1. Introduction
Microtubules (MTs), dynamic polymeric filaments composed of α-tubulin and β-tubulin heterodimers, are key components of the cytoskeleton of eukaryotic cells. Their crucial roles in cell division and physiology mainly rely on their ability to rapidly polymerize or depolymerize. Targeted perturbation of this finely tuned process constitutes a major therapeutic strategy. Drugs interfering with MTs are major constituents of chemotherapies for the treatment of carcinomas. A number of compounds bind to the tubulin-MT system. They can be roughly classified into MT-stabilizers such as taxanes or epothilones, and MT-destabilizers such as vinca alkaloids, combretastatin, and colchicine [
1]. It has been demonstrated that binding of vinca alkaloids or colchicine prevents the curved-to-straight conformational change of tubulin at the tip of the growing MT, necessary for proper incorporation of new tubulin dimers into the MT lattice (see reviews [
1,
2]).
Paclitaxel (PTX) binds to the taxane-site of β-tubulin and stabilizes the MT lattice by strengthening lateral and/or longitudinal tubulin contacts within the MT [
1]. At stoichiometric concentrations, it promotes MT assembly. At low and clinically relevant concentrations, PTX primarily suppresses MT dynamics without significantly affecting the MT-polymer mass [
3,
4]. PTX is one of the most successful chemotherapeutic drugs in history. It is currently used to treat patients with a variety of cancers including lung, breast, and ovarian cancers [
5].
Several mechanisms have been proposed to explain the anti-tumor activity of PTX. It can induce a mitosis dependent cell death, either by producing a mitotic arrest [
6], when applied at high concentrations, or by promoting chromosome mis-segregation at low concentrations [
7]. Alternatively, PTX can act on interphase cells and drive autonomous cell death by perturbation of intracellular trafficking [
8]. It has also been recently proposed that post-mitotic formation of micronuclei induced by PTX can promote inflammation and subsequent tumor regression via vascular disruption and immune activation [
9].
While PTX is a successful anti-cancer drug, its low solubility, its toxicity, and the fact that cells become resistant to this drug, impose serious limits to its use. Cell resistance to PTX is due to the high expression of P-glycoprotein or multidrug resistance-associated proteins, as well as to the overexpression of some β-tubulin isoforms or mutations in β-tubulin that affect the MT polymer mass and/or drug binding [
10]. Another major drawback of PTX in clinical applications is the development of peripheral neuropathies, primarily involving the sensory nervous system. Although the molecular bases of these neuropathies are not completely understood, an inhibition of MT-based axonal transport appears to be a possible mechanism [
11]. It has been recently shown that anterograde kinesin based-axonal transport is specifically affected by PTX, whereas MT destabilizing drugs that bind preferentially to the ends of MTs have much less effect on axonal transport [
12].
An alternative therapeutic solution would be the use of pharmaceutics which, when co-administrated with PTX, could potentiate its effect without significantly increasing its toxicity. Such agents could allow the use of lower doses of PTX in cancer therapy, may limit the occurrence of resistances and reduce MT-independent adverse effects.
To identify such agents, we screened a collection of 8000 original compounds using a cytotoxicity assay and selected a derivative of the carbazole series (Carba1) able to sensitize cells to a low, non-toxic dose of PTX. We demonstrated that Carba1 exerts synergistic cytotoxic effects with PTX. In cells, Carba1 has no major effect on the total MT mass in interphase cells and shows moderate cytotoxicity. We found that Carba1 targets the colchicine-binding site of tubulin, inhibits in vitro tubulin polymerization and promotes catastrophes, similar to other MT-destabilizing agents. A combination of Carba1 and PTX causes synergistic perturbation of MT growth in vitro. Carba1-induced modulation of MT dynamics increases the binding of fluorescent taxane, Fchitax-3 to MTs, similar to what we have described previously for vinblastine [
13], providing a biochemical explanation of the observed synergy between Carba1 and PTX.
