Treosulfan, combined with fludarabine is used clinically to condition patients with acute myeloid leukemia (AML) or myelodysplastic syndrome (MDS) for hematopoietic cell transplantation. Preliminary data suggest that the addition of low dose total body irradiation (TBI) to that regimen reduces relapse incidence. We determined, in vitro, whether treosulfan enhanced the effect of radiation.
AML-and MDS-derived myeloid cell lines as well as primary marrow cells from patients with MDS and healthy donors were exposed to treosulfan, radiation or both, and the extent of apoptosis was assessed. Pathways possibly mediating and enhancing treosulfan effects were probed.
The extent of cell kill in cell lines varied considerably. Cell kill was greater in MDS-derived than in healthy marrow cells. The increase in cell kill with treosulfan plus radiation exposure was greater in CD34+ blasts than in total marrow cells. There was dysregulation of miR9, miR7g and NF-kB, but pre-exposure to treosulfan did not clearly enhance radiation-induced cell kill.
These results do not support the hypothesis of treosulfan- induced radiosensitization in myeloid cells. Reduced relapse incidence observed clinically with a treosulfan/total body irradiation regimen is likely achieved by different mechanisms, possibly involving cell populations that were not present in vitro.
Keywords: Treosulfan; Radiation; MDS; Myeloid Cell Lines; Apoptosis
Treosulfan (Treo) is a bifunctional alkylating agent originally approved for the treatment of ovarian carcinoma . The observed dose-limiting toxicity of myelosuppression led to investigations on its usefulness as a conditioning agent in preparation for hematopoietic cell transplantation (HCT) . Pilot studies yielded encouraging results, and reports from several institutions show that Treo combined, for example, with fludarabine, can provide effective pre-transplant conditioning with minimal toxicity [3,4]. However, in patients with myeloid malignancies, including acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS), post-transplant relapse has remained an issue, particularly in patients with high-risk cytogenetics . Based on pre-clinical observations that Treo possessed some radiosensitizing capacity  we conducted a trial combining fludarabine plus Treo with 2 Gy of total body irradiation (TBI). This study  showed that patients with AML or MDS with high-risk cytogenetics experienced a lower relapse incidence and a higher relapse-free survival than observed in an earlier trial, without inclusion of TBI in the conditioning regimen . These observations led to a follow-up trial (ongoing) comparing the efficacies of Treo plus fludarabine regimens with and without inclusion of TBI.
In the present study we aimed at further characterizing, in vitro, potential interactions between Treo, fludarabine and TBI, possibly leading to further refinement of Treo-based regimens.
Materials and Methods
Myeloid Cell Lines and Primary Marrow Cells
Myeloid KG1a cells were obtained from ATCC (Rockville, MD), PL-21 cells were a gift from Dr. D. Stirewalt (Fred Hutchinson Cancer Research Center [FHCRC]), and MDS-L cells were a gift from Prof. Tohyama (Hamamatsu University School of Medicine, Japan).All cell lines have been used in previous studies in our laboratory .
KG1a cells were cultured in RPMI medium with 10% heat-inactivated fetal bovine serum (FBS), 1% sodium pyruvate (Invitrogen, Carlsbad, CA.) and 1% Penicillin/streptomycin (P/S). PL-21 cells were cultured in RPMI plus 20% FBS, 1% sodium pyruvate and 1% P/S. MDS-L cells, CD34+ cells (or mononuclear cells) from the marrows of healthy donors or patients with MDS were cultured in Stemspan (STEMCELL Technologies Inc., Vancouver, BC) plus 100 ng/ml of IL-3 (Peprotech, Rocky Hill, NJ, USA). All cell lines were maintained at 37oC in a 5% CO2 containing atmosphere.
Mononuclear cells from fresh marrow aspirates were prepared using Lymphocyte Separation Medium (mediatech Inc., Herndon, VA), according to the manufacturer’s instructions. CD34+ cells were purified by magnetic bead sorting (Miltenyi Biotec, Auburn, CA), following the manufacturer’s protocol.
All healthy donors and MDS patients had given informed consent according to the process required by the Fred Hutchinson Cancer Research Center (FHCRC).
