Reversal of inflammation-associated dihydrodiol dehydrogenases (AKR1C1 and AKR1C2) overexpression and drug resistance in nonsmall cell lung cancer cells by wogonin and chrysin

Authors

  • Hao-Wei Wang,

    1. Department of Surgery, Faculty of Medicine, School of Medicine, National Yang-Ming University, Taipei, Taiwan
    2. Division of Thoracic Surgery, Department of Surgery, Taipei Veterans General Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Chin-Ping Lin,

    1. Department of Medical Research, Mackay Memorial Hospital, Taipei, Taiwan
    Search for more papers by this author
  • Jen-Hwey Chiu,

    1. Institute of Traditional Medicine, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author
  • Kuan-Chih Chow,

    1. Institute of Biomedical Sciences, National Chung-Hsing University, Taichung, Taiwan
    Search for more papers by this author
  • Kuang-Tai Kuo,

    1. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author
  • Chen-Sung Lin,

    1. Institute of Clinical Medicine, National Yang-Ming University, Taipei, Taiwan
    Search for more papers by this author
  • Liang-Shun Wang

    Corresponding author
    1. Division of Thoracic Surgery, Department of Surgery, Taipei Veterans General Hospital, Taipei, Taiwan
    2. Institute of Emergency and Critical Care Medicine, National Yang-Ming University, Taipei, Taiwan
    3. Department of Surgery, Far Eastern Memorial Hospital, Taipei, Taiwan
    • Department of Surgery, Far Eastern Memorial Hospital, 21, Sec.2, Nan-Ya S. Rd., Pan-Chiao, Taipei 220, Taiwan
    Search for more papers by this author
    • Fax: +886-2-89664355.


Abstract

Dihydrodiol dehydrogenase (DDH) is a member of the aldo-keto reductases superfamily (AKR1C1–AKR1C4), which plays central roles in the metabolism of steroid hormone, prostaglandin and xenobiotics. We have previously detected overexpression of DDH as an indicator of poor prognosis and chemoresistance in human non-small lung cancer (NSCLC). We also found DDH expression to be closely related to chronic inflammatory conditions. The aim of this study was to investigate the links between inflammation, DDH expression and drug resistance in NSCLC cells. We showed that pro-inflammatory mediators including interleukin-6 (IL-6) could induce AKR1C1/1C2 expression in NSCLC cells and increase cellular resistance to cisplatin and adriamycin. This effect was nullified by Safingol, a protein kinase C inhibitor. Moreover, the expression of AKR1C1/1C2 was inversely correlated to NBS1 and apoptosis-inducing factor (AIF). We also showed that IL-6-induced AKR1C1/1C2 expression and drug resistance were inhibited by wogonin and chrysin, which are major flavonoids in Scutellaria baicalensis, a widely used traditional Chinese and Japanese medicine. In conclusion, this study demonstrated novel links of pro-inflammatory signals, AKR1C1/1C2 expression and drug resistance in NSCLC. The protein kinase C pathway may play an important role in this process. Overexpression of AKR1C1/1C2 may serve as a marker of chemoresistance. Further studies are warranted to evaluate wogonin and chrysin as a potential adjuvant therapy for drug-resistant NSCLC, especially for those with AKR1C1/1C2 overexpression. © 2007 Wiley-Liss, Inc.

The link between inflammation and cancer has been proposed for more than 1 century. Early in 1863, Virchow hypothesized that chronic inflammation could enhance cell proliferation and promote cancer development, based upon the observation that cancer originated from sites of chronic inflammation.1 It was not until the past decades that Virchow's theory became more widely accepted because several lines of evidence have indicated the associations between cancer and inflammatory mediators.1, 2, 3 Nonetheless, many of the underlying mechanisms remain unresolved.

Dihydrodiol dehydrogenase (DDH) is a member of the aldo-keto reductase (AKR) superfamily. It plays central roles in the metabolism of steroid hormone, prostaglandin and xenobiotics.4 DDH is also involved in the activation of carcinogenic polycyclic aromatic hydrocarbons (PAHs) by catalyzing NADP+-dependent oxidation of the proximate carcinogenic metabolites PAH trans-dihydrodiols.5 At least 4 isoforms of DDH have been identified in humans, including AKR1C1-AKR1C4.6, 7 The DDH isoforms also exhibit PGF synthase activity by catalyzing the formation of 9α,11β-PGF2 from PGD2 and PGF from PGH2.8, 9 We have previously detected overexpression of DDH in human primary nonsmall cell lung cancer (NSCLC) and esophageal squamous cell carcinoma (ESCC), and found it indicative of poor prognoses.10, 11 High expression of DDH was also observed in tissue from severe esophagitis.11 Besides, the expression of DDH was also found to correlate with inflammatory mediators including cyclooxygenase-2 (COX-2) and interleukin-6 (IL-6) in hepatocellular carcinoma.12 These findings indicated that DDH may play a role in the link between chronic inflammation and cancer. Our previous work has demonstrated upregulation of IL-6 and COX-2 in NSCLC and ESCC,13, 14, 15 and we also showed that IL-6 could enhance COX-2 expression in ESCC cells.16 Therefore, it is valid to investigate the roles of pro-inflammatory mediators, such as IL-6, in the regulation of DDH expression in cancer cells.

