Sensitivity and Acquired Resistance of BRCA 1 ; p 53-Deficient Mouse Mammary Tumors to the Topoisomerase I Inhibitor Topotecan

Authors' A Cancer Ins


Introduction
Topotecan is used to treat ovarian, cervical, and small cell lung cancer patients.Topotecan inhibits topoisomerase I (Top1)-mediated religation of ssDNA breaks (SSB).When drug-stabilized covalent Top1-DNA complexes cause stalling of the replication machinery, SSBs are converted into doublestranded DNA breaks (DSB).Such replication-mediated DSBs seem to be the primary cytotoxic mechanism of Top1 inhibitors in dividing cells (1).Cell lines defective in proper DSB repair pathways show an increased sensitivity to Top1 poisons (reviewed in ref. 1).An example is a defect in BRCA1 (2), which is required for error-free repair of DSB by homologous recombination (HR).HR is frequently impaired in basal-like breast cancer, and BRCA1 dysfunction is one of the causes of this defect (3)(4)(5).About 15% of clinical patients are diagnosed with basal-like breast cancer, which is usually negative for ERBB2 and hormone receptors ("triple negative"; refs.6,7).Patients with this disease face a poor prognosis, as there is no targeted therapy available (6,7).Topotecan is not applied in the clinic to breast cancer patients; hence, we have investigated in a mouse model whether the use of this Top1 inhibitor is an alternative therapeutic approach for patients with basal-like breast cancer.We used the K14cre;Brca1 F/F ; p53 F/F genetically engineered mouse model (GEMM), in which Brca1 −/− ;p53 −/− mammary adenocarcinomas arise, that recapitulate key features of human BRCA1-associated breast cancer (invasive ductal carcinoma not otherwise specified) and can be transplanted orthotopically into syngeneic mice without loss of their genomic profile, morphology, or sensitivity to drug (8,9).In this model inhibition of SSB repair by the poly(ADP-ribose) polymerase (PARP) inhibitor olaparib (AZD2281; KU-0059436) induces DNA damage resulting in synthetic lethality of Brca1 −/− ;p53 −/− tumor cells and a strong increase in overall survival (10).Remarkable responses to olaparib treatment with only few adverse effects were also observed in a phase I clinical trial with patients that carry BRCA1 or BRCA2 mutations (11).Because in our model olaparib treatment alone did not result in tumor eradication (10), we have tested whether increasing the induction of SSBs by a Top1 poison would enhance antitumoral efficacy of this PARP inhibitor.
Here we show that topotecan alone or in combination with olaparib may be another therapeutic option for breast cancer patients who carry cancers with HR defects.However, we find that Brca1 −/− ;p53 −/− tumors relapse and eventually acquire resistance.Of the various mechanisms reported in cultured cells to cause resistance to Top1 inhibitors (1,12,13), we identified only two in our model: an increased expression of the drug efflux transporter ABCG2 and reduced protein levels of the drug target Top1.We confirmed the in vivo relevance of ABCG2 for topotecan resistance through orthotopic transplantation of Abcg2 −/− ;Brca1 −/− ;p53 −/− tumors into syngeneic wild-type mice.The generation of spontaneous mammary tumors deficient for a drug efflux transporter that might protect tumor-initiating cells from chemotherapy (14,15) is novel and cannot be achieved in humans.Our data strongly suggest that ABCG2 is dispensable for the survival of tumor-initiating cells in this model.
Drugs and treatment of tumor-bearing animals.Topotecan was dissolved in glucose (5%, w/v) to yield a solution of 0.4 mg/mL (of active compound) and administered at 10 μL/g of body weight by i.p. injection.When tumors reached a size of ∼200 mm 3 , 4 mg/kg topotecan were administered i.p. daily for 4 (Fig. 1A) or 5 (Fig. 1B) consecutive days or on days 0 to 4 and 14 to 18 (Figs.1C, 3A and 4A).Tumors that responded [volume, <50% 14 d after the first topotecan injection of the first treatment course (Fig. 1A, B) or second course (Figs.1C, 3A and 4A)] were left untreated until they relapsed to 100%.If tumors did not respond (volume remaining, ≥50%), treatments were continued after a recovery time of 9 d after the last treatment.Olaparib and tariquidar were applied as reported (10,17).In topotecan-olaparib combination-treated animals (Fig. 6B and C), the topotecan dose was lowered 8-fold to 0.5 mg/kg and administered on days 0 to 4, 14 to 18, and 28 to 32, whereas 50 mg/kg olaparib were injected i.p. daily for 42 d and, when combined with topotecan, administered 15 min after topotecan injection.Tariquidar (2 mg/kg; Fig. 6C) was administered i.p. every second day during this 42-d period.