Carba1 has no major anti-tumor effect when administrated alone in animals and no detectable toxicity. The administration of a combination of Carba1 and a low, ineffective, dose of PTX showed, however, a significant effect on tumor growth, indicating that Carba1 and PTX act synergistically in vivo. Our results pave the way for new therapeutic strategies, based on the combination of low doses of MT targeting agents with opposite mechanisms of action. These combinations may have reduced toxicity compared to high therapeutic PTX doses.
3. Discussion
Our initial aim was to discover an agent that would allow lowering the dose of PTX while obtaining the same anti-tumor efficacy as the currently used therapeutic dose of PTX. We thus screened a chemical library to detect compounds able to sensitize cells to a low, non-toxic dose of PTX. The test we used was a cytotoxicity test, therefore probing all vital cell functions. Whereas such a whole cell-based assay screens molecules having multiple potential targets and allows the biology to dictate the best targets [
23], it may not be insignificant to have selected Carba1, an agent that targets tubulin and impairs MT dynamics. Indeed, this indicates that the most sensible target, in this specific context, is tubulin.
Recently, a series of carbazole-based MT targeting agents has been reported [
24]. These acyl-substituted derivatives, conceived as analogs of nocodazole, represent other examples of carbazole scaffolds able to interact with the colchicine site of tubulin. Unlike Carba1, they exert potent killing activities in human glioblastoma cells. A possible explanation for the difference in cytotoxicity observed between Carba1 and the compounds described by Diaz et al. [
24] may reside in a lower affinity of Carba1 for the colchicine binding site of the tubulin dimer. Indeed, we found that the affinity of Carba1 for the colchicine site is not very high, in the micromolar range. This difference in affinity could be due to the presence of the pyrol ring on Carba1, which could generate steric hindrance and decrease the affinity of the compound for the colchicine site. In addition, the work of Diaz et al. emphasizes the importance of substitutions at the level of the nitrogen atom of the carbazole moiety. Such substitutions are absent from Carba1 and thus may impact binding of the compound to the colchicine site.
The Carba1 scaffold is a versatile one and we are currently synthesizing modified analogs for medicinal chemistry optimization.
The PTX binding site at the interior of the MT has been characterized at the atomic level: PTX binds to a pocket in β-tubulin that faces the MT lumen and is near the lateral interface between protofilaments (for review see [
1]). The binding of PTX results in the expansion of the taxane binding pocket [
25] of the tubulin dimer. Moreover, PTX binding inhibits, in the protofilament, the compaction at the longitudinal interdimer interface, induced by GTP hydrolysis [
26]. This allosteric mechanism would strengthen the longitudinal tubulin contacts leading to a stabilization of the MTs [
1]. In this context, it is conceptually counterintuitive that an agent that depolymerizes MTs acts in synergy with PTX, an agent that stabilizes them.
A possibility is that the binding of Carba1 to the tubulin dimer modifies its affinity for PTX. However, although it has been shown that the covalent occupancy of the taxane site can affect the structure of the colchicine site [
27], the reverse has not yet been described. Moreover, in cells, due to the low affinity of Carba1 for tubulin and the nanomolar concentration of PTX that was used, it can be assumed that the probability that a single tubulin dimer has both a molecule of Carba1 and another of PTX bound is very low. Thus, an allosteric effect at the level of the tubulin dimer, due to such a simultaneous binding, cannot be responsible for synergistic cytotoxicity.
Another possibility is that the binding of Carba1 can induce conformational changes of the growing MT ends that can facilitate the subsequent binding of PTX to the MT lattice. Recently, using TIRF analysis, it has been shown that non-saturating doses of vinblastine induce a switch to catastrophe and convert the MT plus end to a state that allows more efficient taxane accumulation [
13]. Indeed, we conducted the same type of experiment, replacing vinblastine with Carba1 and observed an increase in the rate of catastrophes associated with more frequent formation of accumulation “hotspots” of fluorescent taxane. Moreover, we observed a synergistic effect of Carba1 and PTX on MT growth inhibition in vitro. Although the underlying structural mechanism is yet unknown, it is highly probable that Carba1 acts similarly to vinblastine to facilitate PTX accumulation. Our experimental in vitro data thus strongly suggest that Carba1 acts by modifying the growing tip of microtubules, favoring the accumulation of PTX.
