Assessment of the Chemosensitizing Activity of TAT-RasGAP317-326 in Childhood Cancers

Although current anti-cancer protocols are reasonably effective, treatment-associated long-term side effects, induced by lack of specificity of the anti-cancer procedures, remain a challenging problem in pediatric oncology. TAT-RasGAP317-326 is a RasGAP-derived cell-permeable peptide that acts as a sensitizer to various anti-cancer treatments in adult tumor cells. In the present study, we assessed the effect of TAT-RasGAP317-326 in several childhood cancer cell lines. The RasGAP-derived peptide-induced cell death was analyzed in several neuroblastoma, Ewing sarcoma and leukemia cell lines (as well as in normal lymphocytes). Cell death was evaluated using flow cytometry methods in the absence or in the presence of the peptide in combination with various genotoxins used in the clinics (4-hydroperoxycyclophosphamide, etoposide, vincristine and doxorubicin). All tested pediatric tumors, in response to at least one genotoxin, were sensitized by TAT-RasGAP317-326. The RasGAP-derived peptide did not increase cell death of normal lymphocytes, alone or in combination with the majority of the tested chemotherapies. Consequently, TAT-RasGAP317-326 may benefit children with tumors by increasing the efficacy of anti-cancer therapies notably by allowing reductions in anti-cancer drug dosage and the associated drug-induced side effects.


Introduction
Cancer represents the second cause of death in children after accidents in industrialized countries [1,2]. Our understanding of childhood cancers has benefited from significant advances over the four last decades. Standard treatments to cure pediatric tumor include surgery, radiation therapy and intensive multi-agent chemotherapy such as etoposide, vincristine, doxorubicin, and cyclophosphamide [3].
In developed countries, eighty percent of children who are diagnosed with cancer are expected to survive within 5 years following the treatment. However, most of them will suffer from chronic diseases by 40 years of age [4]. Extended surveillance of pediatric cancer survivors shows a high risk for life-threatening late effects from second malignancies, cardiac conditions and pulmonary diseases [5,6]. The risk of early mortality is mostly determined by treatmentspecific factors such as the cumulative dose of chemotherapy [7]. Therefore, treatmentassociated long-term side effects induced by damage to non-tumor cells are a challenging problem and remain largely unresolved. As children are by definition long-term survivors, there is a strong need to develop low-toxicity, better targeted and efficient new therapeutic strategies for all types of childhood cancers [8]. Strategies to circumvent such obstacles include the improvement of anti-cancer drug sensitivity and specificity toward cancer cells [9][10][11].
In this context, we previously reported that a cell-permeable peptide derived from the p120 RasGAP protein, called TAT-RasGAP 317-326 , is a tumor-sensitizer to various anti-cancer drugs. Indeed, although it does not show any toxicity toward cells on its own, it efficiently and specifically sensitizes adult tumor cells in vitro and in vivo to various anti-cancer therapies, including chemotherapy [12,13], and photodynamic therapy [14]. Importantly, it displays specificity to cancer cells as it does not sensitize non-tumor cells to genotoxin-induced apoptosis [12,14]. TAT-RasGAP 317-326 appears to have additional anti-cancer activities than tumor cell sensitization as it has been recently demonstrated that this peptide can hamper cell migration and invasion in vitro [15,16] and that this activity can inhibit the metastatization process in vivo [17]. This indicates that the RasGAP-derived peptide has the ability to act as an anti-metastatic compound on top of its tumor sensitization effects. Recently, it has been shown that the anti-cancer properties of TAT-RasGAP 317-326 are dependent on two tryptophan residues at position 317 and 319 [16]. However, the mode of action of TAT-RasGAP 317-326 is not fully characterized. It has been previously shown that this peptide does not favor the death of tumor cells by modulating Ras activity, MAPK signaling pathways, NF-κB transcriptional activity, Akt protein levels and phosphorylation status [18,19]. Moreover, the Bcl-2 family members, which regulate mitochondrial-dependent cell death, were shown to be individually dispensable for the sensitizing activity of the peptide [20].
