Ectopic NGAL expression can alter sensitivity of breast cancer cells to EGFR, Bcl-2, CaM-K inhibitors and the plant natural product berberine

Abstract

Neutrophil gelatinase-associated lipocalin (NGAL, a.k.a Lnc2) is a member of the lipocalin family and has diverse roles. NGAL can stabilize matrix metalloproteinase-9 from autodegradation. NGAL is considered as a siderocalin that is important in the transport of iron. NGAL expression has also been associated with certain neoplasias and is implicated in the metastasis of breast cancer. In a previous study, we examined whether ectopic NGAL expression would alter the sensitivity of breast epithelial, breast and colorectal cancer cells to the effects of the chemotherapeutic drug doxorubicin. While abundant NGAL expression was detected in all the cells infected with a retrovirus encoding NGAL, this expression did not alter the sensitivity of these cells to doxorubicin as compared with empty vector-transduced cells. We were also interested in determining the effects of ectopic NGAL expression on the sensitivity to small-molecule inhibitors targeting key signaling molecules. Ectopic NGAL expression increased the sensitivity of MCF-7 breast cancer cells to EGFR, Bcl-2 and calmodulin kinase inhibitors as well as the natural plant product berberine. Furthermore, when suboptimal concentrations of certain inhibitors were combined with doxorubicin, a reduction in the doxorubicin IC₅₀ was frequently observed. An exception was observed when doxorubicin was combined with rapamycin, as doxorubicin suppressed the sensitivity of the NGAL-transduced MCF-7 cells to rapamycin when compared with the empty vector controls. In contrast, changes in the sensitivities of the NGAL-transduced HT-29 colorectal cancer cell line and the breast epithelial MCF-10A cell line were not detected compared with empty vector-transduced cells. Doxorubicin-resistant MCF-7/Dox^R cells were examined in these experiments as a control drug-resistant line; it displayed increased sensitivity to EGFR and Bcl-2 inhibitors compared with empty vector transduced MCF-7 cells. These results indicate that NGAL expression can alter the sensitivity of certain cancer cells to small-molecule inhibitors, suggesting that patients whose tumors exhibit elevated NGAL expression or have become drug-resistant may display altered responses to certain small-molecule inhibitors.

Keywords: :

• NGAL
• Lcn2
• lipocalins
• siderocalins
• targeted therapy
• inhibitor sensitivity
• EGFR
• rapamycin
• berberine
• BCL-2
• calmodulin kinase
• breast cancer
• colorectal cancer

Introduction

Identification of the signaling pathways that are critical for sensitivity to targeted therapy as well as conventional therapy is essential for improved cancer treatments. In fact, it has recently been demonstrated that in order for some targeted cancer therapy treatments to be effective, specific target genes need to be either mutated or overexpressed. Moreover, cells that are initially sensitive to targeted therapy often develop resistance. Certain types of cancers, namely melanoma,^1-17 chronic myeloid leukemia (CML)^18,19 and non-small cell lung cancer (NSCLC)^20,21 have been intensively investigated for the mechanisms of sensitivity and resistance to small-molecule inhibitors. Thus, it is essential to understand why certain cancer patients are sensitive or develop resistance to various therapeutic approaches and whether the sensitivity results from intrinsic or extrinsic events.

A common phenomenon that occurs after treatment of cancer patients with chemotherapeutic drugs is drug resistance. The mechanisms behind these developments are many and include increased expression of drug transporters, amplification of critical survival genes, genetic mutations and deletions or increased activation of certain signal transduction pathways.^22-26 Therefore, it is also important to understand how cancers become drug-resistant and whether or not their drug resistance can be reversed.

Over the past 35 y, many genes have been identified which can cause or contribute to the formation of cancer.^27,28 These include two major classes of genes, the oncogenes^29-33 and the tumor-suppressor genes such as retinoblastoma (RB),^34-37 TP53,^38-40 BRCA1,^41,42 PTEN,⁴³ TSC1 and TSC2.^44-47 Many of these oncogenes and tumor suppressors are often critical regulators of cellular senescence.^48-53 Moreover, microRNAs (miRNAs)^54-60 and epigenetic modifications^61-63 have been shown to play important roles in regulating cancer progression. Certain miRNAs may be induced by drugs such as metformin, which can control cancer growth.⁶⁰ Some miRNAs may be regulated by epigenetic mechanism by tumor-suppressor genes such as BRCA1.⁶³

In some cases, the genetic culprit involved in a particular cancer may be known [e.g., BCRABL in chronic myeloid leukemia,^18,19 HER2 in certain forms of breast cancer,^64-66 EGFR in certain lung cancers,^25,27,28 BRAF in melanomas, thyroid cancers, non-small cell lung cancers and colorectal cancers (CRC)^{1-18,25,27,28}]. However, in most cases, there are multiple genetic and epigenetic events occurring that can interact and result in a cancer cell capable of becoming metastatic and/or drug resistant. In addition, there are other important metabolic contributions by the tumor microenvironment that aid in the progression of the cancer cell as well as the development of sensitivity/resistance to various therapeutic approaches and the survival of cancer-initiating cells (CICs).^67-69

One factor that may be important for cancer survival and metastasis is neutrophil gelatinase-associated lipocalin (NGAL). One of the genomic responses to common cancer treatments such as radiation and chemotherapy is the induction of NGAL expression.^70-75 NGAL may act to stabilize MMP-9 and increase its ability to degrade the extracellular matrix, thereby promoting metastasis.

NGAL expression is regulated by the transcription factors NFκB, CEBP and others.^76-79 Radiation and chemotherapy can induce reactive oxygen species (ROS) that result in NFκB activation^80-83 and subsequent downstream NGAL transcription. In addition, the tumor microenvironment can alter intracellular NFκB activity.⁸³ Chemo- and radiotherapy could result in the synthesis of NGAL in cancer cells, which may lead to the development of therapy-resistant cells. These cells can contribute to the reemergence and metastasis of the cancer, as increased NGAL expression may allow the cells to persist under conditions where therapy-sensitive cancer cells would not normally survive.