Treo was provided by Medac (Wedel, Germany) and used at concentrations of 1 to 100 μM. Etanercept was purchased from Genentech (Oceanside, CA) and used at concentrations of 4 to 40 μM. IKK inhibitor 4 (IKK4) was obtained from EMD Biosciences (San Diego, CA) and was used at a concentration of 300 nmol.
Fluorescein isothiocyanate (FITC) conjugated Annexin V was purchased from Becton Dickinson, San Jose, CA.and was used to identify apoptosing cells. Antibodies to label CD34+ cells [phycoerythrin (PE)-conjugated] were obtained from Becton Dickinson, San Jose, CA.
Radiation (RT) was delivered from a 137Cs irradiator at a rate of 185 cGy/min for total doses of 200 to 2,000 cGy.
Assays for in Vitro Toxicity of Chemotherapy and Radiation
To test the effects of Treo and RT on apoptosis and gene expression, cell lines or primary CD34+ marrow cells, placed in culture medium, were exposed to RT only, to Treo only, or to both, either Treo preceding RT or RT preceding the addition of Treo. Cells were analyzed over a post-treatment time course of 1 to 5 days (aimed at mimicking the duration of conditioning in clinical protocols prior to donor cell infusion). In some experiments we also used fludarabine, which in clinical practice is frequently combined with either Treo, TBI or both.
To assess the extent of apoptosis, primary marrow cells or cell lines, after treatment with Treo, RT or both (or sham-treated), were labeled with Annexin-FITC to assess the extent of apoptosis. In addition, propidium iodide (PI) was used to assess the total amount of dead cells. In some experiments cells were labeled with PE-conjugated anti-CD34 antibody as described  to allow for sorting and assessment of the impact of Treo and RT on this specific cell population in comparison to unfractionated marrow mononuclear cells.
Blockade of Treatment-Induced Cell Injury
In some experiments, etanercept, to block TNF-mediated signals, or IKK2-IV, to inhibit NF-κB activation, were added to cultures prior to Treo or RT exposure.
RNA Extraction and Real Time PCR
Total RNA was isolated from MDS-derived primary bone marrow CD34+ cells and myeloid cell lines using the RNAeasy Minikit (Qiagen, Foster City, CA). cDNA for mRNA and miRNA (miR) was synthesized from total RNA using the MicroRNA reverse transcription kit (Applied Biosystems, Foster City, CA). Levels of let-7g (hs002282, Applied Biosystems, Foster City, CA) and miR9 (hs000583, Applied Biosystems, Foster City, CA) expression were determined (in biological triplicates) using Taqman and PCR Master mix (Applied Biosystems, Foster City, CA) on the ABI 7500 Real time PCR system (Applied Biosystems, Foster City, CA) for 40 cycles. β-actin and U6B served as internal controls for mRNA and microRNA (miR), respectively. The levels of mRNA and miRs in each sample were normalized to controls and recorded as relative expression levels.
Ninety to 300 ng of RNA were used for each determination using specific primers for let7g, miR9 and U6B .
qPCR results were analyzed using cycle-threshold (ct) values. Briefly, the difference between the Ct values (ΔCt) of the gene of interest and the housekeeping gene was calculated for each experimental sample. Then, the difference in the ΔCt values between the experimental and control samples, ΔΔCt, was calculated. The fold-change in expression of the gene of interest relative to control equals 2 (-ΔΔCt).
Total RNA integrity was tested using an Agilent 2200 TapeStation (Agilent Technologies, Inc., Santa Clara, CA) and was quantified using a Trinean DropSense96 spectrophotometer (Caliper Life Sciences, Hopkinton, MA). Briefly, high-quality RNA samples were converted to cDNA and biotin-labeled for microarray analysis, carried out using Ambion’s Illumina TotalPrep RNA Amplification kit (Life Technologies, Grand Island, NY) . Labeled cDNAs were processed on a HumanHT-12 Expression BeadChip (Illumina, Inc., San Diego, CA) and imaged using an Illumina iScan system. Microarray data were assessed for quality and quantile-normalized using the Bioconductor package lumi. Initial filtering included flagging probes that were below a signal “noise floor,” which was calculated as the 75th percentile of the negative control probe signals within each array. We subsequently filtered the data set by employing a variance filter using the “shorth” function of the Bioconductor package genefilter. As previously described,  differential gene expression was determined using the Bioconductor package limma, and a false discovery rate (FDR) method was used to correct for multiple testing. Significant differential gene expression was defined as |log2 (ratio) | ≥ 0.585 (±1.5- fold) with the FDR set to 5%.