The effect of this peptide in childhood cancer is however not known. The molecular biology of pediatric tumors is distinct from cancers in adults in many ways. As the genesis of most childhood cancers seems to come from disruptions of normal early development, they accumulate fewer mutations than adult tumors. On the other hand, it appears that development of pediatric tumors rely heavily on epigenetic modifications [21][22][23]. In the present study, we have therefore investigated whether TAT-RasGAP 317-326 was able to render childhood tumors more sensitive to clinically relevant anti-tumor drugs.

Cytotoxic assay
For cells growing in suspension (THP-1, M07e, CCRF-CEM, and lymphocytes), three hundred thousand cells were seeded in 6-well plates and directly treated with the indicated doses of 4-HC, etoposide, vincristine, or doxorubicin in the presence or in the absence of 10 μM TAT-RasGAP 317-326 , TAT, TAT-Scrambled or TAT-Mutated. For adherent cell lines (NB1-NBM, NB1-FBS, NB1-FBS-Re, LAN-1, EW-11, TC252, and A673), two hundred thousand cells were allowed to adhere for 48 hours in 6-well plates and were then treated like cells growing in suspension. Twenty-four hours after drug incubation, cells were subjected to flow cytometry to evaluate cell death. Briefly, adherent cell lines were detached by trypsin-EDTA (Gibco) and single-cell suspensions were processed and stained with 1 μg/ml 7-AAD and 50 ng/ml Annexin V-FITC in 500 μl binding buffer. A minimum of ten thousand cells were analyzed. The 7-AAD dye-derived signal could not be used when cells were treated with doxorubicin because 7-AAD and doxorubicin have similar fluorescent properties (7-AAD emission wavelength: 525 nm; doxorubicin emission wavelength: 530 nm [32]).

Statistical analysis
Unless otherwise mentioned, all experiments were derived from three or more independent experiments. The results were expressed as mean ± 95% confidence intervals. Significance was assessed using t-tests performed with Microsoft Excel 2010 followed by Bonferroni corrections (i.e. differences were considered significant when p values < 0.05/n, where n is the number of comparisons made). Asterisks represent statistically significant differences. In the last figure, significance was assessed by one-way ANOVAs followed by Bonferroni (Dunn) t post-hoc tests using the SAS 9.2 software (SAS Institute Inc., Cary, NC, USA).

Results
In this study, three types of childhood cancers were analyzed: neuroblastoma, leukemia, and Ewing sarcoma. Among childhood malignancies, neuroblastoma is the most frequent extracranial solid cancer. Deaths resulting from neuroblastoma account for 15% of all tumor related death in children [33]. Multidrug resistance, acquired during exposure to chemotherapies, plays a major role in this poor outcome [34,35]. Leukemia is the most common childhood cancer type and represents one third of all pediatric tumors [36]. Although less prevalent than leukemia, Ewing sarcoma is an aggressive cancer with a poor prognosis and a high rate of relapse with metastatic disease [37]. The drugs employed here are all used in the clinics. Their properties are presented in Table 1.

TAT-RasGAP 317-326 sensitizes acute lymphoblastic and myeloid leukemia cells to various chemotherapeutic agents without showing any effect toward tumor cells by itself
To investigate the sensitization effect of TAT-RasGAP 317-326 in leukemia cells, three different cell lines were subjected to increasing concentrations of cytostatic agents in the absence or in the presence of 10 μM TAT-RasGAP 317-326 . Etoposide, vincristine, doxorubicin, and 4-hydroperoxycyclophosphamide (4-HC), the active form of cyclophosphamide, were the drugs used here and are each currently employed in the clinic. Cell death was assessed using 7-AAD that labels dead cells that have permeabilized plasma membrane and with Annexin V that binds to phosphatidylserine exposed on dead cells. Some of the anti-cancer drugs used here induced a concomitant appearance of the 7-AAD and Annexin V signals ( Fig 1A shows the representative case of 4-HC). In other words, as soon as a cell became Annexin V-positive it also picked up the 7-AAD dye. Hence, in our hands, anti-cancer drugs such as 4-HC induced a necrosislike type of death. In the next figures, the 7-AAD data are reported as we found them to be associated with a lower variance than those obtained with Annexin V staining ( Fig 1B). The only exception was when doxorubicin was used. Indeed, this dye fluoresces at similar wavelengths as 7-ADD. In this case therefore, the Annexin V data are presented. These experiments revealed that TAT-RasGAP 317-326 significantly sensitizes leukemia cells to almost all tested drugs (Fig 2A). However, in some conditions (for example when THP-1 cells are treated with vincristine) tumor cells did not respond well to TAT-RasGAP 317-326 . A limited sensitization effect of the RasGAP-derived peptide is not necessarily a consequence of intrinsic resistance to a drug as the CCRF-CEM cell line, which is vincristine-resistant, was efficiently sensitized by TAT-RasGAP 317-326 . This point is of particular clinical relevance as it indicates that TAT-Ras-GAP 317-326 can exert a sensitization effect in chemo-resistant cells. TAT-RasGAP 317-326 by itself did not induce cell death in the tested leukemia cell lines, even at a two-fold higher concentration (20 μM) than the one used to induce a sensitization effect (10 μM) (Fig 2A).