Cancer cells have increased demands for intracellular iron. NGAL is a member of the lipocalin family and, as such, is capable of serving as a siderocalin or molecule involved in the transport of iron and other molecules.⁸⁴ Iron is essential for many key processes, including the rate-limiting step in DNA synthesis performed by ribonucleotide reductase.⁸⁵ Iron (Fe²⁺) is also required for cells to progress through the cell cycle from G₁ to S phase. Tumor cells have a high requirement for iron and express elevated levels of the transferrin receptor-1.^86-89 Novel chelators of iron are being considered for cancer treatment.⁹⁰ Iron chelators, such as Desferrioxamine (DFO), inhibit cellular iron transport and have been evaluated in various cancer clinical trials.⁹¹ Oxygen and iron concentrations may be altered in the tumor microenvironment due to drastic tumor growth.^92-94 In order for a cancer cell to survive, invade and metastasize, it may have to have increased iron transport as well as elevated glycolysis.^67-69,95,96 The role of iron transport in chemotherapeutic drug resistance of cancer cells is complex and may depend on the particular drug and cancer type investigated.^97,98 Interestingly, some iron depletors have been shown to decrease resistance of certain cancer cells to chemotherapeutic drugs including doxorubicin.^99,100

Increased levels of NGAL have been detected in the urine of patients with various types of cancer (i.e., brain, breast, colon, ovarian, pancreatic and prostate). Novel non-invasive urine-based tests could prove useful for the detection and/or prognosis of many cancer types.^101-104 The role(s) of NGAL in chemotherapeutic drug resistance, invasion, cancer metastasis and sensitivity to targeted therapy have not been fully elucidated. Targeting NGAL could result in decreased cancer cell survival and tumor regression as well as improve the effectiveness of radiation and chemotherapy in cancer therapy. NGAL is considered by some scientists to possess characteristics of an oncogene. In some studies, NGAL has been shown to increase the mobility, invasion, metastasis and tumorigenesis of certain cancer cells (breast, CRC).^105-108 Elevated expression of NGAL increases the invasiveness of certain cancer cell types, while inhibition of NGAL expression decreases their invasiveness and metastasis.^72-75,105 Novel approaches to target MMP-9/NGAL are needed, as MMP-9 inhibitors have not performed well in clinical cancer trials,¹⁰⁹ and NGAL has functions which are independent of MMP-9.

NGAL may exert many different effects that are important in invasion and metastasis. NGAL can stabilize MMP-9 at the cell surface,^110-112 and this complex, in association with CD44, may promote the cleavage of E-Cadherin (E-Cad) into soluble(s) E-Cad thereby inducing epithelial-mesenchymal transition (EMT).^112,113 Alternatively, NGAL may be important in the transport of inhibitors and natural products into cells or preventing their efflux from cells. In the following studies, we examined the effects of ectopic NGAL expression on the sensitivity of breast cancer and CRC to several small-molecule inhibitors targeting critical molecules in signal transduction pathways. Ectopic expression of NGAL increased the sensitivity of MCF-7 breast cancer cells to EGFR, CaM-K and Bcl-2 inhibitors as well as the natural plant product berberine. In contrast, ectopic expression of NGAL did not alter the sensitivity of the CRC line or the breast epithelial MCF-10A line to the various small-molecule inhibitors or natural products. We chose to examine the effects of ectopic NGAL expression on sensitivity to doxorubicin on two different types of cancers: breast cancer cells, which are generally sensitive to doxorubicin therapy, and CRC cells, which are considered resistant to doxorubicin therapy. Addition of suboptimal doses of some inhibitors lowered the IC₅₀ of doxorubicin in MCF-7 breast cancer cells with the exception of rapamycin. In contrast, doxorubicin appeared to inhibit the effects of rapamycin on the normally sensitive MCF-7/NGAL but not MCF-7/pLXSN cells. These results could have clinical significance, as NGAL is often expressed at high levels in certain advanced cancer patients, and its expression is induced after chemo- and radiotherapy. NGAL expression could also alter the sensitivity of cancer and other patients to various small-molecule inhibitors and natural products such as berberine, which are used in traditional medicine.

Results

Effects of enforced NGAL expression on sensitivity to the EGFR inhibitor AG1478

MCF-10A, MCF-7 and HT-29 cells were infected with a retrovirus encoding NGAL or the empty retrovirus pLXSN. NGAL was detected in the supernatants from NGAL retrovirus-infected cells^114,115 but not in the empty vector pLXSN virus-infected cells.¹¹⁵

We examined the effects of elevated NGAL expression on the sensitivity of all of the cell lines to the various inhibitors and doxorubicin by MTT assays. Graphs in Figure 1 represent the effects of varying concentrations of the EGFR inhibitor AG1478 (panels A, D and G) doxorubicin (panels B, E and H), and varying concentrations of doxorubicin with a constant concentration the EGFR inhibitor (panels C, F and I).

Figure 1. Sensitivity of NGAL- and pLXSN-infected cells and doxorubicin-resistant MCF-7/Dox^R cells to the EGFR inhibitor AG1478, doxorubicin and the combination of doxorubicin and a constant dose of 250 nM AG1478. Cells were collected and seeded (2,000 cells/well) in 96-well plates. The following day, serial 2-fold dilutions of AG1478 (A, D and G), doxorubicin (B, E and H) or serial 2-fold dilutions of doxorubicin and a constant dose of 250 nM AG1478 (C, F and I) were added to the wells. Four days later, MTT assays were performed. (A–C) MCF-10A/pLXSN (solid squares), MCF-10A/NGAL (solid upright triangles), (D–F) MCF-7/pLXSN (solid squares), MCF-7/NGAL (solid upright triangles) and 25 nM doxorubicin-selected MCF-7/Dox^R cells (solid downward triangles), (G–I) HT-29/pLXSN (solid squares), HT-29/NGAL (solid upward triangles). A hatched horizontal line is present at the 50% relative growth mark from which the IC₅₀ can be calculated. A vertical arrow indicates the IC₅₀. The statistical significance was determined by the unpaired t-test (***, p < 0.001). All the experiments in this figure were performed at the same time (set up on the same day). These experiments were repeated multiple times, and similar results were obtained.