In preparation for protein extraction, 5 × 106 cells/10 mL/ well were plated in 12-well plates. Total cell lysates and nuclear extracts were assayed using the Nuclear Extraction Kit, according to the manufacturer’s instructions (Panomics, Fremont, CA). Fractions were cleared by centrifugation at 13,000g for 10 minutes. Protein concentrations were quantified by bicinchoninic acid assay (Pierce Biotechnology Inc., Rockford, IL, USA), and equal amounts of protein (50 μg) from each lysate were diluted in Laemmli sodium dodecyl sulfate sample buffer and resolved by electropheresis on NuPAGE® 4−12% Bis-Tris Mini Gels (NOVEX by Life Technologies) in running buffer (50mM 2-(N-morpholino), ethane sulfonic acid, 50mM Tris [tris(hydroxymethyl)aminomethane]) base, 0.1% sodium dodecylsulfate, and 1mM EDTA [ethylenediaminetetraacetic acid]). For immunoblotting the proteins were transferred to polyvinylidene difluoride membranes. The membranes were blocked with 5% nonfat dry milk diluted in Tris-buffered saline containing antibodies for the following targets: NFkB (p65, p105, p50) and β Actin. All methods have been described before .
Experiments were designed to uncover biologic patterns that may deserve more in depth investigations. Those results are presented here. The focused approach with limited numbers of experiment was not amenable to rigorous statistical analysis.
Apoptosis in Myeloid Cell Lines
We first determined the extent to which Treo and RT alone, and combinations of both modalities induced apoptosis in KG1a, PL21, and MDS-L cell lines. Next, Treo was added to cells plated in 24-well plates in complete medium on day 1 of culture, followed by exposure to RT on days 2 or 4, and flow cytometric analysis 24 hours later (day 5). In further experiments, the sequence of RT and Treo was reversed, i.e. cells were irradiated on day 1, and Treo was added on day 2, with flow cytometric analysis carried out on day 5.
Spontaneous apoptosis in KG1a cells was in the range of 8.8±2.1%. Addition of Treo at doses of 1, 3, 10, 30, or 100 μmol progressively increased the rate of apoptosis to as high as 74.6±26.7%.
RT alone, given at 2, 3, or 20 Gy induced apoptosis in up to 16.3% of cells. Administration of RT to Treo-exposed cells consistently increased the extent of apoptosis, but only minimally, by 7–20%, above that observed with Treo alone, dependent upon the dose of RT (Figure 1A).
The pattern of apoptosis was similar for PL21 cells, although cells exhibited greater sensitivity, particular to RT. As with KG1a cells, the addition of RT resulted in only minimal increments above the extent of apoptosis observed with Treo alone (Figure 1B,1C).
Figure 1. Extent of cell death (apoptosis) induced by Treo, radiation, fludarabine or combinations thereof. (A) KG1a apoptosis induced by Treo alone (T1…T100), radiation alone (R200…R2000) or combination of both. (B) PL21 apoptosis induced by Treo (T 0…T100). (C) PL21 apoptosis induced by Treo (T10), alone, radiation (300cGy) alone (R300) or by R followed by Treo and Treo followed by R. (D) MDS-L apoptosis induced by Treo or radiation (symbols as in (A)) or fludarabine (Flud/F 0.1, 0.3, 1.0) alone or combined with radiation (R300). (E) MDS-L apoptosis, time course day 1 to day 5, after exposure to Treo, radiation ® or fludarabine (Flud). Shown are the mean ± SD of 3–4 experiments. No Tx = no treatment.
Similarly, MDS-L cells first exposed to RT (day1), then to Treo on day 2, and analyzed by flow cytometry on day 3, showed an increase in apoptosis by approximately 7% above that observed with RT alone. The maximum extent of apoptosis was 24±11%. This result did not support an enhancing of RT on Treo-induced apoptosis and, in fact, suggested the possibility that initial RT exposure modified gene expression in a direction that would lead to a reduced sensitivity of cells to Treo.