TAT-RasGAP 317-326 does not display toxicity toward non-tumor lymphocytes from healthy subjects.
It was shown previously using non-tumor immortalized human keratinocytes (HaCaT) and umbilical vascular endothelium (HUV-EC-C) cells that TAT-RasGAP 317-326 , in combination with various genotoxins, did not sensitize non-tumor cells [12]. In the present study, we used isolated lymphocytes from whole blood of healthy patients to investigate the selectivity of the RasGAP-derived peptide toward cancer cells. We used the same dosages of chemotherapeutic agents to treat healthy lymphocytes as those used to treat the CCRF-CEM T-acute . Alternatively (first column), the cells were treated with increasing concentrations of TAT or TAT-RasGAP 317-326 alone. After 24 hours, the evaluation of cell death was carried out using 7-AAD staining. B. Isolated lymphocytes from three distinct healthy subjects were treated as described in Fig. 2A. The dosages of chemotherapeutic agents used to treat the healthy lymphocytes from the three subjects are similar to those used to treat the T-ALL CCRF-CEM cells. Note that the graphs are derived from single experiments where lymphocytes are immediately used after their isolation (if cultured in vitro, they would experience high levels of spontaneous apoptosis that would prevent accurate measurement of anti-cancer drug-and peptide-induced death). T-ALL, T-acute lymphoblastic leukemia; AML, acute myeloid leukemia; 4-HC, 4-hydroperoxycyclophosphamide. * p<0.05 t-test after Bonferroni correction. lymphoblastic leukemia (T-ALL) cell line (Fig 2A). Fig 2B shows the response of peripheral blood lymphocytes (PBLs) derived from three different healthy subjects (subjects A-C) to the RasGAP-derived peptide in combination or not with the genotoxins used in Fig 2A. The Ras-GAP-derived peptide by itself did not display any toxicity toward the PBLs. The sensitivity of the non-tumor lymphocytes for the tested chemotherapies alone was similar to the sensitivity of the CCRF-CEM cell line (even slightly increased in the case of doxorubicin for the healthy subjects A and C). The observation that cancer and normal lymphocytes were similarly sensitive to genotoxins was somehow surprising because the rationale to use a given chemotherapy is that it will target preferentially the malignant cells. Nevertheless, the important information drawn from Fig 2B is that TAT-RasGAP 317-326 does not sensitize PBLs to etoposide-, vincristine-and doxorubicin-mediated death. The peptide however did sometimes sensitize PBLs to 4-HC (Fig 2B), suggesting that TAT-RasGAP 317-326 can affect the viability of normal cells in some treatment combinations.

TAT-RasGAP 317-326 can potentiate genotoxin-induced cell death in Ewing sarcoma
To extend the investigation on TAT-RasGAP 317-326 in non-leukemia childhood cancer, the efficacy of TAT-RasGAP 317-326 was also tested in Ewing sarcoma cells. The results show that the RasGAP-derived peptide is also able to sensitize these cells to various genotoxins, although to a lower extent than in leukemias (Fig 3). In some cases, no sensitization was observed (e.g. when vincristine was used in the A673 and TC252 cell lines). Here again, TAT-RasGAP 317-326 alone did not display any toxicity toward the Ewing sarcoma cells (Fig 3).