Ectopic NGAL expression did not appear to alter the sensitivity of MCF-10A/NGAL cells to either the EGFR inhibitor AG1478 (panel A) or doxorubicin (panel B). Interestingly, a dose of the EGFR inhibitor at approximately the IC₅₀ completely eliminated the growth of these cells, which are normally cultured in medium containing rEGF (panel C). These results demonstrate that the toxic effects of doxorubicin can be enhanced by the EGFR inhibitor in the breast epithelial MCF-10A line.

In contrast to the results observed with the breast epithelial MCF-10A/NGAL cells, ectopic NGAL expression did alter the sensitivity of MCF-7/NGAL cells to the EGFR inhibitor AG1478 (panel D) but not doxorubicin (panel E). In addition, the MCF-7/Dox^R line was very sensitive to the EGFR inhibitor AG1478 (panel D) but was highly resistant to doxorubicin (panel E). Interestingly, a dose of the EGFR inhibitor reduced the concentration of doxorubicin required to reach the IC₅₀ at least 10-fold in all the MCF-7 cells (panel F). The MCF-7/Dox^R cells were included in these studies as they were derived directly from MCF-7 cells by selection in medium containing 25 nM doxorubicin for prolonged periods of time.¹¹⁴

NGAL expression did not significantly alter the sensitivity of HT29/NGAL cells to either the EGFR inhibitor AG1478 (panel G) or doxorubicin (panel H). In fact, the HT29 cells appeared to be very resistant to the effects of the EGFR inhibitor. A constant dose of 250 nM EGFR inhibitor did not produce any additive effects to the doxorubicin IC₅₀ in HT29 cells (panel I), similar to what was seen in the other cell lines. The IC₅₀s for the drug-treated cells are presented in Table 1.

Table 1. Effects of ectopic NGAL expression on the sensitivity of cells to doxorubicin, small-molecule inhibitors and the natural plant product berberine.¹

Drug treatment	MCF-10A/pLXSN	MCF-10A/NGAL	MCF-7/pLXSN	MCF-7/NGAL	MCF-7/Dox^R	HT-29/pLXSN	HT-29/NGAL
AG1478 (EGFR Inh)	300 nM	300 nM	> 2,000 nM	30 nM	6 nM	> 2,000 nM	> 2,000 nM
ABT-737 (Bcl-2 Inh)	400 nM	300 nM	1,000 nM	1 nM	1.5 nM	600 nM	600 nM
Rapamycin (mTORC1 Inh)	0.2 nM	0.2 nM	0.15 nM	0.2 nM	> 100 nM	> 100 nM	> 100 nM
KN-93 (CaMK Inh)	> 10,000 nM	10,000 nM	110 nM	12 nM	2,500 nM	500 nM	550 nM
Berberine	> 2,000 nM	> 2,000 nM	2,000 nM	12 nM	> 2,000 nM	> 2,000 nM	2,000 nM
Doxorubicin	30 nM	25 nM	80 nM	60 nM	2,000 nM	40 nM	40 nM
Dox + 250 nM AG1478	< 1nM	< 1nM	12 nM	12 nM	80 nM	50 nM	40 nM
Dox + 50 nM ABT-737	20 nM	18 nM	7 nM	15 nM	8 nM	1.8 nM	3 nM
Dox + 5 nM Rapa	< 2nM	< 2 nM	1.5 nM	70 nM	2.5 nM	28 nM	28 nM
Dox + 250 nM KN-93	18 nM	12 nM	12 nM	< 2 nM	30 nM	40 nM	40 nM
Dox + 250 nM Ber	28 nM	20 nM	70 nM	10 nM	8 nM	50 nM	50 nM

¹MTT analysis was performed with different unselected cancer lines and certain 25 nM doxorubicin resistant (DoxR). Determined by plating 2,000 cells/well in 96-well plates in phenol red-free RPMI , + 10% FBS. Serial 2-fold dilutions (n = 12 dilutions) of doxorubicin were dispensed into eight wells per each doxorubicin concentration after the first day. MTT analysis was performed after four additional days of incubation and results were normalized to untreated cells as described.¹¹⁴ All IC50s are estimated values derived from the graphs presented in Figures 1–5.

Effects of enforced NGAL expression on sensitivity to the Bcl-2 inhibitor ABT-737

The effects of varying concentrations of the Bcl-2 inhibitor ABT-737 (panels A, D and G), doxorubicin (panels B, E and H) or varying concentrations of doxorubicin and a constant concentration of ABT-737 (panels C, F and I) on growth in NGAL-expressing cells are presented in Figure 2. The results with the doxorubicin titrations are presented again to allow direct comparison with the results obtained with doxorubicin titrations combined with a constant dose of the Bcl-2 inhibitor.