Effect of Fludarabine
Since clinically Treo has been used frequently in combination with fludarabine (as well as TBI) [3,7], we also tested the potential effects of this drug on Treo responses.
As shown in Figure 1D, at concentrations as low as 1 μmol, fludarabine by itself induced apoptosis in 46.3±26.5% of MDS-L cells (background apoptosis in the absence of drugs or RT was 11.6±2.8%). The proportion increased to 60.0±21.2% with 3 Gy of RT given after fludarabine.
In KG1a cells the rate of apoptosis in response to fludarabine, at comparable molar concentrations, was substantially higher than with Treo (Figure 1E) and was additive when both drugs were combined (not shown). The addition of RT did not appear to further increase the rate of apoptosis beyond the maximum cell kill achieved with fludarabine alone or with the fludarabine plus Treo combination.
Primary Marrow Cells
Based on clinical observations , we speculated that MDS blasts/hematopoietic precursor cells may be particularly sensitive to a combination of Treo and RT. As shown in Figure 2A-B, using CD34 positivity as a marker for blasts/hematopoietic precursor cells, apoptosis following exposure to Treo, RT or both occurred in a smaller proportion of those cells than in the total mononuclear cell population. In MDS-derived marrow cells, however, exposed to 10 μmol Treo followed by 3 Gy RT, there was a proportionally greater increase in apoptosis in sorted CD34+ cells than in unfractionated marrow mononuclear cells. (26.5±20.3 - 35.5±14.7% vs. 71.3±10.7 - 78.4±10.1%, i.e. an increment of 34.0% vs. 10.0% increment relative to Treo alone.
Overall the rate of apoptosis was lower in marrow cells from healthy donors than observed in MDS marrow cells (Figure 2B). Exposure of CD34+ cells to Treo (10 μmol), followed by RT (3 Gy) resulted, in fact, in a lesser increase in apoptosis than that observed in total mononuclear cells (23.2±24.6 to 24.6±27.4% vs. 48.2±19.2 to 56.0±11.0%, i.e. an increment of 6.0% vs. 16.2%).
The addition of etanercept (to neutralize TNF α, which is upregulated in MDS) or IKK4 (to interfere with NFκB activation), did not significantly alter the rates of apoptosis (data not shown).
Figure 2. Cell death (apoptosis) in primary marrow cells.
(A) MDS marrow (N=6). (B) Healthy marrow (N=4). Cells were exposed to Treo at 10 μM (T10), radiation at 300 cGy (R300), or both combined, either Treo followed by radiation (T>R) or radiation followed by Treo (R>T). In addition, the soluble TNF receptor etanercept (E, 20μM) or IKK4 (IKK4 300 nmol) was added; shown are the means ± SD of the extent of apoptosis of 4–6 experiments. No Tx = no treatment. The “Blast” population (shown in black) accounted for 5% to 30% of all mononuclear cells; more than 90% of blasts expressed CD34.
Analysis of gene expression (on Illumina chips) in the MDS-derived cell line MDS-L showed clear differences between cells exposed only to Treo or RT, compared to a combination of both; the overall landscape, however, appeared to be determined primarily by Treo exposure. There was significant upregulation of CHK1, CHK2, P21, Cyclin E and Cyclin B, suggesting a central involvement of P53-dependent signaling pathways (not shown), and consistent with a DNA damage response and regulation of cell death.
While the expression of numerous genes was not measurably affected by RT alone, nor modified by RT given after Treo, the expression of other genes or gene products, including the NFκB complex and TNFRS14, was clearly enhanced when RT was administered following Treo exposure (Table 1). Such a pattern was consistent with the altered cytokine profile present in MDS (and MDS-derived cells) and observed after transplant conditioning .
Table1. Microarray Analysis (TNF kit) Of Apoptosis-Related Genes in the MDS-L Cell Line#.
# All experiments were carried out in triplicates; shown are “fold” changes induced by Treo (treosulfan) alone, Radiation alone or both modalities combined.