The NB1 neuroblastoma derived-cell lines are directly killed by TAT-RasGAP 317-326
Acquired chemoresistance is an important cause of failure of treatment of neuroblastoma. To assess the effect of TAT-RasGAP 317-326 in cells that were exposed to high doses of chemotherapies and therefore that potentially have acquired resistance to these cytotoxic agents, we selected primary tumor cells of a single patient at different stages of the disease. The NB1 cells are high risk stage 4 primary neuroblastoma cells derived from bone marrow samples that were established at initial diagnosis (NB1-NBM and NB1-FBS) and at subsequent relapse after multiagent chemotherapy (NB1-FBS-Re) (Fig 4A). NB1-FBS and NB1-FBS-Re cell lines were cultured in 10% fetal bovine serum (FBS)-containing DMEM medium, while NB1-NBM cells were established in neural basic medium (NBM). NBM is a stem cell permissive serum-free, bFGF-, EGF-and B27-supplemented medium to support the growth of neural crest cells, the tissue of origin of neuroblastoma [38,39]. Fig 4A shows that the NB1-NBM and NB1-FBS cell lines do not display the same sensitivity toward the tested chemotherapies. Moreover, the efficacy of TAT-RasGAP 317-326 to improve the killing efficiency of 4-HC varies between these two cell lines: the NB1-FBS cells were found to be very sensitive to the sensitization effect of the peptide, while the NB1-NBM cells were not affected by the presence of TAT-RasGAP 317-326 . As the only difference between the NB1-NBM and NB1-FBS cell lines is the medium in which they are cultured, one could hypothesize that the metabolism of the NB1 cells varies from one culture condition to the other and that this differentially modulates the sensitivity of the tumor cells to TAT-RasGAP 317-326 .
To investigate whether the sensitization effect of TAT-RasGAP 317-326 was still efficient on relapsed neuroblastoma cells, we compared its efficacy between the NB1-FBS cells and the NB1-FBS-Re cells. Both cell lines are cultured in the same culture medium. The treatment regimen that was used to treat the patient is illustrated in Fig 4A. The patient, the relapsed cells of which were derived, was exposed to the four drugs (4-HC, etoposide, vincristine and doxorubicin) tested in this study. However, except for the 4-HC-treated conditions where we can observe a slight resistance to this drug in the NB1-FBS-Re cells, the relapsed cells are not more resistant to etoposide, vincristine and doxorubicin than the cells derived from initial diagnosis (Fig 4A). On the contrary, NB1-FBS-Re cells show an increased sensitivity to doxorubicin compared to the NB1-FBS cells. Nevertheless, TAT-RasGAP 317-326 sensitizes the three primary NB1-derived cell lines, although to a different level of efficacy according to the cell line and the tested chemotherapy (Fig 4A).
In most cases, TAT-RasGAP 317-326 does not induce tumor cell death by itself. However, we found out that the three NB1 derived-cell lines can be directly killed by the RasGAP-derived peptide (Fig 4A). The dose needed to induce cell death (20 μM) is nevertheless higher than the dose used to sensitize the cells to genotoxins (10 μM).
To complete the investigation of the effect of TAT-RasGAP 317-326 in neuroblastoma tumor type, we also tested the peptide on LAN-1, another neuroblastoma cell line derived from an aggressive stage 4 neuroblastoma. Fig 4B shows that the RasGAP-derived peptide displays a slight sensitization effect toward this cell line. Unlike the NB1 derived-cell lines, TAT-RasGAP 317-326 alone was not able to kill LAN-1 cells (Fig 4B), meaning that TAT-RasGAP 317-326 -induced cell death is not a specificity shared by all neuroblastomas.

