resistant (Z138-CytR) was established by continuous exposure of wild type Z138 Cytarabine Naïve Sensitive cells (Z138-CytNS) to increasing concentrations (0.005–0.3 μM) of cytarabine. Using this model, we could identify the approximate time to resistance development, and utilize this information for developing a novel time-controlled cytarabine resistant model, described below.
Z138-CytNS, with viability above 85%, were exposed to 0.005 μM cytarabine and kept at log phase (1–2 × 106 cells/ml). Concentrations were increased two- or ten-fold, and samples for immunoblotting were taken when viability reached above 85%. When reaching a concentration of 0.2 μM cytarabine, cells were expanded and frozen as a cell biobank, a sub-clone called Z138 Cytarabine Exposed Sensitive (Z138-CytES). This cell biobank could then be used for further cytarabine exposure experiments and the establishment of a Cytarabine Resistant 21 days (Z138-CytR21) sub-clone.
Effect of cytarabine on sensitive and resistant cell lines
Cells were seeded in a 48 well plate and incubated with 0, 0.5, 5 and 50 μM of cytarabine at 37 °C (5% CO2) for 24–96 h. Duplicates from each concentration were counted in an automatic cell counter (Countess™, Invitrogen) at each time point, and trypan blue exclusion method was used to monitor viability.
Assessment of cell proliferation by [methyl-14C]-thymidine incorporation
Cells were seeded in a Cytostar-T 96 well plate (Perkin Elmer, Waltham, MA, USA) and cultured for up to 72 h in presence of 0.5 μCi/ml [methyl-14C]-thymidine (PerkinElmer). Cell proliferation was measured at indicated time-points using a Wallac 1450 MicroBeta liquid scintillation counter (Perkin Elmer). Prior to all measurements, cells were centrifuged to allow contact to the scintillation liquid.
Re-introduction of dCK into dCK negative resistant cells
To assess the importance of dCK in relation to resistance, the protein was transiently re-introduced into resistant cells. The Amaxa protocol (Amaxa Biosystems Cologne, Germany) for nucleofection of suspension cell lines was followed, using program CM-138 and Cell Line Nucleofector Solution SF (Amaxa Biosystems). For the re-introduction experiments, 2.5 × 10
6 cells were mixed with 2 μg of OmicsLink™Expression Clone for dCK (EX-C0081-M46, vector information can be found in Additional file
1) in each reaction and a GFP vector was used as a positive control (both from GeneCopoeia, Germantown, MD, USA).
Gene expression analysis
Triplicate cultures of cells were harvested at different time-points and lysed in TRIzol (Invitrogen). Preparation of tRNA was performed as previously described [
15]. Gene expression was assessed using Affymetrix Human Transcriptome Array 2.0 (HTA 2.0; Affymetrix Inc., Santa Clara, CA, USA), and acquired data was pre- processed at the SCIBLU Genomics Centre (Lund University, Sweden) involving quality control and normalization, using the Expression Console software (Affymetrix Inc). Normalized and log
2 transformed data was imported into Qlucore Omics Explorer 3.0 (Qlucore AB, Lund, Sweden) for statistical analysis. For confirmation of mRNA expression in different samples, TaqMan probe-based RT-PCR was performed, using TaqMan® Fast Universal PCR Master Mix (Applied Biosystem, Waltham, MA, USA) and the TaqMan assay Hs01040726_m1 (dCK, Applied Biosystem) and Hs00162150_m1 (SPIB, Applied Biosystem). 18S (Hs99999901_s1, Applied Biosystem) was used as reference gene. All data were analyzed using the 7500 software v2.0.5 (Applied Biosystem). Functional annotation of individual genes was obtained from NCBI/Gene (
http://www.ncbi.nlm.nih.gov/gene), GeneCards (
http://www.genecards.org/) or UniProt (
http://www.uniprot.org/).