Figure 2. Sensitivity of NGAL- and pLXSN-infected cells and doxorubicin-resistant MCF-7/Dox^R cells to the Bcl-2 inhibitor ABT-737, doxorubicin and the combination of doxorubicin and a constant dose of 50 nM ABT-737. Cells were collected and seeded (2,000 cells/well) in 96-well plates. The following day, serial 2-fold dilutions of ABT-737 (A, D and G), doxorubicin (B, E and H) or serial 2-fold dilutions of doxorubicin and a constant dose of 50 nM ABT-737 (C, F and I) were added to the wells. Four days later, MTT assays were performed. (A–C) MCF-10A/pLXSN (solid squares), MCF-10A/NGAL (solid upright triangles), (D–F) MCF-7/pLXSN (solid squares), MCF-7/NGAL (solid upright triangles) and 25 nM doxorubicin-selected MCF-7/Dox^R cells (solid downward triangles), (G–I) HT-29/pLXSN (solid squares), HT-29/NGAL (solid upward triangles). A hatched horizontal line is present at the 50% relative growth mark from which the IC₅₀ can be calculated. A vertical arrow indicates the IC₅₀. The statistical significance was determined by the unpaired t-test (***, p < 0.001). All the experiments in this figure were performed at the same time (set up on the same day). These experiments were repeated multiple times and similar results were obtained.

Overexpression of NGAL did not alter the IC₅₀ of MCF-10A/NGAL cells to ABT-737; however, MCF-10A/NGAL cells were more sensitive to lower concentrations of the Bcl-2 inhibitor than MCF-10/pLXSN cells (panel A). The addition of the Bcl-2 inhibitor decreased the concentration of doxorubicin required to reach the IC₅₀ in both MCF-10A/pLXSN and MCF-10A/NGAL by less than 2-fold (panel C).

In contrast to the results observed with the breast epithelial MCF-10A cells, NGAL expression did exhibit an effect on the sensitivity of MCF-7/NGAL cells to the Bcl-2 inhibitor ABT-737 (panel D) but not to doxorubicin only (panel E). MCF-7/pLXSN cells were at least 1,000-fold more resistant to the Bcl-2 inhibitor than either MCF-7/NGAL or MCF-7/Dox^R cells. The MCF-7/Dox^R line was also highly sensitive to the ABT-737 (panel D) but not to doxorubicin (panel E). Treatment with ABT-737 did significantly diminish the concentration of doxorubicin required to reach the IC₅₀ in pLXSN- or NGAL-expressing cells (11- and 3.3-fold, respectively). Moreover, the IC₅₀ for MCF-7/Dox^R cells was considerably reduced (200-fold) when the Bcl-2 inhibitor was added (panel F).

In HT29 CRC cells, the effects of NGAL overexpression did not significantly alter the sensitivity of HT29/NGAL cells to either the Bcl-2 (panel G) or doxorubicin (panel H). However, a suboptimal dose of the Bcl-2 inhibitor did reduce the doxorubicin IC₅₀ for both HT29/LXSN and HT29/NGAL cells greater than 30-fold (panel I).

Effects of enforced NGAL expression on sensitivity to the mTORC1 inhibitor rapamycin

Breast epithelial, breast cancer and CRC cell lines were exposed to varying concentrations of the mTORC1 inhibitor rapamycin (panels A, D and G) and varying concentrations of doxorubicin combined with a constant dose of rapamycin (panels C, F and I) to determine the effects of NGAL expression. Once again, the results with the doxorubicin titrations are presented (panels B, E and H) to allow for direct comparison of treated cells.

MCF-10A cells treated with rapamycin are presented in Figure 3 (panels A and C). Ectopic NGAL expression did not alter the IC₅₀ of MCF-10A/NGAL cells to the mTORC1 inhibitor (panel A), although both MCF-10A/pLXSN and MCF-10A/NGAL cells were very sensitive to the mTORC1 inhibitor, with the IC₅₀ for rapamycin at approximately 0.2 nM. Furthermore, no additive effects in sensitivity were seen with the addition of rapamycin when combined with doxorubicin (panel C).

Figure 3. Sensitivity of NGAL- and pLXSN-infected cells and doxorubicin-resistant MCF-7/Dox^R cells to the mTORC1 inhibitor rapamycin, doxorubicin and the combination of doxorubicin and a constant dose of 5 nM rapamycin. Cells were collected and seeded (2,000 cells/well) in 96-well plates. The following day, serial 2-fold dilutions of rapamycin (A, D and G), doxorubicin (B, E and H) or serial 2-fold dilutions of doxorubicin and a constant dose of 5 nM rapamycin (C, F and I) were added to the wells. Four days later, MTT assays were performed. (A–C) MCF-10A/pLXSN (solid squares), MCF-10A/NGAL (solid upright triangles), (D–F) MCF-7/pLXSN (solid squares), MCF-7/NGAL (solid upright triangles) and 25 nM doxorubicin-selected MCF-7/Dox^R cells (solid downward triangles), (G–I) HT-29/pLXSN (solid squares), HT-29/NGAL (solid upward triangles). A hatched horizontal line is present at the 50% relative growth mark from which the IC₅₀ can be calculated. A vertical arrow indicates the IC₅₀. The statistical significance was determined by the unpaired t-test (***, p < 0.001). All the experiments in this figure were performed at the same time (set up on the same day). These experiments were repeated multiple times and similar results were obtained.

Similar to MCF-10A cells, both MCF-7/pLXSN and MCF-7/NGAL cells were very sensitive to the mTORC1 inhibitor rapamycin (panel D). In contrast though, the doxorubicin-resistant MCF-7/Dox^R cells were highly resistant to rapamycin. The combinatorial effects of different concentrations of doxorubicin and a constant dose of rapamycin were assayed, and as expected, the MCF-7/pLXSN cells were highly sensitive to the constant 5 nM dose of rapamycin in the presence of different concentrations of doxorubicin (panel F). Moreover, the dose of 5 nM rapamycin synergized with doxorubicin and lowered the doxorubicin concentration required to reach the IC₅₀ greater than 10-fold in the doxorubicin-resistant MCF-7/Dox^R cells. Interestingly, overexpression of NGAL appeared to prevent the effects of rapamycin when combined with different concentrations of doxorubicin in MCF-7/NGAL cells (panel F). These results suggest that doxorubicin neutralized the effects of rapamycin in MCF-7/NGAL cells.