Genome-wide expression profiling.RNA extraction, amplification, microarray hybridization, data processing, and statistical analyses were performed as described (9).Twocolor duplicate hybridizations on mouse microarrays (NKI Central Microarray Facility) were used, containing 38,784 70-mer probes representing 23,527 genes and 35,172 gene transcripts (MEEBO, Illumina BV).The microarray data reported in this article have been deposited in the Array Express database of the European Bioinformatics Institute (Cambridge, United Kingdom; accession no.E-NCMF-28).

Results
Brca1 −/− ,p53 −/− mammary tumors respond to the maximum tolerated dose of topotecan but relapse and eventually acquire resistance.We established the maximum tolerated dose (MTD) of topotecan in 12-to 14-week-old female K14cre,Brca1 F/F ,p53 F/F mice to maximize antitumoral drug efficacy and thereby mimicking a clinically relevant dose (Supplementary Fig. S1A).Topotecan (4 mg/kg) i.p. on 4 consecutive days was tolerated and resulted in 10% to 20% weight loss.When two animals carrying "spontaneous" mammary tumors (∼200 mm 3 in volume) were treated with this dose, tumors shrank (Fig. 1A, T1 and T2).Despite this good response, tumors were not eradicated and relapsed.Once the tumors grew back to a volume of ∼200 mm 3 (100%) or showed progressive growth after a recovery period of 14 days after the first topotecan injection, treatments were resumed.Response of Brca1;p53-deficient tumors to the MTD of topotecan.A, two individual K14cre,Brca1 F/F ,p53 F/F animals carrying a spontaneous mammary tumor of ∼200 mm 3 (T1 and T2) were treated with 4 mg/kg topotecan i.p. on days 0 to 3. For comparison, the growth of untreated spontaneous tumors in three individual K14cre,Brca1 F/F ,p53 F/F animals (C1-C3) is also shown.B, animals with four individual orthotopically transplanted Brca1;p53-deficient mammary tumors (volume, ∼200 mm 3 ) were left untreated (pink line) or treated as in A, but on days 0 to 4. C, animals bearing 14 individual orthotopically transplanted Brca1;p53-deficient mammary tumors (volume, ∼200 mm 3 ) were left untreated (pink line) or treated as in A, but on days 0 to 4 and 14 to 18 (dose-intensified treatment).When tumors in A to C relapsed or showed progressive growth (tumor size, ≥50%) after a recovery time of 14 d after the day 0 injection, treatment was resumed as indicated by the arrows.
Eventually, both tumors acquired resistance to the MTD of topotecan.In an attempt to eradicate the tumors by intensifying the topotecan therapy, we used wild-type animals with orthotopically transplanted Brca1;p53-deficient tumors, because wild-type animals could take an additional dose of topotecan (4 mg/kg drug i.p. on 5 consecutive days; Supplementary Fig. S1B).The orthotopically transplanted tumors responded to topotecan like the primary tumors, but also the intensified topotecan therapy was unable to eradicate the tumors (Fig. 1B, T3-T6).We therefore added a second MTD schedule after a 14-day recovery period after the first topotecan injection and tested this treatment in 14 individual Brca1 −/− ;p53 −/− tumors (4 mg/kg drug i.p. on days 0-4 and 14-18; Fig. 1C, T7-T20).Although some tumors became nonpalpable after this dose-dense treatment, eventually all tumors relapsed and acquired topotecan resistance.Interestingly, the time until resistance developed varied substantially between individual tumors, showing the response heterogeneity of these tumors.