Library preparation, hybridization capture and MPS sequencing
DNA from Z138-CytNS, Z138-CytES and Z138-CytR cells were purified using RNeasy Plus Mini Kit (Qiagen, Hilden, Germany) and thereafter quantified using the Qubit system (Life Technologies, Carlsbad, CA, USA). Two μg of DNA were fragmented using the Covaris S2 Ultrasonicator (Covaris, Woburn, MA, USA) and DNA fragments from 64 target genes, including TP53 were captured using SureselectXT Custom 3–5.9 Mb library kit (Agilent Technology, Santa Clara, CA, USA). Before capture, eight samples were pooled, and the molarity of the pooled library was determined based on and DNA fragment size distribution measured on a Bioanalyzer (Agilent) and concentration measured by Qubit. Sequencing was performed on the Illumina HiSeq 2500 (Illumina, San Diego, CA, USA) with 2 × 101 bp paired end reads.
Analysis of sequencing data
Picard Extract IlluminaBarcodes and IlluminaBasecallsToSam
(https://broadinstitute.github.io/picard/) was used for format conversion and demultiplexing of raw Illumina sequencing data and sequence reads were aligned to the human reference genome hs37d5ss (1000 genome with decoy sequences) using Novoalign (
http://www.novocraft.com). Picard MarkDuplicates were used to identify and exclude PCR duplicates in subsequent analyses and quality scores were recalibrated and indels realigned using the Genome Analysis Tool Kit (GATK) [
16]. GATK UnifiedGenotyper with a call confidence cutoff of 10 were used to identify genetic variants and genotypes. Variants were annotated for their effect on protein coding transcripts using snpEff and Annovar using RefSeq reference transcripts. Bases of coding exons and 20 bp of adjacent introns were covered by at least 30 reads and variants affecting coding exons and 20 bp of adjacent introns were evaluated for pathogenicity.
Silencing of SPIB in the resistant cell line
Cell Line Nucleofector Solution SF was used with program CM-138 following the Amaxa protocol for suspension cell lines (Amaxa Biosystems). Cells were mixed with 1000 nM of siRNA (Ambion, Austin, TX, USA) or a scrambled sequence. GFP-producing plasmid was used as control for the transfections (Amaxa Biosystems).
Effect of bortezomib on resistance development
Based on the set up presented above, Z138-CytES cells were co-treated with 0.3 μM of cytarabine and 0.001 or 0.01 μM of bortezomib during the 21 days expected for resistance to develop. A positive control with only 0.3 μM of cytarabine was grown in parallel (as visualized in Fig.
10c). Z138-CytNS cells and Z138-CytR cells were subjected to the same treatment during the same period. Lysates for immunoblotting were sampled continuously during the treatment period and after completed treatment. After 21 days of treatment, proliferation in cytarabine containing medium was assessed by [methyl-
14C]-thymidine incorporation as previously described.
Immunoblotting
Cells were harvested for western blot and lysed on ice with lysis buffer (1% NP40 (Sigma Aldrich) in PBS supplemented with 1× complete protease inhibitor (Roche Applied Sciences, Indianapolis, IN, USA)) for 30 min followed by centrifugation at 4 °C, 1300 rpm for additional 30 min. Supernatants consisting of protein lysates were collected and protein concentrations were measured using a bicinchoninic acid kit (Sigma Aldrich). For western blot, 25 μg protein was loaded on a Bis-Tris gel (Life Technologies, Carlsbad, USA). Following electrophoresis using an XCell Surelock Mini- Cell system (Life Technologies), the proteins were immediately blotted onto a PVDF membrane using program P3 on the iBlot Dry Blotting System (Life Technologies). The membranes were then blocked for 60 min with 5% milk in PBS, before incubation with primary antibodies targeting dCK (TA502698, OriGene Technologies, Rockville, MD, USA), ENT1 (ab 11,337–1-AP, Proteintech, Chicago, IL, USA), SPIB (Cell Signaling), NF- κB (D14E12, Cell Signaling), IκBα (44D4, Cell Signaling) and/or GAPDH (G8795, Sigma Aldrich). Horseradish peroxidase-conjugated rabbit anti-mouse immunoglobulin (P0260, Dako, Glostrup, Denmark) and swine anti-rabbit immunoglobulin (P0217, Dako) were used as secondary antibodies. Protein levels were visualized in a ChemiDoc™ MP Imaging System (Bio-Rad Laboratories, Hercules, CA, USA) with SuperSignal West Femto Maximum Sensitivity Substrate (ThermoFisher Scientific). Quantification of the results was performed using the Image Lab software (Version 5.2.1, Bio-Rad Laboratories).
