Cross resistance to diverse anticancer nicotinamide phosphoribosyltransferase inhibitors induced by FK866 treatment

Cross-resistance to drugs remains an unsolved problem in cancer chemotherapy. This study elucidates a molecular mechanism of cross-resistance to diverse inhibitors of nicotinamide phosphoribosyltransferase (NAMPT) with anticancer activity. We generated a variant of the human colon cancer cell line HCT116, HCT116RFK866, which exhibited primary resistance to the potent NAMPT inhibitor FK866, and was approximately 1,000-fold less sensitive to the drug than the parental HCT116. HCT116RFK866 was found to be cross-resistant to diverse NAMPT inhibitors, including CHS-828, GNE-617, and STF-118804. Whole-exon sequencing revealed two point mutations (H191R and K342R) in NAMPT in HCT116RFK866, only one of which (K342R) was present in the parental HCT116. Importantly, the protein level, NAMPT enzyme activity, and intracellular NAD+ level were similar between HCT116RFK866 and HCT116. Hence, we investigated NAMPT-binding partners in both cell lines by focused proteomic analyses. The amount of NAMPT precipitated with anti-NAMPT monoclonal antibody was much higher in HCT116RFK866 than in the parental. Furthermore, in HCT116, but not in HCT116RFK866, NAMPT was revealed to interact with POTE ankyrin domain family member E and beta-actin. Thus, these results suggest that NAMPT usually interacts with the two partner proteins, and the H191R mutation may prevent the interactions, resulting in resistance to diverse NAMPT inhibitors.


INTRODUCTION
NAD + is an essential redox cofactor in energy metabolism, including glycolysis, the TCA cycle, and oxidative phosphorylation [1,2]. In addition, it serves as a substrate for multiple enzymes, such as sirtuin and poly (ADP-ribose) polymerase, involved in cellular signaling processes [2], calcium homeostasis [3], gene regulation [4], genome integrity [5], and cell survival or death [6]. The main pathways for NAD + biosynthesis include the de novo pathway, from tryptophan, and two salvage pathways, from nicotinamide (NAM) and nicotinic acid (NA) [2,7]. Many cancer cells have a much higher rate of NAD + turnover, and primarily use the NAM salvage pathway. Accordingly, nicotinamide phosphoribosyltransferase (NAMPT), the rate-limiting enzyme of the salvage pathway, is considered an attractive target for the development of anticancer drugs [1,8].

Research Paper
Oncotarget 16452 www.oncotarget.com Continuous exposure of cancer cells to NAMPT inhibitors can result in acquired resistance to these drugs, often caused by mutations in NAMPT [19][20][21]. For example, the G217R point mutation, first detected in inhibitor-resistant HCT116 human colon cancer cells, results in a 2,500-fold shift in the 50% effective concentration (EC 50 ) of CHS-828 relative to parental HCT116 cells, with no associated change in the level of NAMPT protein [20]. NAMPT mutations that confer resistance to specific NAMPT inhibitors, such as FK866 and CHS-828, include G217R, H191R, D93del, and Q388R [21]. Based on the wild-type NAMPT structure, the side chains of the mutated residues in G217R and H191R protrude into the inhibitor-binding pocket tunnel in NAMPT, whereas the D93del and Q388R mutations are located on the dimer interface [20,21]. Recently, Wang et al. reported six point mutations (D93del, S165F, S165Y, G217R, G217A, G217V) in rhabdosarcoma RD, pancreatic cancer MIAPaCa-2, and non-small cell lung cancer NCI-H460 cells that became resistant to GNE-618 [19]. Some mutated NAMPT alleles are thought to be resistant to NAMPT inhibitors due to lack of occupancy of the tunnel-shaped cavities near the NAM-binding sites [20,21]. The S165F mutant is 1,000-fold more resistant to GNE-618, but only 10-fold and 100-fold more resistant to FK866 and GMX1778, respectively, suggesting that NAMPT mutants are differentially affected by distinct classes of NAMPT inhibitors [19]. Moreover, cell lines harboring S165Y and G217Y are preferentially resistant to GNE-618 in comparison with GMX1778 and FK866 [19]. The precise molecular mechanisms by which cancer cells become cross-resistant to NAMPT inhibitors remain to be elucidated.
To address this issue, we established an FK866resistant HCT116 cell line (HCT116R FK866 ) and analyzed its characteristics. Importantly, HCT116R FK866 cells were found to be cross-resistant to diverse classes of NAMPT inhibitors, including CHS-828, GNE-617, and STF-118804. To elucidate the molecular cause of the drug resistance, we performed whole-exon sequencing to compare the NAMPT gene between HCT116R FK866 and parental HCT116 cells. The results revealed that the point mutation H191R was present in HCT116R FK866 , but not in parental HCT116 cells. Importantly, NAMPT protein level and enzyme activity were similar in HCT116R FK866 and HCT116 cells.
Next, we used a focused proteomic approach, immunoprecipitation with anti-NAMPT monoclonal antibody (mAb) followed by mass spectrometry, to identify NAMPT-binding partners in HCT116R FK866 and HCT116 cells. The level of precipitated NAMPT was much higher in HCT116R FK866 than in HCT116 cells. We also identified two NAMPT-binding proteins, POTEE and beta-actin, in HCT116 but not in HCT116R FK866 cells. These findings suggest that the NAMPT H191R mutation in HCT116R FK866 cells abolishes interactions with the two binding partners, thereby decreasing the protein's affinity for various NAMPT inhibitors.