No tx.= no treatment; base = baseline
MicroRNAs (miRs) are involved in the regulation of hematopoiesis and in the pathophysiology of MDS [8,15]. Some miRs, such as miR9 and miR7g have been reported to mediate resistance or sensitivity to gamma radiation.10 In PL21 cells, exposure to 3 Gy RT had no significant effect on miR7g and reduced levels of miR9 (0.67-fold expression) compared to baseline levels. Conversely, Treo-exposure alone (at 3 μmol) induced a 2.17-fold increase of miR7g, and a 1.40-fold increase of miR9, compared with exposure to Treo followed by RT, which resulted only in a minimal 0.65-fold increase in miR7g and a 0.51- fold increase in miR9, suggesting that RT (rather than Treo) was the modality determining the overall response.
However, patterns differed between cell lines. In MDS-L cells, 3 Gy of RT resulted in a 1.69-fold increase in miR7g and a 2.45- fold increase in miR9. There was a slightly further increase, to 3.45 for miR9 if Treo was given before RT. In KG1a cells, 3 Gy of RT resulted in 1.57-fold increase of miR7g and a 0.43-fold increase in miR9. The effect was reduced to 0.83 for miR7g and 0.25 for miR9 if Treo was given before RT (Figure 3A-B). Thus, cell line data showed great heterogeneity, cautioning against extrapolation to primary cells.
Primary MDS Marrow Cells
As illustrated in Figure 4A-B, in primary MDS marrow cells, Treo induced a 2.21-fold increase in miR9, compared to baseline. Following 3 Gy of RT, there was a decrease in miR9 to 0.20 of baseline levels.
Treatment with Treo followed by RT induced miR9 levels similar to those seen after Treo alone, representing a substantial increase in the ratio of expression compared to RT alone.
MiR7g (Figure 4A) showed little change after Treo alone, but there was up-regulation following RT; there was no increase when RT was preceded by Treo, suggesting an inhibitory effect of Treo.
Healthy Donor Marrows
In marrow cells from healthy donors, both Treo (???) and RT treatment resulted in up-regulation of miR9 (2.29 fold), which was not further enhanced when Treo was followed by RT (Figure 4B). There was increased miR7g expression following either Treo or RT exposure, which appeared to be further enhanced with sequential exposure to Treo and RT (7.57) (Figure 4A).
Thus, while apoptotic responses to Treo and RT in myeloid cell lines were heterogeneous, primary marrow cells from healthy individuals showed up-regulation of miR7g and miR9, consistent with increased sensitivity to ionizing radiation (see ref. ). In cells derived from MDS marrow, neither Treo nor RT (nor a combination of both) substantially altered miR7g levels, while miR9 showed increased expression after Treo but not after RT; the effect of Treo followed by RT was not different from that of Treo alone. As NFκB has been shown to be down-regulated by miR7g and miR9 (the expression of which was modified by radiation exposure as shown above), we exposed KG1a cells to 2 Gy RT and determined levels of NFκB. As shown in Figure 5, NFκB mRNA was increased by ~20% following RT exposure, while Treo alone had no measurable effect. In MDS-L cells, in contrast, NFκB levels declined by 20%. There was a slightly greater decrease when RT was preceded by Treo. The data suggest that additional pathways are likely to be involved in these responses.
Figure 3. Expression of microRNAs in myeloid cell lines KG1a, PL21 and MDS-L exposed to Treo, Radiation or both. A) miR7g; B) miR9. Shown are “fold” changes in comparison to untreated cells (No tx), mean ± SD of three experiments. T=treosulfan (3 μmol); R=radiation (300 cGy). T+R vs R indicates the ratio of combined treatment versus radiation exposure alone.
Figure 4. Expression of microRNAs in primary marrow cells from MDS patients (n=5) and healthy donors (n=4).
A) miR7g; B) miR9. Cells were untreated (No tx) or exposed to Treo (3μmol), radiation (R, 300 cGy) or both (T+R). T+R vs R indicates the ratio of combined treatment Theversus radiation exposure alone. Shown are the means ± SE of “fold” changes in 4–5 experiments after exposure relative to untreated cells. microRNA levels were determined at the end of the experiment (see Methods). NL BM = marrow from healthy (nl) donors; MDS BM= marrow from patients with myelodysplastic syndrome.