Screening of compound libraries and validation of selected drugs
Z138-CytNS or Z138-CytR cells were seeded in a Cytostar-T 96 well plate as previously described, and treated with different concentrations of chemotherapeutic and epigenetic drugs. Non-treated cells re-suspended in R10 medium were considered as R10 controls, and DMSO (0.01%) treated cells as vehicle controls. Proliferation was measured 0, 24 and 48 h after addition of chemotherapeutic and/or epigenetic drugs. To evaluate the additive effect of epigenetic drugs to cytarabine, 50 μM cytarabine was added 6 h after pre-incubation of cells with epigenetic compounds.
Patients, cohorts and treatment protocols
Materials from patients included in the Nordic Lymphoma Group MCL2 and MCL3 trials at hospitals in Sweden, Finland, Norway and Denmark, were selected for TMA construction as previously described [
17]. The treatment protocols for MCL2 and MCL3 both include high-dose cytarabine, rituximab and ASCT as previously described [
9].
Immunohistochemistry staining and digital scoring
Immunohistochemistry was performed as previously described [
2]. The sections were stained for dCK (TA502698; OriGene Technologies) and visually analysed using a Nikon ECLIPSE 80i microscope (Nikon Instruments Inc., Melville, NY, USA) at a magnification of 20× (Plan Flour 20× DIC M/N2, Nikon) with a numeric aperture of 0.5. Images were captured using a Nikon DS-U2/L2 USB (Nikon) camera, and NIS Elements BR 3.10 (Nikon) as acquisition software. For digital scoring, dCK stained slides were scanned at an absolute magnification of 20× (resolution of 0.493 μm per pixel) and digitally scored using HALO™ (Indica Labs, Corrales, NM, USA). Positive areas (tumour) and negative areas (stroma) were separated and quantified based on a pattern recognition algorithm in the HALO platform. Image analysis based on RGB (red, green, blue) spectra was used to detect all cells by counterstaining with hematoxylin (blue). All analysis settings including thresholds set for weak, intermediate and strong nuclei staining were maintained throughout the whole study (Additional file
2: Table S1).
Discussion
Resistance to cytarabine is a significant clinical problem, as this agent is part of the backbone of treatment in a wide range of malignancies, including MCL. Thus, ways of predicting and possibly prevent cytarabine resistance are needed. The aim of the present study was to further explore the mechanisms of cytarabine resistance in MCL, to identify drugs suitable for treatment of patients with relapsed/recurrent disease and to identify preventive measures. To address these important clinical questions, we have established a unique MCL resistant model, based on the cell line Z138, in which cytarabine resistance repetitively in a timely and molecularly reproducibly manner can be induced. This cellular model allows for the first time resistance to be studied over time, and thus constitute an important research tool, despite the obvious drawback of being based on a single cell line It is known that dCK which is a key enzyme in converting cytarabine to active drug is dependent on functional p53 [
24]. To ensure that the developed resistance was not caused by introduction of TP53 mutations due to cellular stress and selection pressure, the resistant cells were sequenced and found to contain only TP53 WT alleles.
Transportation and activation of cytarabine is mediated by several classes of transporters and activating enzymes, such as ENT [
11] and dCK [
12], as described in more detail elsewhere [
25]. Using molecular profiling, we show that in our model, down-regulation of ENT1 is associated with adaptation to cytarabine exposure, but no further decrease is seen upon final development of resistance. Thus, in our model, ENT1 is not the limiting factor for resistance development, but rather a stand-by effect related to exposure to cytarabine. Also, previous studies using MCL cell lines show that there is a correlation between the expression of ENT1 and sensitivity to nucleoside analogues such as gemcitabine [
26]. In childhood AML patients, sensitivity towards nucleoside analogues could be linked to ENT1 mRNA levels, with decreased levels of ENT1 in cytarabine resistant compared to sensitive patients [
27]. However, it is not clear from these studies whether the down-regulation of ENT1 is a bystander effect, as our results indicate, or the limiting factor for drug sensitivity.