Ectopic expression of NGAL did not alter the sensitivity of HT29/NGAL cells to either rapamycin (panel G) or doxorubicin (panel H). HT29 cells were exceedingly resistant to the mTORC1 inhibitor by itself as compared with the breast cell lines. A suboptimal dose of rapamycin was able to reduce the doxorubicin IC₅₀ for HT29/LXSN and HT29/NGAL cells approximately 2-fold (panel I).

Effects of enforced NGAL expression on sensitivity to the CaM-K inhibitor KN-93

The CaM-K inhibitor KN-93 was also used to examine the potential effects NGAL overexpression may have on cell growth in response to chemotherapeutic drugs. Effects of KN-93 (panels A, D and G) and KN-93 in combination with varying concentrations of doxorubicin (panel C, F and I) were analyzed compared with the effects of doxorubicin alone (panels B, E and H).

Figure 4 (panels A, B and C) represents results with the breast epithelial MCF-10A cells. Ectopic NGAL expression appeared to slightly alter the sensitivity of MCF-10A cells to the CaM-K inhibitor, although a clear IC₅₀ could not be established but was estimated to be close to 10,000 nM KN-93 in the MCF-10A/NGAL and greater than 10,000 nM in the MCF-10A/pLXSN cells (panel A). In general, both MCF-10A/NGAL and MCF-10A/LXSN cells were very resistant to the CaM-K inhibitor. These slight effects of KN-93 on MCF-10A/NGAL cells were similarly observed with the addition of varying concentrations of doxorubicin as the IC₅₀s for both MCF-10A/pLXSN and MCF-10A/NGAL cells were marginally decreased (panel C).

Figure 4. Sensitivity of NGAL- and pLXSN-infected cells and doxorubicin-resistant MCF-7/Dox^R cells to the CaM-K inhibitor KN-93, doxorubicin and the combination of doxorubicin and a constant dose of 250 nM KN-93. Cells were collected and seeded (2,000 cells/well) in 96-well plates. The following day, serial 2-fold dilutions of KN-93 (A, D and G), doxorubicin (B, E and H) or serial 2-fold dilutions of doxorubicin and a constant dose of 250 nM KN-93 (C, F and I) were added to the wells. Four days later, MTT assays were performed. (A–C) MCF-10A/pLXSN (solid squares), MCF-10A/NGAL (solid upright triangles), (D–F) MCF-7/pLXSN (solid squares), MCF-7/NGAL (solid upright triangles) and 25 nM doxorubicin-selected MCF-7/Dox^R cells (solid downward triangles), (G–I) HT-29/pLXSN (solid squares), HT-29/NGAL (solid upward triangles). A hatched horizontal line is present at the 50% relative growth mark from which the IC₅₀ can be calculated. A vertical arrow indicates the IC₅₀. The statistical significance was determined by the unpaired t-test (***, p < 0.001). All the experiments in this figure were performed at the same time (set up on the same day). These experiments were repeated multiple times and similar results were obtained.

Both MCF-7/pLXSN and MCF-7/NGAL were very sensitive to the CaM-K inhibitor KN-93 and had IC₅₀s of approximately 110 nM and 12 nM, respectively (Fig. 4, panel D). In contrast, the doxorubicin-resistant MCF-7/Dox^R was more resistant to KN-93, and an IC₅₀ of approximately 2,500 nM was observed. The effects of combining different concentrations of doxorubicin with a constant dose of KN-93 were assessed. The MCF-7/NGAL cells were highly sensitive to the constant dose of KN-93 in the presence of different concentrations of doxorubicin (panel F). Moreover, the dose of 250 nM KN-93 synergized with doxorubicin and lowered the doxorubicin IC₅₀ in both MCF-7/pLXSN and the doxorubicin-resistant MCF-7/Dox^R cells.

As seen in panel G of Figure 4, ectopic NGAL did not significantly alter the sensitivity of HT29/NGAL cells to the CaM-K inhibitor KN-93, and the HT29/pLXSN and HT29/NGAL cells had IC₅₀s of approximately 500 nM and 600 nM, respectively to KN-93. Unlike the results observed with the MCF-7 cell lines, a suboptimal dose of the CaM-K inhibitor did not reduce the doxorubicin IC₅₀ for HT29/LXSN and HT29/NGAL cells (panel I).

Effects of enforced NGAL expression on sensitivity to the natural plant product berberine

The natural plant derivative berberine was also examined in our cell lines, and the results are presented in Figure 5. Berberine affects the regulation of many genes, including suppressing NFκB, ERK and MMP-9 expression while also inducing 5′ adenosine monophosphate-activated protein kinase (AMPK) expression.

Figure 5. Sensitivity of NGAL- and pLXSN-infected cells and doxorubicin-resistant MCF-7/Dox^R cells to the natural product berberine, doxorubicin and the combination of doxorubicin and a constant dose of 250 nM berberine. Cells were collected and seeded (2,000 cells/well) in 96-well plates. The following day, serial 2-fold dilutions of berberine (A, D and G), doxorubicin (B, E and H) or serial 2-fold dilutions of doxorubicin and a constant dose of 250 nM berberine (C, F and I) were added to the wells. Four days later, MTT assays were performed. (A–C) MCF-10A/pLXSN (solid squares), MCF-10A/NGAL (solid upright triangles), (D–F) MCF-7/pLXSN (solid squares), MCF-7/NGAL (solid upright triangles) and 25 nM doxorubicin-selected MCF-7/Dox^R cells (solid downward triangles), (G–I) HT-29/pLXSN (solid squares), HT-29/NGAL (solid upward triangles). A hatched horizontal line is present at the 50% relative growth mark from which the IC₅₀ can be calculated. A vertical arrow indicates the IC₅₀. The statistical significance was determined by the unpaired t-test (***, p < 0.001). All the experiments in this figure were performed at the same time (set up on the same day). These experiments were repeated multiple times and similar results were obtained.