In drug-sensitive tumors we identified a strong increase in nuclear DNA damage foci 24 hours after a 5-day topotecan schedule using γ-H2AX as marker (Supplementary Fig. S2).γ-H2AX foci are associated with DSBs (18); hence, our finding supports the notion that topotecan-induced SSBs are converted into DSBs in the tumors.Moreover, we observed a significant decrease in the proliferation marker Ki-67 and an increase in senescense-associated β-galactosidase activity, as well as increased nuclear p19 ARF staining (Supplementary Fig. S3) after topotecan treatment.On the basis of published findings in cell lines (19), we expected to find many apoptotic cells.However, only a modest increase in apoptosis-related cell death [cleaved caspase-3 and terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL) positive] was detected in shrinking tumors (Supplementary Fig. S3A).
Abcg2 is frequently upregulated in topotecan-resistant tumors.We looked for alterations in gene expression using a two-class paired significance of microarray analysis (SAM) comparing 39K oligo expression arrays of resistant to matched untreated control tumor samples.We found elevated expression of 150 genes at a δ of 1.058 (false discovery rate of 2.21%; Fig. 2A and Supplementary Table S2).Of these 150 genes only Abcg2 has previously been linked to topotecan resistance.Abcg2 encodes the ATP-binding cassette (ABC) drug efflux transporter ABCG2/BCRP, which is known to cause topotecan resistance in cultured cells (20,21).Quantification of Abcg2 transcript levels confirmed that the level of Abcg2 was at least 2-fold elevated in 9 of 20 tumors (Fig. 2B, T1, T2, T4, T6, T9, T11, T13, T15, and T16).Moreover, ABCG2-positive tumor cells could be detected in situ (Fig. 2C) with the ABCG2-specific antibody Bxp53 (16), and immunoreactivity correlated with Abcg2 transcript levels (Supplementary Table S3).The increased Abcg2 gene expression was not due to DNA amplification: we quantified the DNA content in whole tumor lysates by MLPA and found no copy number alterations of Abcg2 (Supplementary Fig. S4).
Doxorubicin-resistant and Abcb1a/b-overexpressing tumors respond to topotecan after orthotopic transplantation.To investigate whether the mouse P-glycoproteins ABCB1A and ABCB1B can cause topotecan resistance in our tumor model, three individual doxorubicin-resistant and docetaxel cross-resistant Brca1 −/− ,p53 −/− tumors were orthotopically transplanted into syngeneic wild-type hosts and treated with the MTD of several drugs (Fig. 3A).The tumors maintained their elevated Abcb1a and Abcb1b RNA levels (Fig. 3B) and accordingly their resistant phenotype after transplantation but were completely sensitive to topotecan.This shows that upregulation of the mouse P-glycoproteins by itself does not lead to topotecan resistance in this tumor model.
Introduction of Abcg2 null alleles into the mouse model.To study ABCG2-independent topotecan resistance mechanisms, Abcg2 null alleles were introduced into the K14cre, Brca1 F/F ,p53 F/F mouse model (Supplementary Fig. S5).Abcg2 −/− ,K14cre,Brca1 F/F ,p53 F/F animals with spontaneous mammary tumors cannot be treated with the topotecan MTD established in Abcg2 +/+ mice, because the ablation of ABCG2 alters topotecan pharmacokinetics (28,29).This problem was solved by orthotopic transplantation of tumors into syngeneic wild-type animals (Fig. 4A).In comparison with Abcg2-proficient tumors, we found an increased topotecan sensitivity in Abcg2-deficient tumors resulting in an augmented overall survival (P < 0.001; Fig. 4B).This confirms that ABCG2 is contributing to topotecan resistance in vivo.Nevertheless, ablation of Abcg2 did not result in tumor eradication.Eventually, most of the Abcg2 −/− ;Brca1 −/− ; p53 −/− tumors developed topotecan resistance after a median latency of 137 days (n = 11) after start of treatment.