In the current study, we show that the final step of resistance development is associated with a complete down-regulation of the dCK protein. To investigate if the cellular levels of dCK protein is regulated at the transcriptional level, we performed gene expression analysis followed by RT-PCR validation. We show that dCK mRNA is down-regulated upon exposure and adaptation to cytarabine, but with minor additional down-regulation upon resistance development, indicating that post-translational modification may be involved in the final steps of resistance development. The relation between the dCK mRNA level and enzymatic activity has been explored previously. For example, using gemcitabine treatment in tumours of different origin such as pancreas and lung, it was shown that resistance only is predicted by dCK activity and protein level, and not by dCK mRNA level [
28], in agreement with our data. DCK enzymatic activity is controlled by phosphorylation at Ser-74 [
29], and dephosphorylation decreases enzyme activity [
30]. A more active mutant of dCK, with a 10,000-fold increased sensitivity to nucleoside analogues has been created and could be of interest for suicide gene approaches [
10,
31]. Of note, in our hands forced overexpression of dCK had a direct effect on sensitivity to cytarabine in our model, indicating that protein levels govern resistance but that post-translational modification adjust the level of resistance.
To further molecularly characterize the two distinct phases of (i) cytarabine adaptation and (ii) resistance development, global gene expression analysis was performed. Analysis reveals that major changes (68 genes) in gene expression could be associated with the initial adaptation to growth in cytarabine containing media. In contrast, only a few genes (FABP5P7, FABP5P1, FABP5, SPIB and TCEA3) showed expression changes upon final resistance development. Among those genes, SPIB showed major changes on the protein level. It is known from previous reports that SPIB is frequently overexpressed in diffuse large B cell lymphomas (DLBCL), and is a poor prognostic factor [
22]. Of note, SPIB and IRF4 have been shown to amplify NF-ĸB signalling by transactivating CARD11 [
32]. SPIB has not previously been described in relation to MCL and the connection to increased NF-ĸB activity in our model is of major interest. NF-ĸB is highly active in MCL and is considered to be a key feature for the aggressiveness of MCL [
33]. Constitutively active NF-κB in MCL may be caused by different factors such as, chronic activation of the B cell receptor (BCR) or mutations in the Toll like receptor (TLR) signalling pathway. When the activation is a result of somatic mutations on inhibitors of the alternative pathway, the activation is unaffected by BCR inhibitors [
4]. To pursue if the increased aggressiveness of the resistant cells could be related to enhanced NF-ĸB activity, related proteins were investigated. Analysis showed that the resistant cells have elevated levels of both NF-ĸB total protein as well as IκBα, indicating increased engagement of the pathway in dCK
negative/SPIB
high cells exhibiting resistance to cytarabine. Of major interest, forced overexpression of dCK led to decreased SPIB and NF-ĸB levels showing for the first time that dCK not only have an impact on the conversion of pro-drugs into active drugs, but also contribute to resistance through direct transcriptional control of pathways involved in proliferation and apoptosis. DCK as a key-driver was further pin-pointed by the fact that knock-down of SPIB did not affect dCK or NF-ĸB levels, although cytarabine resistance was slightly affected. Thus, SPIB does not regulate dCK levels but is likely associated with increased resistance through other means.
As NF-ĸB seem to play a major role in resistance we assessed if drugs known to be able to affect NF-ĸB activity, such as the immune-modulating drug lenalidomide [
34] and the Bruton tyrosine kinase (BTK) inhibitor ibrutinib [
35], had increased activity in the resistant cells. However, it was only at high concentrations of lenalidomide and Ibrutinib that the resistant cells had a significant increased sensitivity to cytarabine treatment.
In order to identify clinically relevant drugs effective in cytarabine resistant cells, a library of chemotherapeutic compounds was used. Eleven drugs, some predominantly used in treatment of solid cancers, and to less extent in treatment of leukemias and lymphomas, showed strong anti-proliferative effect on cytarabine resistant cells. Interestingly, four of the eleven drugs have already been considered in treatment of MCL, either as part of standard therapy (doxorubicin and vincristine) or as part of clinical trials including refractory MCL (mitoxantrone [
36] and etoposide [
37]). Also previous studies on cytarabine resistant leukaemia cells show that such cells are sensitive to vincristine and mitoxantrone [
38], supporting the clinical usefulness of these drugs for treatment of cytarabine-resistant disease. Novel drugs that may have a clinical role for treatment of cytarabine-resistant disease included topotecan HCl, teniposide, oxaliplatin, paclitaxel, docetaxel and cerubidine.