Establishment of FK866-resistant HCT116 cells
To elucidate the mechanisms underlying crossresistance to diverse NAMPT inhibitors, we generated a variant of the human colon cancer cell line HCT116 that was resistant to FK866, a potent inhibitor of NAMPT in vitro that has efficacy against various human tumors in in vivo xenograft models [11,12]. FK866, also known as APO866 and WK175, is a clinical-stage first-line candidate [11,12]. As in previous experiments aimed at obtaining cells resistant to NAMPT inhibitors, our initial efforts focused on generating a NAMPT inhibitor-resistant HCT116 cell line with one point mutation in the NAMPT protein.
By repeated exposure to stepwise increasing concentrations of FK866 over a period of about 6 weeks ( Figure 1A), we established HCT116R FK866 , an HCT116 variant resistant to the drug. The EC 50 of FK866 in HCT116R FK866 cells was determined by WST-8 assay after continuous exposure for 72 hr. As shown in Figure 1B, the EC 50 value of HCT116R FK866 cells was significantly higher (EC 50 = 3,300 nM) than that of the sensitive parental HCT116 cells (EC 50 = 3.3 nM). The resistance index (RI) was approximately 1,000 (Table 1). In addition, parental HCT116 and HCT116R FK866 cells have almost similar morphological features ( Figure 1C).

Analysis of NAMPT mutations by whole-exon sequencing
To assess the relationship between cross-resistance to diverse NAMPT inhibitors and NAMPT mutations, we www.oncotarget.com identified the point mutations in NAMPT in HCT116R FK866 cells by whole-exon next-generation sequencing. The results showed that two NAMPT mutations, H191R and K342R, were present in HCT116R FK866 cells (Table 1). In contrast, the parental HCT116 cells harbored only the K342R mutation. H191R is one of the six mutations previously reported in NAMPT inhibitor-resistant cells [20,21]. These results suggest that cross-resistance to diverse NAMPT inhibitors is due to this mutation, and further that the H191R mutation may cause similar resistance to CHS-828 and STF-118804, but less resistance to GNE-617, in comparison with FK866. Furthermore, our data suggest that the degree of resistance by H191R mutation to various NAMPT inhibitors depends primarily on the head structures of those compounds, and less on the appended linker and tail groups ( Figure 2).

Biochemical features of the mutated NAMPT in HCT116R FK866
To elucidate the biochemical features of NAMPT in HCT116R FK866 and HCT116 cells, we monitored NAMPT expression levels by western blotting. As shown in Figure  4A and 4B, NAMPT protein levels were almost identical between the cell lines, as were the levels of nicotinic acid phosphoribosyltransferase (NAPRT1) ( Figure 4A, middle) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), used as an internal control ( Figure 4A, lower). We then measured NAMPT enzyme activities and intracellular NAD + levels in HCT116R FK866 and HCT116 cells. As shown in Figure 4C and 4D, the NAMPT enzyme activities in cell lysates from HCT116R FK866 and HCT116 were almost the same. Similarly, the intracellular NAD + levels were approximately equal in both cell lines ( Figure 4E). These findings suggest that the NAMPT mutation in H191R primarily affects the NAMPT dimer conformation without impacting activity.