Topotecan-olaparib combination therapy does not eradicate tumors either.Brca1 −/− ,p53 −/− tumors are sensitive to the PARP inhibitor olaparib, and its combination with cisplatin or carboplatin increased the overall survival of tumor-bearing mice (10).Topotecan might even be more effective in combination with olaparib, because topotecan increases SSBs, which require PARP-mediated repair.In the absence of functional PARP, topotecan should drastically increase SSBs and thus increase synthetic lethality with BRCA1 deficiency.When we tested 4 mg/kg topotecan i.p. on days 0 to 4 in combination with 50 mg/kg olaparib i.p. daily, the mice had to be killed due to diarrhea with accompanying weight loss that could not be compensated by additional fluid supplementation (Fig. 6A).This is not surprising, as PARP deficiency is known to sensitize mammalian cells to Top1 poisons (33).The combination with olaparib was tolerated only when topotecan was lowered to 0.5 mg/kg i.p. on 5 consecutive days (Fig. 6A).We then treated Abcg2-deficient Brca1 −/− ,p53 −/− tumors to avoid ABCG2-mediated topotecan resistance and to maximize the chance of tumor eradication.Nevertheless, the topotecan-olaparib combination did not increase the relapse-free survival of four individual tumors compared with olaparib as single agent (Fig. 6B, TB1, TB5, TB7, and TB11).Even the addition of the P-glycoprotein inhibitor tariquidar to the regimen, to prevent Abcb1a/b-mediated olaparib resistance (10), did not result in the eradication of the tumors (Fig. 6C).

Discussion
We show here that spontaneous and transplanted Brca1 −/− ; p53 −/− mammary tumors are sensitive to topotecan, albeit with considerable variability in drug response.The median overall survival of mice with transplanted tumors under the dose-densest topotecan therapy (T7-T20) was 53.5 days (±39 SD) compared with 8 days (±1 SD) of their matched untreated controls (Fig. 1C).Such intertumoral differences, which did not correlate with a specific histomorphology, were also observed with other drugs in this model (9,10) and seem to mimic the response heterogeneity seen in patients.We know that tumors differ in additional mutations acquired after the initial Brca1 and p53 deletion (8), and these must be responsible for response variability.
Despite their initial sensitivity, tumors eventually acquire resistance to the MTD of topotecan.Research with cell lines has identified several mechanisms of camptothecin/topotecan resistance.These include overexpression of the drug efflux transporters ABCG2, ABCC4, ABCB1, ABCC2, reduced expression of Top1, Top1 mutations, Top1 mislocalization, elevated DNA repair, and resistance to apoptosis (reviewed in refs.1,12,13).Which of these mechanisms contribute to topotecan resistance in real tumors has remained unclear, however (34,35).We show here that increased levels of ABCG2 are a major mechanism of topotecan resistance in vivo, as in cultured cells (21,36,37).The importance of this mechanism is illustrated by the Abcg2 −/− tumors in which the development of topotecan resistance was substantially delayed (Fig. 4B).Abcg2 −/− ;Brca1 −/− ;p53 −/− tumors can still de-velop full topotecan resistance, however, and a substantial decrease in the level of Top1 can explain resistance in several of these tumors.Remarkably, this profound decrease in Top1 protein was not accompanied by a corresponding decrease in Top1 mRNA in most tumors, in sharp contrast with the observations on camptothecin-resistant tumor cell lines (12,32,38).If downregulation of Top1 would also be posttranscriptional in human tumors, its detection by gene expression profiling would be impossible.Top1 degradation is mediated by ubiquitin-or SUMO1-associated posttranscriptional modification (39)(40)(41)(42), and which mechanism applies in our mouse tumors is under investigation.
It is notable that some mechanisms of topotecan resistance identified in cell lines, such as upregulation of ABCC4 or mutations in Top1, have not yet turned up in our tumor model.ABCG2-mediated drug efflux or downregulation of Top1 can explain topotecan resistance in 14 of 20 individual Brca1 −/− ;p53 −/− tumors, and the study of topotecan resistance in the remaining tumors is ongoing.P-glycoprotein is not such an effective transporter of topotecan as ABCG2 (23,43,44), but how effective it is cannot be deduced from the literature.Relative resistance levels up to 20-fold have been reported (25), although most recent papers find much lower P-glycoprotein-mediated resistance or transport (45)(46)(47).Our results show that even a 53-fold upregulation of P-glycoprotein, which makes the tumor completely resistant to doxorubicin, does not result in clear-cut resistance to topotecan.In real tumors, P-glycoprotein is therefore unable by itself to cause topotecan resistance.Nevertheless, we found in many tumors a significant (2-to 13fold) upregulation of the Abcb1 genes (Fig. 2B).Although this could be incidental, we think that P-glycoprotein, in conjunction with other weak resistance mechanisms, could allow the tumor cell sufficient time to develop more robust resistance.Effective cooperation between ABCG2 and P-glycoprotein in the defense against topotecan penetration into the brain has been documented by de Vries and colleagues (48).Whereas only minor increases (up to 2-fold) of brain topotecan were found in Abcg2 −/− or Abcb1a/1b −/− mice, the compound knockout mice showed 12-fold increased brain penetration relative to wild-type mice.This shows that P-glycoprotein can significantly contribute to the prevention of cellular topotecan accumulation, if it gets help.