As expected, cytarabine resistant cells showed cross-resistance to other nucleoside analogues, including gemcitabine, cladribine and fludarabine, emphasizing the specificity of our model, development of nucleoside analogue-resistant cells and not a general increased drug resistance [
12,
14].
It has been proposed that the down-regulation of dCK is mediated by epigenetic mechanisms [
39,
40] and in order to evaluate if epigenetic drugs have a potent (i) stand-alone effect or potentially, (ii) a sensitizing effect, a library with epigenetic compounds was evaluated. Of interest, most of the compounds that showed inhibitory effect on cytarabine resistant cells were HDAC inhibitors. The potential of these compounds to re-sensitize resistant cells to cytarabine was assessed through pre-incubation prior to cytarabine exposure. Some of the compounds showed a minor tendency to increase cytarabine sensitivity after pre-treatment with low concentrations of HDAC inhibitors. Potentially, a longer pre-treatment would have generated a larger impact on cytarabine resistance. HDAC inhibitors generally have low toxicity and although further pre-clinical studies are needed, inclusion of such inhibitors in combination with cytarabine should be considered.
Our present data pinpoints the importance of dCK for resistance development, and thus ways of preventing and restoring dCK activity would have a major clinical impact for treatment of MCL patients. Several previous studies have used different approaches to increase dCK activity and thus revert cytarabine resistance. Among others, it has been shown that dCK was silenced through promoter DNA methylation, and that demethylation can restore dCK levels [
39]. However, in our hands neither decitabine nor hydralazine resulted in restored sensitivity.
From our data, it is clear that dCK activity is governed by mechanisms at the protein level and we hypothesized that co-treatment with bortezomib, a proteasome inhibitor, could prevent down-regulation of dCK and development of resistance. Bortezomib, is a reversible proteasome inhibitor, primarily used in MCL as part of combinatorial therapy and approved for treatment of relapsed MCL [
41]. Inhibition of the NF-κB pathway, leading to apoptosis, has been reported as a major mechanism of action for bortezomib [
42] and that correlates well with our model, showing increased sensitivity to bortezomib (at concentrations above 0.005 μM) in resistant cells compared to sensitive. Interestingly, the effects of bortezomib were much more prominent than for both lenalidomide and ibrutinib. The difference may partly be explained by the fact that the targets of bortezomib affects proteins more directly involved in NF-κB activation and regulation. However, being a proteasome inhibitor the action of bortezomib is not specific for the NF-κB pathway and several other cellular targets and pathways may be affected. Co-treatment with bortezomib prevented the development of cytarabine resistance and constitute a very attractive complement to cytarabine treatment in clinical protocols. Proliferation was assessed and no difference in growth rate was observed between cells co-cultured with both bortezomib (at 0.001 μM) and cytarabine compared to cells treated with cytarabine alone, ruling out the possibility that the resistance was prevented due to decreased proliferation. To our knowledge, this is the first time that it has been shown that cytarabine resistance can be prevented, which is of major clinical importance. As bortezomib is an already approved drug, implementation into the clinic should be rapid. It should be noted however, that the concentrations used in our study is in the low-end of the bortezomib serum levels measured in patients treated within current dosing regiments [
43]. Previous studies have demonstrated a synergy effect of bortezomib and cytarabine both in vitro and in individual case reports [
44] and also as part of a small combinatorial therapy trial [
45]. A clinical trial is currently ongoing, where bortezomib was combined with cytarabine for treatment of relapse patients, and results may shed light on the efficacy of this co-treatment [
46]. Diagnostic/pre-treatment levels of dCK have in a previous study been shown to predict in vivo gemcitabine sensitivity in human tumour xenografts (pancreas, colon, ovarian cancer) [
28]. Thus, using a cohort of patients treated with the combinatorial Nordic MCL2/3 protocol, where 95% of the patients are progression free ≥6 months after therapy including high-dose cytarabine, we investigated the diagnostic levels of dCK. The dCK levels were automatically scored to identify both frequency of positive tumour cells, but also intensity of nuclear dCK levels. Ninety seven percent of the patients showed high or intermediate expression of dCK, corresponding to the high clinical response to the cytarabine-containing combinatorial protocol. Thus, the number of dCK negative/weak patients were too few to evaluate impact on survival. Today, material from relapsed MCLs is not available but constitute an important work material to study to validate our finding of down-regulation of dCK and development of cytarabine resistance.