It is known that PTX accumulates intracellularly [
4], reaching intra-tumor concentrations that are higher in the tumors than in the plasma [
7]. It is thus remarkable that the synergistic effect is observed not only at the MT level, but also at the cellular level, as well as when both drugs are applied systemically in animals to exert their anti-tumor action. Although the most probable hypothesis is that the same molecular mechanism is at work in these different contexts, it cannot be excluded that Carba1 has other target(s). For instance, Carba1 could also inhibit a drug export pump and further facilitate PTX accumulation.
We observed a different anti-tumor response in the two models used. In one case (allogeneic orthotopic 4T1 cell transplant model) the difference in tumor size observed between PTX alone and Carba1/PTX combination therapy was not significant. In the other case (xenograft of HeLa cells), a synergistic effect is clearly demonstrated.
We assume that these different tumor responses reflect the sensitivity of these cells to PTX in vitro (PTX GI50 is 1.5 nM in HeLa cells and 90 nM in 4T1 cells). More trivially, differences in response between the two models may also result from the protocols used, particularly with respect to the injection route—i.p. (4T1) versus i.v. (HeLa)—and the frequency of injections—every day (4T1) versus every other day (HeLa). Nor can we rule out the possibility that the combination is more effective on certain tumor types. A thorough study of the effect of the combination on different cell lines, such as the one performed for Carba 1 alone in the NCI60 screen, will probably answer this question.
These last years, attention has been directed to the combined use of therapeutic agents to target critical cellular pathways involved in carcinogenesis. Accordingly, several small molecules have been reported to synergize with PTX to kill cancer cells. For instance, combined administration of the src inhibitor dasatinib and PTX has a synergistic antiproliferative activity on different cancer cell lines [
28,
29]. Similarly, aberrations in the PI3K/AKT/mTOR pathway are commonly described in aggressive cancers [
30] and inhibitors of this pathway can improve outcomes in some patients when combined with paclitaxel [
30,
31,
32]. We cannot completely exclude that Carba1 exerts its synergistic effect by additionally targeting a pathway involved in cell survival. However, it is less probable that it targets a survival pathway involving some kinases as we have tested Carba1 potential inhibitory activity on a panel of 64 kinases, including src and mTOR, and found that Carba1 did not exhibit any selective activity on the assayed kinases.
As the combined administration of Carba1 and a low dose of PTX can have an anti-tumor effect, one could imagine that this combination should reduce the unwanted side effects observed with high doses of PTX. This has to be tested. For instance, the effect of the combination should be compared to the PTX effect on the kinesin-based anterograde transport, since perturbation upon PTX treatment is thought to be part of the mechanism involved in peripheral neuropathy [
12]. Given the mode of action we have described, with Carba1 facilitating the accumulation of PTX in MTs, we can bet that the combination should diminish MT-independent adverse events. Interestingly, it has been shown that a compound sharing the same carbazole scaffold protects against PTX-induced peripheral neuropathy [
33].
Anti-cancer strategies based on the concomitant administration of taxanes and depolymerizing agents such as vinorelbine have been reported [
34,
35,
36]. However, these approaches used high doses of each of these drugs. Our results suggest that good therapeutic efficacy could be achieved with the combined administration of each of these agents at low doses, which could improve patient comfort.
In addition, reducing the doses of taxane administered to patients could delay the onset of acquired resistances. This work thus paves the way to new therapeutic perspectives that are easy to implement.