The sensitizing activity of TAT-RasGAP 317-326 is carried by the RasGAPderived sequence
To assess the specificity of the chemo-sensitizing activity of TAT-RasGAP 317-326 , three different control peptides were tested: a peptide composed of the TAT sequence only (TAT), a peptide in which the RasGAP sequence was scrambled (TAT-Scrambled), and a peptide in which the first tryptophan of the RasGAP sequence was substituted into an alanine residue (TAT-Mutated). This tryptophan was recently shown to be essential for the sensitizing activity of TAT-RasGAP 317-326 in adult tumors [16]. The effect of TAT-RasGAP 317-326 and the three control peptides was tested in CCRF-CEM cells in combination with various anti-cancer drugs at doses that allowed the greatest sensitization activity to be detected. CCRF-CEM cells were used because this leukemia cell line was the one most efficiently sensitized by the RasGAP-derived peptide. Fig 5A shows that all three control peptides had a tendency to slightly favor the death of CCRF-CEM cells when combined with the different anti-cancer drugs but in most cases this did not reach statistical significance. In contrast, TAT-RasGAP 317-326 markedly, and always significantly, sensitized these cells to the drugs. The slight sensitization effect of the control peptides is most likely due to the cell penetrating activity of the TAT moiety, which has the potential to negatively affect cellular homeostasis [40]. Similar results were obtained when the TC252 Ewing sarcoma cell line and the NB1-FBS neuroblastoma cell line were used: TAT-Ras-GAP 317-326 significantly increased their sensitivity to 4-HC, while the three control peptides did not, or only minimally (Fig 5B and 5C). These data indicate that the tumor sensitizing activity of TAT-RasGAP 317-326 only marginally relies on its ability to penetrate cells via the TAT cell-permeable sequence. Therefore, it can be concluded that the sensitizing activity of TAT-RasGAP 317-326 is mainly carried by the RasGAP-derived sequence.

Discussion
The currently applied therapies to treat childhood solid cancers, such as neuroblastoma and Ewing sarcoma, need to be more efficient and specific to improve clinical outcome. Leukemia has usually a better 5-year survival rate than pediatric solid tumors, but long-term effects of the therapy remain an important cause of morbidity and mortality. Furthermore, acquired resistance to the conventional treatments is responsible for relapses and poor outcome. Accordingly, finding less toxic and more efficient therapeutic agents to treat pediatric cancers is a necessity to decrease drug-induced deleterious side effects and the associated mortality [3].
In the present study, we assessed the effect of TAT-RasGAP 317-326 in several childhood cancer cell lines. This RasGAP-derived peptide was already known to sensitize adult tumor cells in vitro and in vivo to various anti-cancer therapies [12]. Our results show that TAT-RasGAP 317-326 significantly sensitizes childhood cancer cells in the majority of the tested conditions. However, in some cases, tumor cells did not, or minimally, respond to the peptide. The pediatric tumor-sensitizing effect illustrated in this study was reached using half of the peptide concentration used previously to sensitize adult cancer cells (10 μM TAT-RasGAP 317-326 in this study instead of 20 μM in the paper of Michod et al. [12]). Possibly, the dose of TAT-RasGAP 317-326 could be doubled to increase the sensitizing efficacy of the peptide in the conditions where it was less efficient. Since primary tumor cells represent a more clinically relevant condition, we also tested the peptide in primary neuroblastoma cells derived from bone marrow samples from a single patient (the three NB1-derived cell lines showed in Fig 4). The results show that the RasGAP-derived peptide sensitizes primary tumor cells, suggesting that it could exert its anti-cancer activity on tumors in an in vivo context.