Lung cancer is one of the leading causes of cancer-related death worldwide. In Taiwan, it accounts for about 20 percent of total cancer-related deaths.17 In the United States, approximately one third of all cancer-related deaths are due to lung cancer, and the toll exceeds the total death numbers of breast, prostate and colon cancer combined.18 Despite improvements in diagnosis and treatment, the overall 5 year survival rate remains below 15%.19 Resistance to chemotherapy and radiotherapy has been considered as the major obstacle to successful treatment. Recently, we demonstrated that DDH overexpression was closely associated with chemoresistance in various cancers.20, 21, 22 We also found that transfection of the full-length cDNAs of AKR1C1 or AKR1C2 into NSCLC cells could increase cellular resistance to cisplatin, adriamycin and radiation, while no significant effect was observed in AKR1C3-transfected cells.23 Therefore, in this study we aimed to investigate the in vitro effects of pro-inflammatory mediators on the expression of AKR1C1/1C2 in NSCLC cells and the influence on drug resistance.

In addition, we also examined the therapeutic potential of Scutellaria baicalensis, a widely used traditional Chinese and Japanese herb medicine. Scutellaria baicalensis is also known as Huang Qin to the Chinese. It has been used to treat a variety of allergic and inflammatory diseases.24, 25 Many flavonoids with bioactivities have been identified in Scutellaria baicalensis including baicalin, baicalein, wogonin, wogonoside, apigenin, chrysin and scutellarein.26, 27, 28 In recent research, this herbal medicine and its constituents also showed to have antitumor effects.29 In this study, the efficacy of the pure compounds from Scutellaria baicalensis in modulating inflammation-associated AKR1C1/1C2 expression and chemoresistance was investigated.

Material and methods

Cell lines, cell cultures, drug treatment and reagents

The human lung cancer cell lines H1437, H1648, H2009, H2087, H2126, H23 and H838 were purchased from American Type Culture Collection (Manassas, VA). The culture media were obtained from GibcoBRL (Carlsbad, CA). Cells were grown as monolayer in RPMI1640 containing 5% FCS, 1% nonessential amino acid, 3 mM glutamine, 100 IU/ml penicillin and 100 μg/ml streptomycin. All cultures were incubated at 37°C with 5% CO2. The human recombinant IL-6 was purchased from GibcoBRL. The purified components of Scutellaria baicalensis (wogonin, bacalin and bacalein) were purchased from Nacalai Tesque (Kyoto, Japan), and chrysin from Sigma (St. Louis, MO). The pure compounds were dissolved in DMSO to the desired concentrations. Other materials and reagents not specified were obtained from Sigma or Merck (Darmstadt, Germany).

RNA extraction and signal amplification using RT-PCR

RNA extraction and RT-PCR.

The total RNA was extracted from 1 × 107 cancer cells using RNeasy Mini Kit (Qiagen, Santa Clarita, CA). Following spectrophotometric determination of RNA concentration, 5 μg of total RNA was reverse-transcribed using RevertAid™ First Strand cDNA Synthesis Kit (MBI Fermentas, St. Leon-Rot, Germany) with oligo(dT)18 primers according to the manufacturer's protocol. An aliquot of cDNA was then subjected to 35 cycles of PCR, with each cycle consisting of denaturing at 94°C for 30 sec, hybridizing at 52°C for 30 sec, and elongating at 72°C for 30 sec. The primer sequences and the length of the amplified gene fragments are listed in Table I. For AKR1C1 and AKR1C2, we utilized two kinds of PCR primer, the AKR1C1/1C2 common primer, which simultaneously amplifies AKR1C1 and AKR1C2, and the isoform-specific primers as previously described by Ji et al.30 The PCR product was resolved by electrophoresis on 2% agarose gel in Tris-acetate-EDTA buffer and visualized with ethidium bromide staining.

Table I. The Primer Sets Used in RT-PCR
GeneS/AS1Primer Sequence, 5′ → 3′AccessionPositionProduct size (bp)
  • 1

    S, sense; AS, anti-sense.

  • 2

    The AKR1C1/1C2 primer was designed to detect simultaneously AKR1C1 and AKR1C2.

  • Abbreviations: AKR, aldo-keto reductase, sEpH; soluble epoxide hydrolase; mEpH, microsomal epoxide hydrolase; CBR, carbonyl reductase; GST, glutathione-S-transferase; B2M: β2-microglobulin.