4. Materials and Methods
4.1. Chemical Reagents and Cells
The chemical library that was used for the initial screening is part of the collection assembled by the CERMN (Centre d’Études et de Recherche sur le Médicament de Normandie, University of Caen, Caen, France). This collection is composed of original compounds (>19,000), synthesized within the frame of the numerous drug design programs the unit has developed for 40 years. These compounds are mainly small molecules, often including heteroelements and heterocycles, they basically obey to the Lipinski’s rules of five and are considered as druggable. The 8000 compounds screened were selected from this chemical library using a clustering method in order to adapt the size of the panel to the capacity of the screening, while respecting the structural diversity of the collection.
Carba1 was re-synthesized at the CERMN and supplied as a powder. It was dissolved in anhydrous dimethyl sulfoxide (DMSO, Sigma-Aldrich, #D4540, St Quentin Fallavier, France) and stored at −20 °C as 10 mM stock solution. Paclitaxel (PTX) was purchased from Sigma (#T7402) and was dissolved in DMSO and stored at −20 °C as 1 mM stock solution.
The human HeLa and RPE-1 cell lines and the murine 4T1 cell line were obtained from the American Type Culture Collection (ATCC, Gaithersburg, MD, USA), routinely tested and authenticated by the ATCC. HeLa Kyoto cells expressing EGFP-alpha-tubulin and H2B-mcherry were from Cell Lines Service, #300670. HeLa cells and 4T1 cells were grown in RPMI 1640 medium (Gibco, Invitrogen, Carlsbad, CA, USA) and RPE-1 cells were grown in DMEM (Gibco, Invitrogen), supplemented with 1% penicillin/streptomycin and 10% Fetal Bovine Serum, and maintained in a humid incubator at 37 °C in 5% CO2.
4.2. Analysis of Cell Viability Using MTT (Screening of the Chemical Library)
The assay was performed in 96-well microplates. Cells were seeded at a density of 2500 cells per well and allowed to adhere for 24 h before being treated for 48 h with either DMSO (0.1% final concentration) or compounds at 5 μM, with or without 1 nM PTX. Viability was evaluated with a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) colorimetric assay (Sigma, #M5655).
4.3. Analysis of Cell Viability Using Prestoblue Assay
Cell viability was analyzed using the colorimetric Prestoblue assay (Invitrogen, #A13262). Cells were seeded in 96-well microplates (Greiner, #655077, Courtaboeuf, France) at a density of 2500 cells per well and allowed to adhere for 24 h before being treated for 72 h with either DMSO (0.1 % final concentration) or drugs at indicated concentrations. After a 72-h treatment, 11 µL Prestoblue was added to each well and cells were incubated for another 45 min. The absorbance of each well was measured using a FLUOstar Optima microplate reader (Excitation, 544 nm; Emission, 580 nm, BMG Labtech, Champigny/Marne, France).
4.4. Apoptosis Assay
The apoptosis assay was performed with FITC Annexin V Apoptosis Detection Kit I (BD Biosciences, #556547, San Jose, CA, USA) using flow cytometry and analyzed by FCS express software.
4.5. Drug Combination Analysis
The Chou–Talalay analysis on the basis of dose–response curves was used to evaluate the synergism, additivity, and antagonism of the combination drug treatment. Combination index (CI) values were calculated using the CompuSyn software (
www.combosyn.com) which uses the equation:
where CA,× and CB,× are the concentrations of drug A and drug B in the combination to produce a certain effect ×. SA,× and SB,× are the concentrations of drug A and drug B used as a single agent to produce that same effect. CompuSyn also generates a plot of CI values at different fraction affected (Fa) levels referred to as Fa-CI plot, which are widely used to interpret drug combination effects [
37]. A CI value of <0.1 indicates very strong synergism, 0.1–0.3 strong synergism, 0.3–0.7 synergism, 0.7–0.9 moderate to slight synergism, 1 additivity, 1.1–1.45 slight to moderate antagonism, 1.45–3.3 antagonism, and >3.3 strong to very strong antagonism.