Ectopic NGAL expression did not exhibit any effect on the sensitivity of MCF-10A cells to berberine over the concentrations examined (up to 2,000 nM), as the IC₅₀s were not reached. Although the MCF-10A/NGAL cells did display more sensitivity to berberine at the highest concentrations used, they were not statistically different from the MCF-10A/pLXSN. Likewise, the addition of a constant dose (250 nM) of berberine did not appear to significantly lower the concentration of doxorubicin needed to reach the IC₅₀ of doxorubicin alone in either MCF-10A/pLXSN or MCF-10A/NGAL cells (panel C).

MCF-7/pLXSN, MCF-7/NGAL and MCF-7/Dox^R cells were also treated with berberine and are presented in Figure 5 (panels D, E and F). As seen in panel D, MCF-7/NGAL cells were very sensitive to berberine and had an IC₅₀ of approximately 10 nM. In contrast, MCF-7/pLXSN and the doxorubicin-resistant MCF-7/Dox^R were exceedingly more resistant to berberine and exhibited IC₅₀s of approximately 2,000 nM and greater than 2,000 nM, respectively. The combinatorial effects of different concentrations of doxorubicin and a constant dose of berberine were determined and yielded some unexpected results. Although MCF-7/pLXSN and MCF-7/NGAL cells did not demonstrate a significant difference in their doxorubicin IC₅₀s, the sensitivities of MCF-7/NGAL cells to a constant dose of berberine (250 nM) in the presence of different concentrations of doxorubicin were increased approximately 10-fold, while addition of a constant dose of berberine did not increase the sensitivity of MCF-7/pLXSN to doxorubicin (panel F). Moreover, berberine synergized with doxorubicin to dramatically lower the doxorubicin IC₅₀ in the doxorubicin-resistant MCF-7/Dox^R cells.

Similar to the results observed with the MCF-10A cell line, ectopic NGAL expression had a mild effect on the sensitivity of HT29 cells to berberine (Fig. 5, panels G, H and I). HT29/pLXSN cells were slightly more resistant to berberine than HT29/NGAL cells, with both having an IC₅₀ of greater than 2,000 nM (panel G). A suboptimal dose of berberine did not produce a measurable reduction in the doxorubicin IC₅₀s for HT29/LXSN and HT29/NGAL cells (panel I).

Inhibitor and doxorubicin IC₅₀s for various NGAL and pLXSN retrovirally infected cell lines as well as the combination effects of doxorubicin with a constant dose of each inhibitor are presented in Table 1. The fold difference in sensitivities of the MCF-7/pLXSN, MCF-7/NGAL and MCF-7/Dox^R cell lines to the various inhibitors are presented in Table 2. In this table, the values obtained with MCF-7/NGAL and MCF-7/Dox^R are normalized to the empty vector-infected, doxorubicin-sensitive control line, MCF-7/pLXSN.

Table 2. Fold differences in sensitivity to small-molecule inhibitors, berberine and doxorubicin in MCF-7/pLXSN, MCF-7/NGAL and MCF-7/Dox^R cells.¹

Drug treatment	MCF-7/NGAL	MCF-7/Dox^R
AG1478 (EGFR Inh)	67 X↓	333 X↓
ABT-737 (Bcl-2 Inh)	1,000 X↓	667 X↓
Rapamycin (mTORC1)	1.3 X ↑	667 X↑
KN-93 (CaM-K)	9.2 X↓	23 X↑
Berberine (AMPK, others)	166 X↓	—
Doxorubicin (topoisomerase, others)	1.3 X↓	25 X↑
Dox + 250 nM AG1458	—	6.7 X↑
Dox + 50 nM ABT-737	2 X↑	1.1 X↑
Dox + 5 nM Rapa	47 X↑	1.7 X↑
Dox + 250 nM KN-93	6 X↑	2.5 X↑
Dox + 250 nM Berberine	7 X↓	8.8 X↓

¹Fold change in IC50s were normalized to the IC50s detected in MCF-7/pLXSN cells. ↓ indicates a decrease in IC50 in comparison to MCF-7/pLXSN control cells. ↑ indicates an increase in the IC50 compared with MCF-7/pLXSN control cells. Estimated data values obtained from drug titrations derived from graphs in Figures 1–5 and listed in Table 1.

Discussion

These studies were undertaken to determine whether increased NGAL expression altered the sensitivity of breast and CRC cells to certain small-molecule signal transduction inhibitors and the natural plant product berberine. The effects of the inhibitors and natural product were also examined in combination with the chemotherapeutic drug doxorubicin in cancer cell types that are normally sensitive to doxorubicin (breast) and cancer cell types which are normally resistant to doxorubicin (CRC) as well as immortalized breast epithelial cells (MCF-10A), which are not malignant. NGAL may have roles in iron transport, which may be associated with chemosensitivity in certain cancers. Some studies have shown that iron chelators will reduce chemotherapeutic drug resistance.^99,100 NGAL expression has been associated with a poor prognosis in breast cancer. Importantly, our studies document that elevated NGAL expression altered the sensitivity to certain small-molecule inhibitors and the natural plant product berberine, especially in the MCF-7 breast cancer cell line. However, elevated NGAL expression had fewer effects on the immortalized breast epithelial MCF-10A cells and essentially no effects on the HT-29 CRC cell line.