RT-PCR
AKR1C1/1C22SGTG TGA AGC TGA ATG ATG GTC A 203-224815
 ASTCT GAT GCG CTG CTC ATT GTA GCT C 1017-993 
AKR1C1SAGT AAA GCT TTA GAG GCC ACNM_001353277-296591
 ASCAC CCA TGG TTC TTC TCG G 867-849 
AKR1C2SGTA AAG CTC TAG AGG CCG TNM_001354490-508590
 ASCAC CCA TGG TTC TTC TCG A 1079-1061 
sEpHSGCC GCC ATG ACG CTG CGC GNM_00197976-94487
 ASGTT TGA CCA TTC CCA CCT GA 562-543 
mEpHSCAT GTG GCT AGA AAT CCT CCNM_000120275-2941369
 ASTCA TTG CCG CTC CAG CAC CGA CA 1643-1621 
CBR1SCCC CGT TCA GCC ATG TCG TCC GGNM_001757118-140829
 ASTCT TCT CTG AAA CAA ATT GTC CAT GGG G 946-919 
CBR3SGCA GCC GCG TGG CGC TGG TGA CCG GGGNM_001236239-265405
 ASTGA TAT TCA CCA CTC TCC CAT GAG G 643-619 
GSTSCAG AGG AGG TCG CAG TTC AGM99422211-230690
 ASCAT CCC TTA GCC CAG TCA AG 900-881 
MDR1SAAA GAT CAA CTC GTA GGA GTG TCC GTG GAT CANM_0009272412-2443620
 ASTGC TAT TGC AAT GAT GGG TAC AAT TGC TAA G 3031-3001 
β-actinSTGA AGT ACC CCA TCG AGC ACGNM_001101273-293755
 ASAGT GAT CTC CTT CTG CAT CCT GT 1027-1005 
Real-time RT-PCR
AKR1C1/1C22SGGT CAC TTC ATG CCT GTC CT 220-239244
 ASACT CTG GTC GAT GGG AAT TG 463-444 
B2MSAGC AGA GAA TGG AAA GTC AAANM_004048162-182266
 ASATG CTG CTT ACA TGT CTC GAT 427-407 

Quantitative real-time PCR.

Real-time PCR was performed using the LightCycler Instrument and DNA Master SYBR Green I reaction mix (Roche, Mannheim, Germany). For each sample, 2 μl of cDNA was amplified in a 20 μl reaction capillary containing 0.5 μM of each primer, 3 mM MgCl2 and 2 μl of LightCycler DNA Master SYBR Green I (10X). Simultaneously, a blank was performed by incubating H2O PCR grade instead of sample cDNA. The primer sequences are listed in Table I. The PCR conditions were set up as follows: denaturation at 95°C for 30 sec, followed by 45 cycles of 95°C for 10 sec; 55°C for 10 sec; 72°C for 30 sec. Fluorescence was acquired at the end of every 72°C extension phase. The housekeeping gene β2-microglobulin served as an endogenous control. The gene level of AKR1C1/1C2 was normalized to the corresponding β2-microglobulin levels.

Evaluation of protein expression by western blot

Protein extraction and Western blots were performed as previously described in detail.10, 11 The primary antibodies used were as follows: mouse antihuman AKR1C1/1C2 monoclonal antibody from Cashmere Scientific (Taipei, Taiwan), mouse anti-AIF (apoptosis inducing factor) and anti-β-actin monoclonal antibody from Sigma, rabbit anti-NBS1 and anti-Chk2 polyclonal antibody from Novus Biologicals (Littleton, CO). Briefly, 40 μg total protein of each sample was resolved on 10% SDS-PAGE gel and then transferred onto nitrocellulose membrane (Schleicher & Schuell, Dassel, Germany). Membranes were blocked with 5% skimmed milk/TBS and incubated overnight at 4°C with each primary antibody. After washing, the blots were incubated with horseradish peroxidase-conjugated secondary antibody for 1 hr, developed using enhanced chemiluminescence reagent (Pierce) and exposed on an X-Omat film (Eastman Kodak, Rochester, NY).

Colony formation assay

Drug sensitivity was determined by colony formation assay. In brief, cells were plated at 100, 1,000 or 10,000 cells/well onto a 6-well plate and incubated for 18 hr to allow cells to adhere. Then the cells were treated with various doses of drugs for 2 hr. The control groups were treated with the same dilution of DMSO. After treatment, drug-containing medium was removed, followed by wash with HBSS for twice and replacement with fresh culture medium. After culture for 12 days at 37°C, the colonies were washed with PBS, stained with crystal violet, and counted under a magnifier. Colony formation was expressed as percent of control. IC50 was defined as the concentration of drug that resulted in a 50% inhibition in cell growth.