Figure 5. NF-kB Protein Expression in MDS-L Cells. Protein lysates from MDS-L cells, were separated on 4%–12% Bis-Tris gel and immunoblotted with antibodies against phosphorp50, p105 and β-actin (loading control), respectively. Shown is one of two similar experiments.
Treo has been introduced into the clinic as part of conditioning regimens for HCT. Results have been encouraging in regards to relapse prevention. Preliminary outcome data suggested that the addition of 2 Gy of TBI to a Treo plus fludarabine combination, compared to Treo plus fludarabine without the addition of TBI, improved relapse-free survival for patients with high risk cytogenetics . Additional preclinical work suggested that Treo has a radio-sensitizing effect . The current experiments were aimed at further characterizing potential interactions of radiation and Treo that might improve our understanding of the clinical results. Our data in cell lines reveal a complex picture. While sensitivity to either Treo or RT varied from cell line to cell line, cell kill occurred in a dose- and time-dependent pattern. At “sub-optimal” concentrations of Treo, the addition of fludarabine significantly increased cell kill, which was further enhanced with the incorporation of RT. The presence of etanercept (to block TNFα) or IKK4 (to interfere with activation of NFκB) did not significantly impact the extent of cell kill, in either MDS –derived or healthy donor marrow cells. There was evidence, however, at least in MDS-L cells, that treatment with RT increased levels of NFκB, a response that was further enhanced if RT was preceded by Treo exposure. Differences in sensitivity to the cytotoxic effects of Treo, RT or both, are, of course, not surprising as clinical results in many studies show very broad variability in regards to eradication of clonal cells in patients with MDS . This heterogeneity and complexity of cytotoxic response regulation was further emphasized by the studies on the effect of microRNAs, which varied substantially between cell lines.
More re-assuring were results in primary human marrow cells, which showed greater consistency in support of interactions between Treo and TBI (± fludarabine). However, contrary to our hypothesis, there was no differential effect on cell kill in MDS precursor cells in comparison to marrow cells from healthy donors, and the literature pertaining to interactions of Treo with RT and potential cell selectivity is rather limited. While Sender et al. proposed a radiosensitizing effect of Treo in pre-clinical models, the mechanism has remained unclear.6 Even clinical data are limited as most transplant regimens have combined Treo with chemotherapeutic agents, such as fludarabine or cyclophosphamide, but not RT [4,16,17]. There are no data from a prospective randomized trial comparing the efficacy of conditioning regimens including Treo with or without the addition of RT. A randomized phase II trial, including molecular analysis of marrow cells exposed to Treo alone or Treo plus RT is currently underway at our Center. Results from that trial may shed light on the interactions of Treo and RT at the molecular level.
In conclusion, these experiments characterize cellular and molecular effects of Treo and RT on myeloid cell lines and primary MDS-derived and healthy marrow cells. Results illustrate differential effects on MDS-derived and healthy donor marrow cells, showing a higher total cell kill in MDS marrows compared to healthy marrow cells. However, these studies fail to show, in vitro, a clear radio-sensitizing effect of Treo on cells in the “blast gate”, presumably containing the relevant clonal precursor cells, in comparison to all marrow cells, and further work, addressing the mechanism of action, is required. Ongoing studies analyzing gene expression in marrow cells before transplant conditioning, and again after completion of conditioning with either Fludarabine + Treo or Fludarabine + Treo + RT, are expected to provide further insights on signaling pathways affected by the two conditioning strategies.
We would like to thank Helen Crawford and Bonnie Larson for help with manuscript preparation and Joachim Baumgart, PhD, for providing the treosulfan for these experiments. Dae Zang, MD, PhD was supported by a grant from Hallym University, Anyang, South Korea. Supported in part by Grants no. NIDDK K08 DK085156-01 (AMM), HL036444 (HJD).
Conflict of Interest
Treosulfan for these studies were provided as a gift by Dr J. Baumgart from medac GmbH, Wedel Germany. The authors declare no other conflict of interest. (DISCLSURES: NONE)
This article does not contain any studies with human participants or animals performed by any of the author