Looking at the
TOP1 gene copy number in the SN-38 resistant cells using FISH, we identified changes in two of the three resistant cell lines (HCT116-SN38 and HT29-SN38). The first was a
TOP1 gene copy gain in a subpopulation of cells (about half of the cells) independently of centromere-20, a marker for the chromosome bearing
TOP1 (HCT116-SN38), while the other was a loss of one copy of chromosome 20 (including one copy of
TOP1) (HT29-SN38). Two cell line studies by McLeod and Keith [
21] and Romer et al. [
17] demonstrated a positive correlation between
TOP1 gene copy number and sensitivity to SN-38 and between gene copy number and protein expression, respectively. The positive association between
TOP1 amplification and expression of Top1 mRNA and protein was confirmed by another study [
22], and Top1 mRNA and protein expression have been shown to correlate well in the NCI-60 cancer cell line panel [
14].
Using primers covering the full coding region of
TOP1 we did a mutational analysis of the gene in the SN-38 resistant and parental cells. Two of three resistant cell lines did not harbor any mutations, while the third (HCT116-SN38) harbored two mutations. Both identified mutations were non-synonymous and heterozygous; one located at amino acid residue position 364 (c.1336C > T, R364K, arginine to lysine change) in the Top1 core domain and the other at position 717 (c.2395G > A, G717R, glycine to arginine change) in the C-terminal domain [
27,
46]. Both were shown by multiple sequence alignment to be located at highly evolutionarily conserved positions in the Top1 protein. Using
TOP1 wild-type and mutant specific primers, we showed by PCR that the mutated cells (HCT116-SN38) expressed mutant
TOP1 mRNA at a markedly higher level than wild-type
TOP1 mRNA, even though the sequencing analysis suggested heterozygosity of the mutations.
TOP1 containing a single mutation (R364K or G717R) appeared to be expressed at a higher level than double-mutant
TOP1, suggesting that the mutations are present on separate alleles or in separate subpopulations in the HCT116-SN38 cells, each of them conferring resistance. This is in accordance with the FISH analysis, which detected two subpopulations in the HCT116-SN38 cell line. In line with these findings, we demonstrated that Top1 activity in nuclear extract from the HCT116-SN38 cell line was largely unaffected in the presence of high doses of SN-38, which completely eliminated the Top1 activity in parental cells, while the Top1 activity in drug absence was the same in HCT116-SN38 and parental cells. Furthermore, we measured the degree of formation of Top1-DNA cleavage complexes by alkaline elution following SN-38 treatment of cells, and showed that far fewer complexes were formed in HCT116-SN38 cells compared to parental cells. These findings suggest that SN-38 binding to Top1-DNA is hindered in the mutation-harboring cells. In previous studies, identified
TOP1 mutations have clustered in regions close to the structural site where Top1 binds DNA and camptothecin, i.e. the regions 361–364 (DNA minor groove), 503–533 (minor groove) and 717-729 (major groove) [
11,
47]. Some mutations have been shown to hinder binding of drug, while others destabilize the drug-Top1-DNA cleavage complex or enhance Top1 DNA religation. Other mutations have also been identified in relationship to the linker region of Top1 [
38,
48]. In a study by Li et al. [
49] amino acids in the 361–364 region was demonstrated to be involved both in enzyme catalysis and camptothecin resistance. More specifically, experimental substitution of amino acids in the 361–364 range, i.e. R362L (arginine to leucine) and R364G (arginine to glycine) was shown to affect the catalytic activity of Top1, however R364G only slightly reduced the activity compared to wild-type enzyme [
49]. Furthermore, Li et al. [
49] showed that R364G Top1 was able to bind DNA with the same affinity as wild-type enzyme, however camptothecin was largely unable to bind R364G Top1-DNA cleavage complexes and cause DNA breaks. In addition, a R364H (arginine to histidine change) mutation was previously described in two camptothecin resistant prostate cancer cell lines [
50]. The R364H mutation did not affect the catalytic activity, but rendered the cells resistant to camptothecin [