4.6. Cell Cycle Analysis
Cells were harvested and washed by centrifugation in PBS. Then, 105 cells were fixed in 1 mL of 70% ethanol at 4 °C overnight. Following two washes with PBS the cells were incubated with 50 µg/mL propidium iodide and 0.2 mg/mL RNase A (Sigma, #10109142001)/PBS for 30 min at 37 °C before analysis. The percentage of cells in the specific cell-cycle phases (G0, G1, S, G2, and M) was determined using an Accuri C6 flow cytometer (Becton Dickinson, San Jose, CA, USA).
4.7. Analysis of Carba1 Effect on Kinases
The analysis of Carba1 effect was performed on a panel of 64 recombinant protein kinases. The assays were performed at 10 μM ATP in the presence of 10 μM Carba1 using the Upstate Kinase profiler panel service (Millipore, Molsheim, France). Inhibition, expressed as the percent of activity determined in the absence of inhibitor, was calculated from the residual activity measured in the presence of 10 μM Carba1.
4.8. Immunofluorescence Microscopy and Live Cell Imaging
HeLa cells at a density of 20,000 cells were grown for 48 h on glass coverslips placed in a 24-well microplate. When cells reached 70% confluence the medium was replaced with a fresh one supplemented with Carba1. After a 5-h exposure to Carba1, cells were fixed and permeabilized with −20 °C absolute methanol for 6 min. After washing and saturation with 3% BSA (Bovine Serum Albumin; Sigma, #A7906)/PBS (Phosphate Buffered Saline; Dutscher, #L0615-500, Brumath, France), cells were incubated for 45 min at room temperature (RT) with anti-alpha-tubulin antibody (clone α3A1, 1:4000), produced by L. Lafanechère [
38]. Cells were washed twice again and subsequently incubated with Alexa 488 conjugated anti-mouse antibody (1:1000, Jackson immunoresearch, #115-545-4637, Cambridgeshire, UK) for 30 min at RT. DNA was stained with 20 µM Hoechst 33342 (Sigma, #23491-52-3) and coverslips were mounted on glass slides with Mowiol 4-88 (Calbiochem- Sigma-Aldrich, #475904, St Quentin Fallavier, France). Images were captured with a Zeiss AxioimagerM2 microscope equipped with the acquisition software AxioVision (Marly le Roi, France) and analyzed using the Fiji software (imagej.net). For live-cell imaging, HeLa Kyoto cells expressing EGFP-alpha-tubulin and H2B-mcherry were seeded on 2-well glass-slides (Ibidi, #80297, Gräfelfing, Germany) at a density of 7000 cells per well and allowed to grow for 24 h prior to imaging. After treatment, the slide was placed on a 37 °C heated stage, at 5% CO
2, and images were acquired every 2.5 min by a spinning disk confocal laser microscope (Andromeda iMIC, TILL I.D. GmbH, Martinsried, Germany) equipped with a Plan-Apochromat 20×/0.75 WD610 objective and an EMCCD camera (iXon 897, Oxford Instrument, Belfast, UK). For each time point, a stack of seven planes (thickness: 1 µm) was recorded. Acquisition (LA), off-line analysis (OA) and Fiji software programs were used.
4.9. Transfection of GFP-EB3
To label MT plus ends, GFP-EB3 plasmids were used because EB3 has a strong binding affinity to MT plus ends. Cell transfection was performed using electroporation (AMAXA®, Köln, Germany). In total, 2 µg of purified plasmid DNA were used for each transfection reaction.
4.10. Fluorescence Time-Lapse Videomicroscopy of MT Plus Ends
Live imaging of MT plus ends was performed as described in Honoré et al. [
39], on transiently GFP-EB3 transfected-HeLa cells by using an inverted fluorescence microscope (ZEISS Axiovert 200M with a 63× objective, Zeiss, Marly le Roi, France); Time-lapse acquisition was performed with a COOLSNAP HQ (Roper Scientific, Ottobrunn, Germany), driven by Metamorph software (Universal Imaging Corp., Molecular Devices, San Jose, CA, USA). Image acquisition was performed at a temperature of 37 ± 1 °C/5% CO
2.
Data are from three independent experiments. For each experiment, six MTs/cell in six cells per condition were analyzed.