It is remarkable that deletion of Abcg2 did not result in tumor eradication by topotecan.ABCG2 has been proposed as a critical factor in the chemotherapy resistance of tumorinitiating cells of various tumor types (14,15,22), but our results argue against this proposal.We cannot exclude that topotecan does not reach remnant tumor cells but we consider this unlikely, as there are no obvious histologic sanctuaries in the remnants.Another possibility is that the tumor-initiating cells are protected by another drug efflux transporter, but there is no evidence for this either.Our .Topotecan response of Abcg2;Brca1;p53-deficient tumors and differential survival of Abcg2 −/− versus Abcg2 +/+ tumor-bearing animals under topotecan therapy.A, response of Abcg2;Brca1;p53-deficient tumors to the MTD of topotecan.Animals with orthotopically transplanted Abcg2;Brca1;p53-deficient mammary tumors (volume, ∼200 mm 3 ) were left untreated (pink line) or treated with 4 mg/kg topotecan i.p. on days 0 to 4 and 14 to 18.When tumors relapsed or showed progressive growth (tumor size, ≥50%) after a recovery period of 14 d after the day 0 injection, treatment was resumed as indicated by the arrows.B, overall survival (%) over time (d) of wild-type animals carrying orthotopically transplanted Abcg2-proficient (T7-T20) or -deficient (TB1-TB11) Brca1 −/− , p53 −/− mammary tumors.Topotecan was administered as in A and Fig. 1C.All animals carrying Abcg2-proficient tumors and 10 of 11 animals carrying Abcg2-deficient tumors had to be killed, because the resistant tumor reached a volume of ∼1,500 mm 3 .One animal (TB11) had to be sacrificed due to a rectum prolapse on day 198, before having acquired full topotecan resistance.The P value was calculated using the Wilcoxon rank sum test.current hypothesis is that remnant tumor cells arrest in the cell cycle and thereby escape anticancer drug-mediated cell death (49).This hypothesis is supported by the observation that cycling cells (Ki-67 positive) are drastically reduced after treatment.Because we found many senescent-like cells in the topotecan-sensitive tumors after treatment, it is possible that therapy-induced senescence allows escape from topotecan-induced cell death.
How cells in topotecan-sensitive tumors are killed is unclear.We found only a slight increase in apoptosis-related cleaved caspase-3 and TUNEL immunoreactivity, as was also observed in reponse to olaparib (10).Necrosis or mitotic catastrophe ( 50) may be more important mechanisms of cell death in this tumor, but we cannot exclude that cleaved caspase-3mediated cell cycle-dependent apoptosis is only detectable for a short period of time and therefore underestimated in tumor sections.In vitro studies with Brca1 −/− ;p53 −/− cells are in progress to settle this issue.In any case, our data show that in vivo imaging of apoptosis is not a suitable readout for therapy response in this tumor model.