Molecular mechanisms of cross-resistance to diverse NAMPT inhibitors
To explore the molecular mechanisms by which the H191R mutation confers cross-resistance to various NAMPT inhibitors, we analyzed the NAMPT protein in FK866-resistant HCT116R FK866 and parental HCT116 cells by immunoprecipitation with anti-NAMPT mAb followed by proteomics analysis. The anti-NAMPT mAb precipitates were solubilized and analyzed by SDS-PAGE ( Figure 5A). One protein band in the anti-NAMPT mAb The cells were treated as described in the legend of Figure 3.  immunoprecipitate (band 1), indicated by a red dotted line in Figure 5A, differed in intensity between HCT116R FK866 and HCT116 cells; accordingly, this band was cut out of the gel and analyzed by nano-LC-MS/MS. Two proteins, POTE ankyrin domain family member E (POTEE) and beta-actin, were identified in band 1 ( Table 3). These results suggest that POTEE and beta-actin are specific NAMPT-binding partners in parental HCT116 cells. POTEE is a tumor-associated antigen that is expressed in a wide variety of human cancers, including breast, colon, lung, ovarian, pancreatic, and prostate tumors [22][23][24]. The level of POTEE is very low in normal tissues [23]. Actin, a ubiquitous protein in eukaryotes, is one of the major components of the cytoskeleton. Betaactin, also known as a cytoplasmic actin, is predominantly expressed in nonmuscle cells, where it regulates cell structure and cell motility [25].
To validate the interaction between NAMPT and POTEE and/or beta-actin, we performed individual western blots using specific anti-NAMPT ( Figure 5B, upper), anti-POTEE ( Figure 5B, middle), or anti-beta actin antibody (mAbs) (Figure 5B, lower). Importantly, the amount of immunoprecipitated NAMPT protein was significantly higher in HCT116R FK866 cell extracts than in those from parental HCT116 cells ( Figure 5B, upper). Full-length POTEE (around 121 kDa) was also observed (asterisk). Notably, a truncated POTEE (tPOTEE; indicated by a black line in Figure 5B, middle) of around 42 kDa was much more abundant in HCT116 than in HCT116R FK866 cells. Therefore, it is possible that POTEE undergoes post-transcriptional restricted proteolytic processing and post-translational modifications (e.g., ubiqutination or phosphorylation). In our experiments, the anti-rabbit HRP-conjugated secondary antibodies were not cross-reacted with immunoglobulin heavy and light chain of the immunoprecipitated mouse anti-NAMPT monoclonal antibody. Thus, we assume that the smear-like bands immunoblotted by rabbit anti-POTEE antibodies were not due to artifact such as immunoglobulins use for the immunoprecipitation (Figure 5B, middle). In addition, the level of beta-actin in the immunoprecipitants was significantly higher in HCT116 than in HCT116R FK866 cells ( Figure 5B, lower). However, the beta-actin protein levels in whole-cell extracts did not differ significantly between the two cell lines ( Figure 5B, lower). We consider that the different position of beta-actin band in IP or cell lysate lane was caused by differences the proteins composition in IP or whole cell lysate (Figure 5B lower). The H191R mutation in NAMPT is located in the binding site for NAM and NAMPT inhibitors (e.g., FK866, CHS-828, and TP201565) and in the NAMPT dimer interface [8,20,21]. These observations suggest that the NAMPT H191R variant in HCT116R FK866 cells fails to achieve complete dimer formation or interact with its binding partner(s), tPOTEE, and/or beta-actin, resulting in lower affinity for diverse NAMPT inhibitors ( Figure 6). Interestingly, these data shows that the differences of NAMPT protein complex, interactions with or without binding partners, tPOTEE and beta-actin, between parental HCT116 and FK866-resistant HCT116R FK866 cells do not affect the NAMPT activities.