In MCF-7/NGAL cells, it is likely that NGAL is altering either the uptake or efflux of some of the inhibitors and berberine into the cells. This is a novel, but logical function for NGAL. The expression of NGAL increased the toxicity of the EGFR, Bcl-2, CaM-K and berberine. In contrast, elevated NGAL expression antagonized the effects of the mTORC1 inhibitor rapamycin when combined with doxorubicin in MCF-7/NGAL cells. These are important results, as some cancer patients are treated with EGFR, Bcl-2 and mTORC1 inhibitors in combination with chemotherapeutic drugs. Interestingly, our results also identified that the sensitivity of the CRC line HT-29 to doxorubicin could be increased when co-treated with the Bcl-2 inhibitor ABT-737.

We included in our studies the doxorubicin-resistant derivative of the MCF-7 cell line, MCF-7/Dox^R, to serve as a control for determining the effects of doxorubicin and the small-molecule inhibitors. Drug resistant breast cancer cells often express higher levels of proteins involved in drug transport. The doxorubicin-resistance of MCF-7/Dox^R cells was eliminated by EGFR, Bcl-2, CaM-K and mTORC1 inhibitors and the natural plant product berberine.

The doxorubicin-resistant MCF-7/Dox^R cell line was also infected with the retrovirus encoding NGAL;¹¹⁵ however, these cells were not more resistant or sensitive to doxorubicin than the pLXSN empty vector control MCF-7/Dox^R cells (data not presented). NGAL expression was not detected at higher levels in drug-resistant MCF-7/Dox^R/pLXSN cells than MCF-7/pLXSN.¹¹⁵ Moreover, the doxorubicin-resistant MCF-7/Dox^R cells do not normally express NGAL; however, upon treatment with doxorubicin, increased NGAL protein has been detected.¹¹⁴ Elevated NGAL expression does not appear, by itself, to alter the sensitivity to doxorubicin in the cells examined, and the doxorubicin-resistance of MCF-7/Dox^R cells could result from aberrant regulation of various signaling pathways or drug transporters in these cells.

The MCF-7/Dox^R cells were very sensitive to the EGFR inhibitor AG1478 and the Bcl-2 inhibitor ABT-737. It is likely that the drug resistance present in MCF-7/Dox^R is hypersensitive to EGFR pathway activation and Bcl-2 survival signaling. In contrast, MCF-7/Dox^R cells were resistant to mTORC1 inhibition by itself, but the doxorubicin-resistance present in MCF-7/Dox^R was eliminated when rapamycin was combined with doxorubicin in our assays. On the other hand, while MCF-7/NGAL cells were highly sensitive to rapamycin, the effects of rapamycin were eliminated when combined with doxorubicin. These results suggest that NGAL has different effects when certain small-molecule inhibitors are combined with chemotherapeutic drugs and could have clinical implications, as many cancer patients express elevated levels of NGAL.^71-76,116

The doxorubicin-resistance of MCF-7/Dox^R cells was eliminated when the cells were treated with the CaM-K inhibitor KN-93 and doxorubicin, as they exhibited a similar IC₅₀ for doxorubicin as the drug-sensitive MCF-7/pLXSN cells. These results suggest a key role for the CaM-K pathway in the doxorubicin resistance of MCF-7/Dox^R cells. The MCF-7/NGAL cells were also very sensitive to the CaM-K inhibitor. The CaM-K pathway is important in the regulation of the Ras/Raf/MEK/ERK pathway,¹¹⁷ the regulation of NFκB¹¹⁸ and AMPK.¹¹⁹ Also, the CaM-K pathway is activated after doxorubicin treatment and after other treatments which induce reactive oxygen species (ROS).^120-122 A diagram illustrating where some of these drugs interact in signaling pathways is presented in Figure 6.

Figure 6. Overview of targeting of key pathways and effects on NGAL expression. Activation of many signaling cascades can occur after activation of the EGFR receptor or by treatment with doxorubicin, which induces reactive oxygen species (ROS) and the CaM-K cascade. The sites where certain signal transduction inhibitors (black rectangles) and the natural product berberine and the diabetes drug metformin are indicated. Berberine may activate AMPK (red rectangle) as well as inhibit NFκB (black rectangle). NGAL expression may alter the transport and efflux of certain small-molecule inhibitors, which, in some cases, may be deleterious to the cancer cell.

NGAL expression has been associated with a poor prognosis in breast and other cancers.^80,116 While elevated expression of NGAL does not alter the IC₅₀ for the chemotherapeutic drug doxorubicin in the cell lines examined, it did alter their sensitivity to certain small-molecule inhibitors in the breast cancer cell line MCF-7. Furthermore, elevated expression of NGAL did not appear to alter the responses of either the CRC line HT-29 or the immortalized epithelial line MCF-10A to doxorubicin.

Targeting the EGFR/Ras/PI3K/Akt/mTORC1 pathway is a key anticancer and anti-aging approach. There are many sites which are frequently mutated or aberrantly expressed and are being targeted in this pathway, from upstream receptors (e.g., EGFR, HER2)^{25,27,28,123,124} to downstream signaling proteins such as Ras,^29,125,126 PI3K,^{26,30,127-138} PTEN,^43,139,140 Akt,^25-28 TSC1/TSC2,^44-47,144 Rheb,^25-28,145 mTOR^{25-28,146-165} and p70S6K.^25-28,166 This pathway also plays important roles in cell growth and is often aberrantly regulated in diabetes and obesity.¹⁵¹ Many anti-diabetes drugs such as metformin and the traditional drug berebrine interfere with components that feed into this pathway or are regulated by this cascade,e and there is cross-talk between these pathways. It is interesting that rapamycin and berberine both had effects on NGAL-expressing cells. NGAL suppressed the effects that rapamycin had on lowering the doxorubicin IC₅₀ in MCF-7/NGAL cells, while MCF-7/NGAL cells were very sensitive to the effects of berberine.

mTORC1 phosphorylates unc-51-like kinase 1 (ULK1), which results in the suppression of autophagy. The mTORC1 inhibitor rapamycin prevents phosphorylation of ULK1, and autophagy can occur.^167-169 It appears that targeting key molecules which control energy may be approaches to control cancer and aging. Drugs such as rapamycin target mTORC1, and metformin is an indirect inhibitor of mTORC1. Metformin induces AMPK which activates TSC1 and subsequently suppresses mTORC1 activity.^170-186 Berberine may elicit similar effects.^187-189 Metformin may also induce the phosphorylation and inactivation of Raptor,¹⁹⁰ a key regulatory component in the mTORC1 complex which is critical for the translation of many weak “oncogenic” mRNAs important in proliferation. Interestingly, diabetics treated with metformin have lower incidences of cancer and also do not exhibit as much aging¹⁹¹ and mice treated with metformin do not exhibit as much aging as untreated mice.¹⁸³ Moreover, metformin may be able to prevent the survival of certain CICs.