4.11. Dynamic Instability Parameter Analysis
The dynamic instability parameter analysis was performed by tracking MT plus ends over time using the imageJ software (imagej.net). The methods of calculation were as described in Honoré et al. [
39].
4.12. Tubulin Polymerization Assay
Tubulin was prepared from bovine brain as previously described [
40]. Tubulin polymerization assays were carried out at 37 °C in BRB80 buffer (80 mM Pipes, 0.5 mM MgCl
2, 2 mM EGTA, 0.1 mM EDTA, pH 6.8 with KOH) by mixing 7 µM of pure tubulin, 1 mM GTP, 5 mM MgCl
2, and indicated concentrations of drugs (0.2% DMSO, final concentration) in a final volume of 100 µL. The time course of the self-assembly activity of tubulin was monitored as turbidity at 350 nm, 37 °C, during 30 min, using a spectrophotometer (ThermoScientific, Evolution 201, Waltham, MA, USA).
4.13. [3H]-Colchicine Tubulin-Binding Assay
The tubulin was prepared from bovine brain as previously described [
40]. Pure tubulin (3 µM final concentration) in cold BRB80 buffer was mixed at 4 °C with a mix of [
3H]-colchicine (82.6 Ci/mmol, Perkin-Elmer, #NET189250UC, 50 nM final concentration, Courtaboeuf, France) and the competitor Carba1 (100 µM final concentration) in a final volume of 200 µL. Following a 30-min incubation at 30 °C, the samples were deposited onto 50 µL of presedimented DEAE Sephadex A25 in BRB80 buffer. All subsequent steps were carried out at 4 °C. Samples were incubated for 10 min with continuous shaking to ensure quantitative binding of tubulin to the gel. Following centrifugation (2400
g, 4 min), supernatants were discarded and the pellets containing the bound molecule-tubulin complexes were washed four times with 1 mL of BRB80 buffer. Pellets were incubated for 10 min with 500 µL of ethanol to solubilize the tubulin-bound tritiated colchicine and 400 µL aliquots of the ethanol solutions were transferred to 5 mL of Ultima Gold scintillant (Perkin-Elmer) for determination of radioactivity.
4.14. Determination of the Binding Constant of Carba1 on Tubulin Using a Competition Assay
Calf brain tubulin was purified as described [
41]. 2-Methoxy-5-(2,3,4-trimethoxyphenyl)-2,4,6-cycloheptatrien-1-one (MTC) [
42] was a kind gift of Prof. T. J. Fitzgerald (School of Pharmacy, Florida A & M University, Tallahassee, FL, USA). The compounds were diluted in 99.8% DMSO-d6 (Merck, Darmstadt, Germany) to a final concentration of 10 mM and stored at −80 °C.
Competition of the compound with MTC was tested by the change in the intensity of fluorescence of MTC upon binding to tubulin. The fluorescence emission spectra (excitation at 350 nm) of 10 μM tubulin and 10 μM MTC in 10 mM sodium phosphate, 0.1 mM GTP, pH 7.0, were measured in the presence of different concentrations (0–20 μM) of the desired ligand with 5 nm excitation and emission slits using a Jobin-Ybon SPEX Fluoromax-2 (HORIBA, Ltd., Kyoto, Japan). The decrease in the intensity of the fluorescence in the presence of the competitor ligand indicated competition for the same binding site. The data were analyzed and the binding constants determined using Equigra V5.0 (available upon request to J.F. Diaz) as described in Díaz and Buey [
43].
4.15. In Vitro MT Dynamics and Analysis of MT Dynamics Parameters
In vitro assay for MT growth dynamics and analysis of MT dynamic parameters in the presence of tubulin (Cytoskeleton Inc., Denver, CO, USA) and EB3 with Carba1 and Fchitax-3 was performed as described previously [
13]. For statistical analysis, graphs were plotted in GraphPad Prism 7 (San Diego, CA, USA) and statistical analysis was done using non-parametric Mann–Whitney U-test.