The mechanism of action of TAT-RasGAP 317-326 remains to be precisely determined [18][19][20]. We initially assumed that working with a large panel of cell lines and cytotoxic agents would allow us to draw a pattern that could be used to predict which type of childhood tumors display sensitivity to TAT-RasGAP 317-326 in response to a given drug. Such information would be helpful to decipher the mode of action of TAT-RasGAP 317-326 . For this purpose, we analyzed how the half maximal inhibitory concentration (IC 50 ) for each chemotherapeutic agent was modified by the presence of TAT-RasGAP 317-326 for a given cell line ( Table 2). The RasGAPderived peptide decreased the IC 50 for most genotoxins in most cell lines. However, no specific pattern based on tumor origin or drug type could be deduced from the data presented in Table 2. We also looked whether there were some common specific mutations in well-known oncogenes or tumor suppressor genes (e.g. HRAS, KRAS, NRAS, HDM2, N-MYC, C-MYC, TAL1, FLI-1, MLL, p53 and Rb1) in the analyzed cell lines. Again, no association could be  found between the presence of specific mutations in a cancer cell line and its ability to be sensitized by the RasGAP-derived peptide. An interesting aspect of this research is that the sensitization effect of TAT-RasGAP 317-326 can be differentially modulated by the cell culture conditions. This feature is illustrated by the fact that the NB1-FBS cells, but not the NB1-NBM cells, are sensitive to the effect of the peptide. These two cell lines are derived from the same neuroblastoma tumor at initial diagnosis (see Fig  4) but they were cultured in different media. These two cell lines appear to have the same genomic alterations. Indeed, they possess identical array-CGH profiles with the same typical neuroblastoma-associated segmental chromosomic alterations (unpublished data). So if these two cell lines have identical array-CGH profiles, what could explain their different sensitivity to TAT-RasGAP 317-326 ? One possibility is that the medium in which they are grown modulate their metabolism differentially. This might in turn affect their sensitivity to anti-cancer treatments. There are indeed accumulating evidence that the metabolic state of cancer cells play an important role in the way they respond to anti-cancer drugs [41]. Another hypothesis is that different cell populations are selected upon passages in different medium. Unlike the FBS-containing DMEM medium, the NBM medium blocks cell differentiation. Consequently, the NB1-FBS cells are potentially more differentiated than the NB1-NBM cells. The state of differentiation of the cells could explain their different behavior in response to TAT-RasGAP 317-326 . Alternatively, passage in different medium may induce differential epigenetic modifications or mutations that impact on the sensitivities of the NB1-FBS and NB1-NBM cell lines toward the RasGAP-derived peptide.
We have previously shown that the genotoxin sensitivity of non-tumor cells is not affected by TAT-RasGAP 317-326 [12]. Moreover, doses that induce tumor sensitization in vivo do not exert any detectable toxic side effects [13]. In the present study, we tested the effect of the peptide in isolated lymphocytes from blood of healthy patients. We confirmed the non-toxic effect of the RasGAP-derived peptide in these cells. The peptide was also unable to sensitize the isolated lymphocytes to etoposide, vincristine or doxorubicin. However, TAT-RasGAP 317-326 was found sometimes to sensitize non-malignant lymphocytes to 4-HC-induced death. This finding is clinically relevant because one of the most frequent dose-limiting side effect of chemotherapy is hematological toxicity [42]. Consequently, if TAT-RasGAP 317-326 increases genotoxin-induced side effects toward non-tumor cells, it would lose its attractivity as an anti-cancer candidate to treat childhood tumors. However, cyclophosphamide (and consequently its active form 4-HC) possesses hematologic toxic effects. This toxicity depends of the cellular expression level of aldehyde dehydrogenase (ALDH), an enzyme responsible for cyclophosphamide detoxification. The expression level of ALDH is very low in mature hematopoietic cells such as lymphocytes, while in hematopoietic stem cells, ALDH is expressed at high levels. Thus, the later are relatively resistant to cyclophosphamide, whereas this alkylating agent is toxic toward mature lymphocytes [43]. This feature explains why white blood cell counts drop in patients treated with cyclophosphamide [44]. However this drop is transient and a rapid hematologic recovery invariably occurs after cyclophosphamide therapy. Consequently, the observed sensitizing effect of TAT-RasGAP 317-326 in combination with 4-HC in healthy lymphocytes may not be clinically incompatible, provided that the negative effect of the peptide is limited to mature hematopoietic cells and does not affect hematopoietic stem cells.
Although in most cases TAT-RasGAP 317-326 by itself does not kill tumor cells, we found out that it induced the cell death of the three NB1-derived cell lines tested here. This newly discovered feature has two consequences. First, it will give us the opportunity to study the pro-death activity of the peptide alone. This possibly may facilitate deciphering its mode of action by, for example, using genome-scale knockout screenings [45,46]. Second, it indicates that the Ras-GAP-derived peptide, in its own right, corresponds to a new cytotoxic agent in certain tumors.
In conclusion, our work shows that in most cases TAT-RasGAP 317-326 sensitizes childhood cancer cells to genotoxins. However, it was not possible to predict the efficacy of this sensitization based on the tumor type and the drug used. Additional investigation is therefore required to increase our understanding on how TAT-RasGAP 317-326 works and to determine the indications for which it might be useful in the clinic.