Enhanced glycolysis (Warburg effect) is critical for CIC survival.^192-197 Metformin disrupts the glycolytic metabotype and alters the ATM-mediated DNA damage response resulting in the acceleration of stress-induced senescence. Metformin in the presence of suppressed mTOR signaling slows down aging and alters the cellular senescence processes. Hence, metformin can alter the ability of cells to become immortalized into CICs and slows down cellular aging. By reducing the levels of DNA damage signaling, metformin has genoprotective effects¹⁹⁸ and anticancer/tumor suppressor effects.^199-202 Berberine may have similar effects on AMPK. Interestingly, MCF-7 cells that overexpressed NGAL were hypersensitive to berberine.

Materials and Methods

Cell lines and growth factors

Breast cancer cells lines (MCF-7) and the CRC line HT-29 were obtained from the ATCC. Cells were maintained in a humidified 5% CO₂ incubator at 37°C with RPMI-1640 (RPMI) (Invitrogen) supplemented with 10% fetal bovine serum (FBS) (Atlanta Biologicals). This complete RPMI media is abbreviated cRPMI. The immortalized breast epithelial MCF-10A line was obtained from the ATCC and cultured in DMEM/F12 (Invitrogen) medium containing: 2.5 mM L-glutamine, supplemented with 5% heat-inactivated equine serum (Invitrogen), 500 ng/ml hydrocortisone (Sigma-Aldrich), 21.5 ng/ml epidermal growth factor (EGF) (Sigma-Aldrich), 10 μg/ml insulin (Sigma-Aldrich), 100 ng/ml cholera toxin (Sigma-Aldrich) and 15 mM HEPES (Sigma-Aldrich). AG1478 (EGFR inhibitor), KN-93 (CaM-K inhibitor), rapamycin (mTORC1 inhibitor), doxorubicin and berberine were purchased from Sigma-Aldrich. ABT-737 (BCL-2 inhibitor) was obtained from Dr. Michael Andreeff (MD Anderson Cancer Center).

Methylthiazol tetrazolium assay

Methylthiazol tetrazolium (MTT) assays were performed to determine a cell line’s sensitivity to chemotherapeutic drugs. Two thousand cells per well were plated in 96-well plates in 100 μL of cRPMI without phenol red (Invitrogen) and allowed to attach overnight under normal culture conditions. The next day, serial 2-fold dilutions of a chemotherapeutic drug were made, and 100 μL of each dilution were added to a corresponding well on the 96-well plate. Cells were incubated for 4 d under normal culture conditions. On the fourth day, 22.2 μL of a 5 mg/mL solution of thiazolyl blue tetrazolium bromide (Sigma-Aldrich) in 1X PBS was added to each well and incubated for 90 min at 37°C. The media was then removed, and 150 μL of dimethyl sulfoxide (DMSO) (Fisher Scientific) was added to resuspend formazin crystals to produce a purple color, which was subsequently read on a Multiskan EX Microplate photometer (Thermo Scientific) at a wavelength of 570 nm. Colormetric readings were normalized against plates of non-treated cells under identical culture conditions. Relative growth was calculated by dividing normalized cell growth values in the presence of drugs by normalized cell growth values in the absence of drugs, and the results were graphed. Drug concentrations that killed at least 50% of the cells (IC₅₀) were determined from the calculated graphed values.^114,115

NGAL retroviral transduced cells

The construction of the pLXSN/NGAL retroviral expression vector and the infection of cells with this retrovirus have been previously described.¹¹⁵

Acknowledgments

We thank Dr. Michael Andreeff for generously providing the Bcl-2 inhibitor, ABT-737. M.C. and G.M. were supported in part by grants from the Italian “Ministero dell’Istruzione, dell’Università e della Ricerca” (Ministry forEducation, Universities and Research) – MIUR PRIN 2008 and FIRB-MERIT (RBNE08YYBM). M.C. was also supported in part by a grant to the CNR from the Italian Ministry of Economy and Finance for the Project FaReBio di Qualità. M.L. was supported in part by a grant from the Italian Ministry of Health, Ricerca Finalizzata Stemness 2008 entitled “Molecular Determinants of Stemness and Mesenchymal Phenotype in Breast Cancer.” A.M.M. was supported in part by grants from: MIUR PRIN 2008 (2008THTNLC) and MIUR FIRB 2010 (RBAP10447J-003) and 2011 (RBAP11ZJFA_001). M.M. was supported in part from the Italian Association for Cancer Research (AIRC), the Cariplo Foundation and the Italian Ministry of Health. J.P. and R.T. were supported in part by a grant from Fondazione Umberto Veronesi entitled “Neutrophil gelatinase-associated lipocalin (NGAL) and matrix metalloproteinases (MMPS) as biomarkers of bladder cancer development and progression.” A.T. was supported in part by grants from the Italian “Ministero dell’Istruzione, dell’Università e della Ricerca” (Ministry for Education, University and Research) - MIUR - PRIN 2008 and grant from “Sapienza,” University of Rome 2009–11.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.