4.16. Tumor Xenografts in Mice
All animal studies were performed in accordance with the institutional guidelines of the European Community (EU Directive 2010/63/EU) for the use of experimental animals and were authorized by the French Ministry of Higher Education and Research under the reference: apafis#8854-2017031314338357 v1.
In a first exploratory experiment, the effects of PTX and/or Carba1 were evaluated on the allogeneic 4T1-rvLuc2 orthotopic mammary carcinoma model. Five-week-old female NMRI nude mice (Janvier Labs, Le Genest-Saint Isle, France) were anesthetized (4% isoflurane/air for anesthesia induction and 1.5% thereafter) and 20,000 4T1 cells in 1× PBS were implanted into the mammary fat pad. Mice were then randomized in five groups of eight mice each and drugs were administered intraperitoneally every day. A first group received the vehicle (14% DMSO, 14% Tween 80, and 72% PBS). Two groups received PTX at different doses (2 and 8 mg/kg) while one other group received Carba1 at 30 mg/kg. A last group received a combination of Carba1 (30 mg/kg) and PTX (2 mg/kg). Tumor size was measured every 2 days, using a caliper, and the tumor volume was calculated as follows: length × (width)2 × 0.4. At the end of experiment, 300 µL of D-luciferin (Promega, 10 mg/mL in PBS, Madison, WI, USA) was injected intraperitoneally and the tumor cell viability was evaluated using in vivo bioluminescence imaging (IVIS Kinetic, Perkin-Elmer, Courtaboeuf, France). Living Image software (Perkin-Elmer) was used to analyze the results. Kruskal–Wallis test was used to compare the effects of the treatments on tumor size.
In a second series of experiment, the effects of PTX and/or Carba1 were evaluated on a tumor model based on HeLa cells transplantation. First, the effects of PTX or Carba1 when administrated alone were evaluated. To that aim, 5-week-old female NMRI nude mice (Janvier Labs, Le Genest-Saint Isle, France) were anesthetized (4% isoflurane/air for anesthesia induction and 1.5% thereafter) and were injected subcutaneously in the flank with 107 exponentially dividing HeLa cells in 1 × PBS. Tumor size was measured three times a week using a caliper. When tumors reached a volume of about 250 mm3 i.e., 9 days after cell injection, mice were randomized in seven groups of six mice each and drugs were injected intravenously every 2 days. A first group received the vehicle (14% DMSO, 14% Tween 80, and 72% PBS). Three groups received PTX at different doses (2, 4, and 8 mg/kg) while three other groups received Carba1 at different doses (15, 30, and 60 mg/kg). Statistical comparison between mice groups were determined using two-way ANOVA.
Then, the effect of a combination of PTX- Carba1 was evaluated, and compared to the effect of the compounds alone. To that aim, 5-week-old female NMRI nude mice were injected subcutaneously with 107 exponentially dividing HeLa cells into the right flank. When tumors reached a volume of about 200 mm3 i.e., 9 days after cell injection, mice were randomized in four groups of eight mice each and drugs were injected intravenously every 2 days. The first group received PTX at 3 mg/kg, the second group Carba1 at 60 mg/kg, the third group received a combination of Carba1 (60 mg/kg) and PTX (3 mg/kg), and the fourth group received the vehicle (14% DMSO, 14% Tween 80, and 72% PBS). For statistical analysis, we verified the normality of the data using a Shapiro test and the homogeneity of the variances using a Bartlett test. We then used a Student’s t-test to compare the different groups to the control (i.e., vehicle). We found that the p values of the comparisons of the group treated with PTX (3 mg/kg) and of the group treated with Carba1 (60 mg/kg) with the vehicle were, respectively, 0.5355 and 0.5139. These p values are greater than 0.05, indicating that these groups are not different from the vehicle group at the risk beta calculated greater than 90% (power of the test lower than 10%). On the contrary, the comparison of the group treated with the combination to the control group gave a p value of 0.00015, indicating that the combination has an effect on the tumor size, at the risk alpha of 5%.