To explore new therapeutic alternatives for breast cancers associated with BRCA1 defects, we tested the combination of topotecan with the PARP inhibitor olaparib.This combination did not result in tumor eradication, even if we prevented efflux of topotecan by using Abcg2 −/− tumors and P-glycoprotein-mediated olaparib extrusion by inhibiting this transporter with tariquidar.We also observed this lack of complete cell kill when olaparib was combined with platinum drugs (10).It is clinically relevant that the topotecanolaparib combination proved so toxic in mice that we had to reduce the topotecan concentration 8-fold.We have not seen this increased toxicity in the combination of olaparib Figure 6.Response of transplanted Abcg2 −/− ,Brca1 −/− ,p53 −/− tumors to topotecan-olaparib combination therapy.A, relative animal weights (%) in response to topotecan-olaparib combination treatment.Female wild-type (FVB/N × 129/Ola)F1 animals were treated with 4 mg/kg topotecan i.p. on days 0, 1, 2, 3, and 4 in combination with 50 mg/kg olaparib i.p. daily (blue triangles).On day 5 the animals had to be killed due to severe weight loss caused by diarrhea.Weight monitoring in response to the 8-fold lower dose of 0.5 mg/kg is shown as red squares.The average of five animals is presented.Error bars, SD.B, response of four Abcg2 −/− ,Brca1 −/− ,p53 −/− tumors (TB1, TB5, TB7, and TB11) to topotecan (orange) and olaparib (brown) monotherapy or topotecan-olaparib (green) combination therapy (untreated controls in blue).Animals received 0.5 mg/kg topotecan i.p. on days 0 to 4, 14 to 18, and 28 to 32; 50 mg/kg olaparib i.p. daily for 42 d or both antitumor agents in these dosages combined.After a 42-d dosing schedule animals were left untreated and screened for tumor relapse.C, response of the same Abcg2 −/− ,Brca1 −/− ,p53 −/− tumors as in B to tariquidar-topotecan and tariquidar-olaparib treatment and tariquidar-topotecan-olaparib combination therapy.In addition to the same olaparib and topotecan dosages as in B, animals received 2 mg/kg tariquidar i.p. every second day during 42 d of drug administration to prevent ABCB1A/B-mediated olaparib resistance.After a 42-d dosing schedule animals were left untreated and screened for tumor relapse.
with cisplatin or carboplatin (10).We do not think that topotecan toxicity is increased because olaparib inhibits ABCG2, because methotrexate transport by ABCG2-overexpressing cells was not altered by olaparib (data not shown).It seems more likely that the level of endogenous DNA damage requiring PARP-mediated repair is already close to the maximum that normal cells can deal with.A further increase in SSBs induced by topotecan would overtax the DNA repair capacity remaining in the absence of functional PARP.
Our study shows that GEMMs of human cancer are not only useful for studying the response of real tumors to drugs but also for the identification of resistance mechanisms actually occurring in vivo.Eventually, novel strategies to prevent or reverse topotecan resistance can be tested in this model as well.

Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.

Figure 1 .
Figure1.Response of Brca1;p53-deficient tumors to the MTD of topotecan.A, two individual K14cre,Brca1 F/F ,p53 F/F animals carrying a spontaneous mammary tumor of ∼200 mm 3 (T1 and T2) were treated with 4 mg/kg topotecan i.p. on days 0 to 3. For comparison, the growth of untreated spontaneous tumors in three individual K14cre,Brca1 F/F ,p53 F/F animals (C1-C3) is also shown.B, animals with four individual orthotopically transplanted Brca1;p53-deficient mammary tumors (volume, ∼200 mm 3 ) were left untreated (pink line) or treated as in A, but on days 0 to 4. C, animals bearing 14 individual orthotopically transplanted Brca1;p53-deficient mammary tumors (volume, ∼200 mm 3 ) were left untreated (pink line) or treated as in A, but on days 0 to 4 and 14 to 18 (dose-intensified treatment).When tumors in A to C relapsed or showed progressive growth (tumor size, ≥50%) after a recovery time of 14 d after the day 0 injection, treatment was resumed as indicated by the arrows.

Figure 4
Figure 4. Topotecan response of Abcg2;Brca1;p53-deficient tumors and differential survival of Abcg2 −/− versus Abcg2 +/+ tumor-bearing animals under topotecan therapy.A, response of Abcg2;Brca1;p53-deficient tumors to the MTD of topotecan.Animals with orthotopically transplanted Abcg2;Brca1;p53-deficient mammary tumors (volume, ∼200 mm 3 ) were left untreated (pink line) or treated with 4 mg/kg topotecan i.p. on days 0 to 4 and 14 to 18.When tumors relapsed or showed progressive growth (tumor size, ≥50%) after a recovery period of 14 d after the day 0 injection, treatment was resumed as indicated by the arrows.B, overall survival (%) over time (d) of wild-type animals carrying orthotopically transplanted Abcg2-proficient (T7-T20) or -deficient (TB1-TB11) Brca1 −/− , p53 −/− mammary tumors.Topotecan was administered as in A and Fig.1C.All animals carrying Abcg2-proficient tumors and 10 of 11 animals carrying Abcg2-deficient tumors had to be killed, because the resistant tumor reached a volume of ∼1,500 mm 3 .One animal (TB11) had to be sacrificed due to a rectum prolapse on day 198, before having acquired full topotecan resistance.The P value was calculated using the Wilcoxon rank sum test.