Statistical analyses

The IC50 values were determined using the sigmoid-Emax model which could be represented by the following equation modified from the Hill's31:

equation image

where Emax is the full range of inhibitory effect by the drug, C is the drug concentration, and m is the Hill coefficient describing the steepness of the concentration-effect relationship. The curve fitting and parameter estimation were performed by nonlinear regression analysis. Where appropriate, data are presented as mean ± SD. Comparisons between two independent treatment groups were carried out by two-tailed Student's t test. A p value of less than 0.05 was considered as statistically significant. All the analyses were performed with SPSS software ver. 12.0 (SPSS, Chicago, IL).

Results

The expression of AKR1C1/1C2 was associated with drug resistance in NSCLC cells

The expression of AKR1C1/1C2 in various NSCLC cell lines was evaluated by RT-PCR and Western blot (Fig. 1a), and showed high levels in H838, H1437 and H1648, and weak or undetectable levels in H23, H2087, H2009 and H2126. The 815-bp PCR products of AKR1C1 and AKR1C2, using the common AKR1C1/1C2 primer, differed in 18 base pairs and showed distinct restriction fragment length polymorphism after NcoI digestion (Fig. 1b). The patterns of PCR restriction fragment length polymorphism were compatible with the results of isoform-specific RT-PCR, showing concomitant expression of AKR1C1 and AKR1C2 in H838 and H1437 cells. The specificity of the PCR product was verified by DNA sequencing, which matched to AKR1C1 and AKR1C2 (data not shown). Interestingly, the level of AKR1C1/1C2 expression was correlated with chemoresistance in respective cells. As summarized in Table II, cells with high AKR1C1/1C2 expression, that is, H838, H1437 and H1648, tended to be less sensitive to cisplatin and adriamycin treatment, showing relatively higher values of IC50 as compared with those of cells exhibiting low levels of AKR1C1/1C2.

Figure 1.

AKR1C1/1C2 expression was inversely related with DNA repair, cell cycle and apoptosis-related factors in NSCLC cell lines. (a) RT-PCR and Western blots demonstrated AKR1C1/1C2 expression in 7 NSCLC cell lines. AKR1C1/1C2 expression was high in H1437, H1648 and H838, weak or undetected in H2009, H2087, H2126 and H23. (b) The amplicons of AKR1C1 and AKR1C2 produced by the common AKR1C1/1C2 primer (*) exhibited distinct patterns after digestion with NcoI restriction endonuclease. A 657 bp fragment indicated the presence of AKR1C1 amplicon while a 430 bp fragment pointed to AKR1C2. The patterns of the PCR restriction fragment length polymorphism and the isoform-specific RT-PCR (†) both showed concomitant expression of AKR1C1 and AKR1C2 in H838 and H1437 cells. (c) RT-PCR showing the expression of AKR1C1/1C2, soluble epoxide hydrolase (sEpH), microsomal epoxide hydrolase (mEpH), carbonyl reductase-1 (CBR1), carbonyl reductase-3 (CBR3), glutathione-S-transferase (GST) and MDR1 in various NSCLC cell lines. No significant association existed between AKR1C1/1C2 and other chemoresistance-related factors. (d) Expression of AKR1C1/1C2, NBS1, apoptosis inducing factor (AIF) and Chk2 were determined by immunoblotting with antibodies specific to the respective protein. There was an inverse relationship between the expression AKR1C1/1C2 and that of NBS1 and AIF.

Table II. The IC50 Values and Fractions of Colony Formation in Response to Escalating Doses of Cisplatin and Adriamycin in NSCLC Cells
Cell linesColony formation1 (%)IC502 (μM)
 Concentrations of Cisplatin 
0 μM0.5 μM1 μM2.5 μM10 μM
H23100.0 ± 0.054.7 ± 9.3913.5 ± 4.20.9 ± 0.30.6 ± 0.10.526 ± 0.051
H838100.0 ± 0.064.0 ± 10.841.3 ± 18.118.7 ± 8.210.2 ± 3.70.807 ± 0.330
H1437100.0 ± 0.087.3 ± 9.169.3 ± 13.320.9 ± 3.47.0 ± 1.91.267 ± 0.156
H1648100.0 ± 0.080.7 ± 13.978.3 ± 17.018.4 ± 3.210.0 ± 1.21.345 ± 0.287
H2087100.0 ± 0.044.3 ± 13.113.6 ± 4.81.7 ± 0.90.1 ± 0.10.458 ± 0.103
H2126100.0 ± 0.070.0 ± 8.531.0 ± 11.119.6 ± 3.48.0 ± 1.60.752 ± 0.174
 Concentrations of Adriamycin 
0 μM0.1 μM0.5 μM1.0 μM2.5 μM
  • 1

    The colony formation data are shown as percent of corresponding untreated cells; each test point represents the mean of triplicate experiments, presented as mean ± SD.

  • 2

    The IC50 was determined by nonlinear regression analyses using the Sigmoid-Emax model and presented as mean ± SD.