Cell culture
The human colon cancer cell line HCT116 was obtained from the American Type Culture Collection. Parental and resistant HCT116 cell lines were cultured in D-MEM containing 10% heat inactivated fetal bovine serum, 100 units/mL penicillin, and 100 µg/mL streptomycin in a 37°C incubator under an atmosphere containing 5% CO 2 at 100% relative humidity.

Cell viability assay
Cell viability was determined using the WST-8 cell proliferation assay (Takara Bio, Otsu, Japan). Briefly, cells were passed into 96-well plates (1,000 cells per well) in triplicate, and then treated with various concentrations of drugs or ultra-pure water (as a negative control). Following incubation for 72 hr, 20 μL of WST-8 reagent was added to each well, and the plate was placed in a 5% CO 2 incubator at 37°C for an additional 1 hr. Optical density was measured at 490 nm on a Tecan microplate reader. The EC 50 value was defined as the concentration of drug producing 50% inhibition of cell proliferation. The RI was defined as the ratio of EC 50 values between the resistant and parental cell lines. Experiments were repeated at least three times.

Colony formation assay
Colony formation assay was performed as described previously [26,27]. HCT116 and HCT116R FK866 cells were dissociated with Accutase, suspended in medium, inoculated into six-well plates (200 cells per well) in triplicate, and then incubated overnight. The cells were treated with various concentrations of drugs or with solvent (DMSO or ultrapure water) as a negative control. After incubation for 10 days, cells were fixed with 4% formaldehyde solution and stained with 0.1% (w/v) crystal violet, and then the number of colonies in each well was counted.

Analysis of NAMPT activity
NAMPT activity was assayed using whole-cell lysates prepared by incubating cells in lysis buffer (2% CHAPS in PBS containing protease-inhibitor cocktail [Roche]) on ice for 30 min. Cell lysates were sonicated on a Branson Sonifier-250, then centrifuged at 15,000 ×g for 15 min at 4°C. The supernatants containing the solubilized protein fractions were subjected to enzymatic assays. NAMPT activity was measured using a coupled-enzyme reaction system (CycLex NAMPT colorimetric assay kit) in 96-well plate format using the one-step method.

Measurement of intracellular NAD + /NADH levels
Parental HCT116 and HCT116R FK866 cells were dissociated with Accutase and suspended in PBS, and the cell suspensions were subjected to NAD + /NADH assays using the NAD/NADH-Glo Assay kit (Promega). The resulting luminescent signals were measured on a Tecan microplate reader.

Whole-exon sequencing analysis
DNA extraction was performed as described previously [28]. Genomic DNA was extracted from cells (5 × 10 6 cells per sample) using the DNeasy Tissue Kit (QIAGEN). Whole-exon sequencing of parental HCT116 and HCT116R FK866 cells was performed by the

Protein identification
Protein identification was subjected to nano-LC/ MS/MS analysis according to a standard protocol [29,30].

Statistical analysis
The data are presented as the means ± standard deviation. Significance of differences among groups was evaluated using Student's t-test; P < 0.05 was considered to indicate a statistically significant difference.

CONCLUSIONS
In this study, we generated a cancer cell line that is cross-resistant to diverse NAMPT inhibitors, identifying the mutation responsible for resistance, and elucidated the molecular mechanisms of action. Our work revealed that mutation of H191R in the NAMPT molecule is responsible for cross-resistance to diverse NAMPT inhibitors. Interestingly, in immunoprecipitation analysis with anti-NAMPT mAb, the amount of immunoprecipitated NAMPT protein was significantly higher in HCT116R FK866 cell extracts. We also identified two novel NAMPTbinding proteins, POTEE and beta-actin, in parental HCT116 but not in HCT116R FK866 cells. Our data suggest that the NAMPT H191R mutant in HCT116R FK866 fails to achieve complete dimer formation due to diminished interactions with its binding partners, POTEE and/or beta-actin, and consequently has lower binding affinity Figure 6: Predicted structure of NAMPT in HCT116R FK866 and parental HCT116 cells. In HCT116 cells, NAMPT may be present in a homodimer in complex with tPOTEE and beta-actin, and have high affinity to NAMPT inhibitors. The H191R mutant in HCT116R FK866 cannot interact with the binding proteins, and consequently has lower affinity for diverse NAMPT inhibitors.