50]. As lysine is a large positively charged amino acid, similar to histidine, it is very likely that R364K Top1 is functionally similar to R364H Top1. Several other mutations associated with camptothecin resistance have previously been reported in the amino acid region 361–365 of Top1 [
51‐
55]. The C-terminal domain of Top1 is known to be involved in both enzyme catalytic activity and drug binding [
27], and several camptothecin resistance-associated mutations have been reported in the 717–737 region [
56‐
62], including a pair of mutations in a tumor sample from a cisplatin/irinotecan treated lung cancer patient [
29]. A camptothecin-associated mutation at position 717 (G717V, glycine to valine change) has previously been reported together with the mutation T729I [
56]. The authors showed that the mutation-harboring cells displayed similar Top1 catalytic activity as wild-type cells, and that each mutation on its own rendered yeast cells resistant to camptothecin [
56]. Losasso et al. [
62] investigated various amino acid substitutions in the 729 position, and suggested that this position is part of a hydrophobic pocket important for drug sensitivity [
62]. Recently, mutations have also been reported in the linker region (amino acid residues 636–712), between the core and C-terminal domains of Top1 [
38,
48]. Losasso et al. [
48] investigated a resistance-associated mutation at position 653 and suggested that altered Top1 linker flexibility is a likely mechanism of resistance [
48]. Gongora et al. [
38] have subsequently identified other mutations in the linker region, which could confer resistance by this mechanism [
38]. One of the linker region mutations found by Gongora et al. (Glu710Gly) was further analyzed in S. cerevisiae and the data indicated that a fully functional linker region of Top1 is important to confer camtotethecin sensitivity [
63]. We did not detect any
TOP1 mutations in the two other SN-38 resistant cell lines (HT29-SN38 and LoVo-SN38). However, even in these cells Top1 activity was unaffected by the presence of large concentrations of SN-38, and thus other mechanisms must be responsible for this finding.
Lastly, we assessed the three SN-38 resistant cell lines for cross-resistance to two non-camptothecin Top1-targeting drugs in clinical trials (indenoisoquinolines: NSC 725776/LMP776/indimitecan and NSC 743400/LMP400/indotecan) and two clinical Top2-targeting drugs (epirubicin and etoposide). All three SN-38 resistant cell lines displayed no or very little cross-resistance to etoposide, a specific Top2 inhibitor [
12,
13,
46], while showing more but still only partly cross-resistance to epirubicin, a DNA intercalating Top2 inhibitor [
64]. Previous studies have demonstrated that reduced activity of Top1 can be compensated by increased activity of Top2 and thus increased sensitivity to etoposide [
65,
66], and furthermore that camptothecin-resistant cells retain sensitivity to Top2-targeting drugs [
50]. In the three SN-38 resistant cell lines, the indenoisoquinolines displayed an interesting pattern of resistance, from full cross-resistance to no cross-resistance. The two indenoisoquinolines are currently in clinical development [
67]. In the present study, the cell line harboring the R364K-G717R mutations (HCT116-SN38) showed very strong cross-resistance to LMP400 and LMP776. The two SN-38 resistant cell lines, which did not carry any mutations in
TOP1 (HT29-SN38 and LoVo-SN38), were partly cross-resistant to LMP776, while showing either no or only small cross-resistance to LMP400. Studies in our laboratory [
36] showed that HT29-SN38 and LoVo-SN38 both strongly upregulated expression (mRNA; 25- and 60-fold, respectively) of the well-known drug-efflux pump ABCG2 (BCRP) [
68,
69], while HCT116-SN38 did not (data not shown). This suggests that LMP400 can remain active in cancers resistant to SN-38, which display upregulation of this multidrug resistance protein, if
TOP1 is wild-type. This is in line with a previous study by Antony et al. [
35], where LMP776, but not LMP400, was shown to be a weak substrate of the ABCG2 pump. LMP400 was thus effective in cells overexpressing ABCG2, which displayed a 46-fold resistance to SN-38 [
35]. Our experiment measuring Top1-DNA cleavage complexes following treatment with LMP400 supported the cross-resistance findings. These results highlight how the underlying molecular mechanism of camptothecin resistance in cancer cells determines their resistance-profile to new classes of drug such as the indenoisoquinolines.