H23100.0 ± 0.047.6 ± 10.56.9 ± 1.70.7 ± 0.10.0 ± 0.00.095 ± 0.025
H838100.0 ± 0.0108.7 ± 11.588.3 ± 1.084.6 ± 0.384.7 ± 5.514.716 ± 0.480
H1437100.0 ± 0.086.3 ± 6.780.7 ± 7.636.3 ± 8.711.7 ± 5.50.822 ± 0.150
H1648100.0 ± 0.086.0 ± 9.583.0 ± 7.061.3 ± 11.916.7 ± 6.11.196 ± 0.194
H2087100.0 ± 0.057.0 ± 17.17.9 ± 1.40.9 ± 0.10.0 ± 0.00.116 ± 0.042
H2126100.0 ± 0.066.7 ± 9.331.7 ± 4.015.3 ± 6.72.1 ± 0.90.207 ± 0.048

Inverse correlation of AKR1C1/1C2 expression with DNA repair, cell cycle and apoptosis related factors

The association between AKR1C1/1C2 and various factors that had been found to be linked to drug resistance, including soluble epoxide hydrolase (sEpH), microsomal epoxide hydrolase (mEpH), carbonyl reductase-1 (CBR1), carbonyl reductase-3 (CBR3), glutathione-S-transferase (GST) and MDR1, was evaluated by RT-PCR. Nonetheless, no significant association was found (Fig. 1c). We further examined the association between AKR1C1/1C2 and NBS1, AIF or Chk2. It is worth noting that the expression of AKR1C1/1C2 was inversely correlated with NBS1 and AIF, and this reciprocal relationship was closely related to drug resistance. As shown in Figure 1d, the expression of NBS1 and AIF was clearly downregulated in H838, H1437 and H1648 cells, in parallel with overexpression of AKR1C1/1C2 and the drug-resistant phenotypes of these cells (see Table II). These data indicated that NBS1 and AIF might be implicated in AKR1C1/1C2-associated drug resistance.

Pro-inflammatory mediators could induce AKR1C1/1C2 expression in NSCLC cells

In order to further investigate the link of AKR1C1/1C2 expression with inflammation and drug resistance, H23 cells, which exhibited low baseline AKR1C1/1C2 expression and higher drug sensitivity (Fig. 1a and Table II), were chosen for the following experiments. Firstly, to assess the effects of pro-inflammatory factors on AKR1C1/1C2 expression, H23 cells were treated with 5 ng/ml of IL-6, 5 mM of n-sodium butyrate (n-BT), 25 ng/ml of phorbol myristate acetate (PMA), or 30 ng/ml of platelet activating factor (PAF) respectively. In 24 hr, immunoblots showed clearly upregulated AKR1C1/1C2 expression in H23 cells (Fig. 2a). On the other hand, addition of 50 μM of safingol, a protein kinase C inhibitor, could completely nullify the induction effect of IL-6 on AKR1C1/1C2 expression. The IL-6 effect was also partially inhibited by dexamethasone but not by 6-methoxy-2-naphthylacetic acid (6-MNA, 100 μM) (Fig. 2b).

Figure 2.

Pro-inflammatory mediators induced AKR1C1/1C2 expression in H23 cells. (a) H23 cells were treated with IL-6 (5 ng/ml), n-BT (5 mM), PMA (25 ng/ml) or PAF (30 ng/ml) for 24 hr. Immunoblots showed that AKR1C1/1C2 was clearly upregulated by these factors. Control, H23 cells without treatment. (b) Immunoblots showed that IL-6-induced AKR1C1/1C2 expression could be completely inhibited by concomitant treatment with safingol (50 μM), and partially by dexamethasone (Dexa). Control, H23 cells without treatment. (c,d) Time-course and dose-response relationships of IL-6-induced AKR1C1/1C2 expression in H23 cells. H23 cells were plated into T25 flasks at 1 × 105 cells/flask and cultured for 18 hr. The cells were then treated with 5 ng/ml of IL-6 for 0, 6, 12, 24 and 48 hr (c) or with various concentrations of IL-6 (0, 1, 5, 25 ng/ml) for 24 hr (d). RT-PCR showed that AKR1C1/1C2 transcription was highest in 24 hr after treatment and that IL-6 at a concentration greater than 5 ng/ml was sufficient to induce a significant AKR1C1/1C2 expression in H23 cells.

Drug resistance to cisplatin and adriamycin in H23 cells was also enhanced by pro-inflammatory mediators

H23 cells were pre-treated with IL-6 (5 ng/ml), n-BT (5 mM), PMA (25 ng/ml) or PAF (30 ng/ml) before exposure to cytotoxic agents. It is worth noting that the resistance to cisplatin and adriamycin in H23 cells was enhanced by treatment with these pro-inflammatory mediators. (Table III).

Table III. The Effect of Pre-Treatment with Proinflammatory Factors On Drug Resistance to Cisplatin and Adriamycin in H23 Cells
PretreatmentColony formation1 (%)IC502 (μM)
 Concentrations of Cisplatin 
0 μM0.5 μM1 μM5 μM
Control100.0 ± 0.029.3 ± 9.610.9 ± 1.11.2 ± 0.40.289 ± 0.132
IL-6 (5 ng/mL)100.0 ± 0.068.0 ± 11.536.7 ± 7.12.4 ± 1.20.733 ± 0.084
n-BT (5 mM)100.0 ± 0.083.3 ± 5.140.0 ± 7.91.9 ± 1.00.873 ± 0.101
PMA (25 ng/mL)100.0 ± 0.083.3 ± 6.754.0 ± 6.63.5 ± 0.81.069 ± 0.125
PAF (30 ng/mL)100.0 ± 0.052.3 ± 8.130.0 ± 8.51.2 ± 0.40.546 ± 0.126
 Concentrations of Adriamycin 
0 μM0.1 μM0.5 μM2.5 μM
  • 1

    The colony formation data are shown as percent of corresponding untreated cells; each test point represents the mean of triplicate experiments, presented as mean ± SD.

  • 2

    The IC50 was determined by nonlinear regression analyses using the Sigmoid-Emax model and presented as mean ± SD.

Control100.0 ± 0.030.7 ± 7.43.0 ± 0.80.1 ± 0.00.061 ± 0.014
IL-6 (5 ng/mL)100.0 ± 0.053.7 ± 14.324.7 ± 7.01.9 ± 0.50.129 ± 0.066
n-BT (5 mM)100.0 ± 0.060.0 ± 11.531.7 ± 6.42.0 ± 0.40.171 ± 0.060
PMA (25 ng/mL)100.0 ± 0.083.7 ± 8.041.0 ± 8.93.7 ± 1.40.369 ± 0.092
PAF (30 ng/mL)100.0 ± 0.057.7 ± 10.129.7 ± 6.41.7 ± 0.60.153 ± 0.052

Time-course and dose-response of IL-6-induced AKR1C1/1C2 expression

The time-course and dose-response relationships of IL-6-induced AKR1C1/1C2 expression are shown in Figures 2c2d. After IL-6 treatment, the transcription of AKR1C1/1C2 was highest in 24 hr, and a concentration of IL-6 greater than 5 ng/ml was sufficient to induce significant AKR1C1/1C2 expression in H23 cells. Accordingly, subsequent induction experiments were conducted using IL-6 at above concentration (5 ng/ml) and time (24 hr).

Wogonin and chrysin could inhibit IL-6-induced AKR1C1/1C2 expression

H23 cells were treated with IL-6 (5 ng/ml) plus 10 μM of baicalin, baicalein, wogonin or chrysin. As shown in Figure 3a, co-treatment with wogonin and chrysin, but not baicalin and baicalein, could suppress IL-6-induced AKR1C1/1C2 expression. The dose-response effects of wogonin and chrysin are demonstrated in Figure 3b. Real-time RT-PCR analyses confirmed the inhibitory effects of wogonin at a concentration higher than 2.5 μM and chrysin at 10 μM (Fig. 3c).

Figure 3.

Inhibitory effects of wogonin and chrysin on IL-6 responses in H23 cells. (a) H23 cells were treated with IL-6 (5 ng/ml) plus 10 μM of baicalin, baicalein, wogonin or chrysin. Wogonin and chrysin could inhibit IL-6-induced AKR1C1/1C2 expression, while baicalin and baicalein could not. (b) H23 cells were treated with 5 ng/ml IL-6 plus various concentrations of wogonin or chrysin for 24 hr. RT-PCR showed the dose-response effects wogonin and chrysin on inhibition of IL-6-induced AKR1C1/1C2 expression. (c) H23 cells were co-treated with IL-6 (5 ng/ml) and various concentrations of wogonin or chrysin for 24 hr. Real-time quantitative RT-PCR showed that wogonin greater than 2.5 μM and chrysin greater than 10 μM could inhibit IL-6-induced AKR1C1/1C2 expression. The blank controls were represented by the untreated H23 cells, while cells treated with IL-6 alone served as the reference groups. The data are shown as percentage of corresponding reference groups. Each experiment was performed in triplicates. Columns, means of replicate analyses (light column, wogonin; dark column, chrysin); bars, ±SD; †, cells treated with IL-6 alone serving as reference groups; *, statistically significant difference from the corresponding reference groups (p < 0.05) by two-tailed Student's t test.

Wogonin and chrysin could overcome IL-6-induced drug resistance

To investigate the effects of wogonin and chrysin on IL-6-induced chemoresistance, H23 cells were pre-treated with IL-6 plus various concentrations of wogonin or chrysin for 48 hr, followed by exposure to various concentrations of adriamycin for 2 hr. The growth inhibition was assessed by colony formation assay. As shown in Figure 4, IL-6 could render H23 cells more resistant to adriamycin. Pre-treatment with IL-6 resulted in an 8.8-fold increase in the IC50 of adriamycin from 0.018 to 0.159 μM (p < 0.001, Table IV). On the other hand, co-treatment with wogonin (5 μM) in addition to IL-6 could negate the IL-6 effect, reducing the IC50 to 0.027 μM (p = 0.001, compared with IL-6 treatment alone). Similarly, chrysin at 5 μM could also inhibit IL-6-induced chemoresistance (IC50 of adriamycin = 0.026 μM, p < 0.001 as compared with IL-6 treatment alone).

Figure 4.

The effects of IL-6, wogonin and chrysin on the sensitivity to adriamycin in H23 cells. (a) H23 cells were pre-treated with 5 ng/ml IL-6 (□), 5 ng/ml IL-6 plus 1 μM wogonin (⋄) or 5 ng/ml IL-6 plus 5 μM wogonin (▵) for 48 hr. The control (○) and the pre-treated cells were incubated with various concentrations of adriamycin for 2 hr. After 12 days of cell culture, the colonies were stained and counted. Pre-treatment with IL-6 could enhance the cell resistance to adriamycin, while wogonin at 5 μM could negate the IL-6 effect. (b) Colony formation assay showed that chrysin at 5 μM could also nullify IL-6-induced resistance to adriamycin. The data at each test point represents the mean of triplicate experiments; bars, ±SD; *, statistically significant difference (p < 0.05) from the corresponding controls (○) by two-tailed Student's t test.

Table IV. Effects of IL-6, Wogonin and Chrysin on Drug Resistance to Adriamycin in H23 Cells
PretreatmentColony formation1 (%)IC502 of Adriamycin (μM)p value3 (vs. control)p value3 (vs. IL-6)
Concentrations of Adriamycin
0 μM0.02 μM0.1 μM0.5 μM2.5 μM
  • 1

    The colony formation data are shown as percent of corresponding untreated cells; each test point represents the mean of triplicate experiments, presented as mean ± SD.

  • 2

    The IC50 was determined by nonlinear regression analyses using the Sigmoid-Emax model and presented as mean ± SD.

  • 3

    The p values were calculated using two-tailed Student's t test.

Control100.0 ± 0.047.1 ± 4.417.1 ± 2.23.5 ± 1.40.0 ± 0.00.018 ± 0.003<0.001
IL-6100.0 ± 0.0107.7 ± 7.561.9 ± 4.223.2 ± 3.20.0 ± 0.00.159 ± 0.017<0.001
IL-6 + Wogonin (1 μM)100.0 ± 0.089.7 ± 11.260.0 ± 4.214.8 ± 3.30.0 ± 0.00.124 ± 0.0160.0010.102
IL-6 + Wogonin (5 μM)100.0 ± 0.054.8 ± 7.828.4 ± 2.46.5 ± 2.40.0 ± 0.00.027 ± 0.0070.1800.001
IL-6 + Chrysin (1 μM)100.0 ± 0.089.7 ± 10.350.3 ± 4.221.3 ± 4.70.0 ± 0.00.099 ± 0.009<0.0010.012
IL-6 + Chrysin (5 μM)100.0 ± 0.053.5 ± 4.032.9 ± 4.28.4 ± 0.90.0 ± 0.00.026 ± 0.0030.069<0.001

Discussion

We have previously postulated that DDH was linked to chronic inflammation and cancer.11 Our subsequent work further showed that overexpression of DDH in human cancers was closely associated with disease progression, drug resistance and poor prognosis.10, 11, 21 In the current study, we first demonstrated that IL-6, a proinflammatory cytokine, was capable of inducing the expression of AKR1C1/1C2 in NSCLC cells, and consequently rendering the cells more resistant to chemotherapeutic drugs. IL-6 provokes a broad range of cellular and physiological responses. A couple of previous studies have demonstrated the prognostic value of serum IL-6 level in various cancers including NSCLC.13, 32 Recent studies further showed that IL-6 could cause drug resistance in cancer cells.33, 34, 35 IL-6 signals through a cell-surface assembly composed of two different subunits, an alpha subunit that produces ligand specificity and gp130, a subunit that shared in common with other cytokines in the IL-6 family.36 Binding of IL-6 to its receptor initiates diverse signaling pathways, including the JAK/STATs, Ras/MAPK and PI3-kinase signaling.37, 38 The PKC pathway is also found to play important regulatory roles in IL-6 signaling.39

On the other hand, Ciaccio and colleagues first demonstrated that DDH was overexpressed in ethacrynic acid-resistant human colon carcinoma cells, and provided evidence showing that the gene is regulated by an anti-oxidant response element (ARE).40, 41 The ARE is a unique cis-acting sequence found in the 5′-regulatory region of a number of genes encoding enzymes involved in the phase II metabolism of xenobiotics. The activation of ARE is mediated by NF-E2-related factor 2 (Nrf2), a basic leucine zipper transcription factor.42 Subsequent studies showed that PKC plays a critical role in the regulation of ARE-directed gene expression through phosphorylation of Nrf2, leading to the release of Nrf2 from INrf2, an inhibitor of Nrf2, and facilitating its nuclear translocation.43, 44 In this study, we showed that the IL-6 induction of AKR1C1/1C2 could be completely suppressed by the PKC inhibitor safingol, highlighting the central role of PKC in the link of the pro-inflammatory and pro-oxidant signals.

Previous studies showed that forced expression of DDH conferred resistance to platinum drugs in various human cancer cell lines including lung carcinoma cells.20, 45 We also found that transfection of AKR1C1 or AKR1C2 into lung adenocarcinoma cells could enhance cell resistance to cisplatin, adriamycin and irradiation.23 Compatible with previous results, this study also demonstrated the correlation of AKR1C1/1C2 expression and chemoresistance in NSCLC cells. The pharmacologic mechanisms of cisplatin and adriamycin are apparently different. Cisplatin interacts with DNA to form DNA adducts, primarily intrastrand crosslink adducts, and disrupts DNA function.46 Adriamycin inhibits DNA and RNA synthesis through the inhibition of topoisomerase II and intercalation at points of local uncoiling of the double helix.47 The molecular mechanisms of the cross-resistance to chemotherapeutic agents appear to be multifactorial and may include: increased expression of factors involved in reducing intracellular drug concentration; alterations in drug-target interaction; and changes in cellular response, in particular the ability to repair DNA damage, and defects in apoptotic pathways, etc. In the present study, we examined various factors known to contribute to chemoresistance through enhanced drug metabolism, efflux or excretion, including sEpH, mEpH, CBR1, CBR3, GST and MDR1. However, we did not find any association between the expression of AKR1C1/1C2 and these factors. Thus we further investigate several factors involved in DNA repair and regulation of cell cycle and apoptosis, and we first demonstrated an inverse relationship between the expression of AKR1C1/1C2 and NBS1 or AIF, which also showed to correlate with drug resistance. The NBS1 protein, also known as nibrin or p95, is responsible for the repair of DNA double-strand breaks,48 while AIF is a nucleus-encoded mitochondrial oxidoreductase that has, in recent years, emerged as an important protein closely associated with apoptosis.49 The connection of drug resistance with the reciprocal expression of AKR1C1/1C2 and NBS1 or AIF indicated that DDH-associated chemoresistance might involve the DNA repair and apoptosis pathways. However, the exact mechanisms warrant further investigation.

Flavonoids are polyphenolic compounds naturally present in plants. Several studies have demonstrated that flavonoids can inhibit several kinases involved in signal transduction, including PKC and tyrosine kinases.50Scutellaria baicalensis has been widely used as an anti-inflammatory herbal remedy in the traditional Chinese and Japanese medicine. In this study, we tested the major flavonoids in Scutellaria baicalensis and demonstrated that IL-6-induced AKR1C1/1C2 expression and drug resistance could be inhibited by wogonin (5,7-dihydroxy-8-methoxyflavone) and chrysin (5,7-dihydroxyflavone). Wogonin has shown to exert various anti-inflammatory effects. It inhibited lipopolysaccharide (LPS)-induced production of nitric oxide and PGE2 in macrophages.51, 52 Nakamura et al. further showed its capability in inhibiting the expression of cytokines via the suppression of NF-κB binding activities.53 Like wogonin, chrysin also poses multiple anti-inflammatory effects. It was found to exhibit weak peroxisome proliferator-activated receptor-γ (PPAR-γ) agonist activities and inhibit the production of inflammatory cytokines and expression of inducible nitric oxide synthase and COX-2.54 Recently, chrysin was also found to inhibit LPS-induced COX-2 expression via inhibition of nuclear factor for IL6 (NF-IL6).55 Our data suggested that both wogonin and chrysin might serve to affect drug sensitivity of cancer cells by modulating the signaling pathways of inflammatory cytokines.

In conclusion, this study demonstrated, for the first time, that IL-6 can induce AKR1C1/1C2 expression in NSCLC cells and contribute to drug resistances to cisplatin and adriamycin. The mechanism may involve the activation of protein kinase C. The AKR1C1/1C2-associated chemoresistance may be related to altered control of DNA repair and apoptosis. Wogonin and chrysin, the flavonoids present in Scutellaria baicalensis, can suppress IL-6-induced AKR1C1/1C2 overexpression and overcome drug resistance. These data suggested that AKR1C1/1C2 might serve as a marker of inflammation-associated drug resistance. Further studies are warranted to investigate the potential of wogonin and chrysin as an adjunct in the treatment of multidrug resistant NSCLC, especially for tumors with high AKR1C1/1C2 expression.

Acknowledgements

We thank Ms. Li-Ling Yang and Ms. Yi-Hsiu Kuo for their excellent technical assistance, and Ms. Hui-Chen Lee, Biostatistics Task Force, Taipei Veterans General Hospital, for the assistance in statistical analysis.

Ancillary