Resistance to cytotoxic agents has long been known to be a major limitation in the treatment of human cancers. Although many mechanisms of drug resistance have been identified, chemotherapies targeting known mechanisms have failed to lead to effective reversal of drug resistance, suggesting that alternative mechanisms remain undiscovered. Previous work identified midkine (MK) as a novel putative survival molecule responsible for cytoprotective signaling between drug-resistant and drug-sensitive neuroblastoma, osteosarcoma and breast carcinoma cells
Drug resistance poses a major obstacle in the treatment of human cancers. Several mechanisms responsible for resistance to chemotherapy have previously been described. Drug-resistant cells may express one or more energy-dependent transporters like the multidrug resistance gene (
Our laboratory has hypothesized a cytoprotective relationship between drug-resistant and drug-sensitive cells within drug-resistant tumors. We previously demonstrated a mechanism of cytoprotective signaling between drug-resistant and drug-sensitive neuroblastoma, osteosarcoma, and breast cancer cells [
MK is a retinoic acid-induced neurotrophic factor. It is a small cysteine-rich protein which belongs to the family of heparin-binding proteins and is highly expressed during midgestation in mouse embryogenesis. MK is integral to neuronal development, migration, and neurite outgrowth [
Recently, several other studies have implicated MK in drug resistance. Kang et al. used microarray analysis to assess global gene expression in gastric cancer cell lines with acquired drug resistance to 5-fluorouracil, doxorubicin, and cisplatin (CDDP) [
In the present study, the relationship between MK expression and drug resistance has been further investigated. Special emphasis was placed on determining whether the cytoprotective action of MK observed
All materials utilized in this study were purchased from the following companies: Dulbecco’s modified Eagle’s medium (DMEM) (Mediatech, Inc., Pittsburgh, PA); fetal bovine serum (FBS) (HyClone, Logan, UT); doxorubicin; 3-(4,5-dimethyl-2-thiazolyl)2,5-diphenyl tetrazolium bromide (MTT) (Sigma, St Louis, MO); antibodies to midkine (R&D Systems, Minneapolis, MN); antibody to
Human neuroblastoma SK-N-SH and osteosarcoma SJSA-1 (OSA) cells were purchased from ATCC (Rockville, MA) and grown in DMEM supplemented with 10% FBS at 37°C in 5% CO2 atmosphere. The IC50 for SK-N-SH wild type (SK-N-SH WT) cells was determined using the MTT assay. The generation of doxorubicin resistant SK-N-SH cells (SK-N-SH DoxR) was achieved by incubating parental cell lines with incrementally increasing concentrations of doxorubicin ranging from 10−9 to 10−6 M over a period of 6 months. Cells were deemed resistant after surviving 10 passages in a doxorubicin concentration of 10−6 M, approximately 2 Log above the IC50 of the parental cell line. Cells were then continuously selected in DMEM containing doxorubicin 10−6 M. SK-N-SH GFP cells (SK-N-SH GFP-WT) were generated using pLPCX-EGFP retrovirus and selected with puromycin. Human MK overexpressed SK-N-SH cells (SK-N-SH HMK) were derived as previously reported [
The cytoprotective effect of SK-N-SH DoxR cells on SK-N-SH WT cells was quantitatively evaluated using a co-culture system. The Falcon Cell Culture Insert System (BD Falcon, Franklin Lakes, NJ) was used to allow SK-N-SH WT cells to be co-cultured with either SK-N-SH WT or SK-N-SH DoxR cells. For each co-culture setup, the two cell types were grown in separate compartments without physical contact. Small molecules were able to diffuse through the media between compartments by traversing a 0.4
Whole cell lysates, culture medium, or cell membranes were isolated and subjected to electrophoresis. Whole cell lysates were isolated from cells grown to 80% confluence in 25 cm2 flasks. Cells were collected and lysed in a lysis buffer composed of 50 mM HEPES pH 7.4, 150 mM NaCl, 100 mM NaF, 1 mM MgCl2, 1.5 mM EGTA, 10% glycerol, 1% Triton X100, 1
Protein concentrations were determined for whole cell lysates and cell membrane fractions, and equal quantities of protein were separated by electrophoresis on a 4–20% SDS-PAGE gel and transferred to nitrocellulose membranes. The volume of medium loaded on the gel was normalized based on cell count in the flask prior to medium collection. Nitrocellulose membranes were incubated with antibodies for proteins of interest (anti-MK, anti-P-glycoprotein or
Human MK siRNA was synthesized by Dharmacon (Lafayette, CO). The nucleotide sequences for the human MK genes were obtained from the NCBI sequence viewer program. The human MK siRNA was based on sequence 229 to 249 and consisted of the following nucleotides: 5′-AAGAAGGAGTTTGGAGCCGAG-3′. On the day before transfection, 3 × 105 cells were seeded into 6-well plates and grown in 2.5 mL of DMEM supplemented with 10% fetal bovine serum. After 24 hours in culture, 25
Cell proliferation was assessed using the MTT assay. Cells were seeded in 96 well plates and incubated with logarithmic concentrations of doxorubicin, ranging from 10−9 to 10−5 M. Cell proliferation was quantitatively estimated by use of a colorimetric assay using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT). MTT (10
The animal protocol used in this study was approved by the Animal Care and Use Committee of the Lurie Children’s Hospital of Chicago Research Center (no. 2006-29). Severe combined immunodeficiency (SCID) or nude mice (Charles River Laboratories, Wilmington, MA), approximately 4–6 weeks of age and weighing approximately 30 g received subcutaneous (SC) tumor implants performed using various ratios of GFP expressing SK-N-SH wild type cells (SK-N-SH GFP-WT) and SK-N-SH doxorubicin drug-resistant cells (SK-N-SH DoxR) or SK-N-SH HMK (SK-N-SH HMK) cells with a total of 106 cells in 100
When tumors were palpable, the animals were challenged with doxorubicin (2.5 mg/kg). A total of three doxorubicin injections, each separated by 3 days, were performed. Mice were weighed and checked for clinical signs of drug toxicity and lethality. Tumor measurements were made with a caliper three times weekly for 3 to 4 weeks and converted to tumor volume by using the formula
Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) allows detection of genomic DNA cleavage which occurs during apoptosis by incorporating fluorescein labels in nucleotide polymers. TUNEL was used to detect apoptotic cells amongst SK-N-SH GFP-WT cells grown in co-culture with SK-N-SH WT, DoxR, or HMK cells. Labeling was quantified using fluorescence microscopy (
After obtaining Internal Review Board approval (IRB# 2010-14080), pediatric patients treated for neuroblastoma from January 1999 to December 2008 at a free-standing, tertiary care children’s hospital were retrospectively identified. Only patients who completed therapy before December 2008 and who no longer require biopsy tissue for diagnostic purposes were included in the study. Medical records were reviewed for the following clinical features: age at diagnosis, pathology features (favorable versus unfavorable), N-MYC amplification status, tumor stage, and outcome (mortality). Paraffin embedded tissue samples were collected from the pathology bank, and slides were created.
Formalin-fixed paraffin embedded slides were dewaxed in xylene and hydrated through a graded series of alcohols. Endogenous peroxidases were blocked with a 3% hydrogen peroxide treatment; antigen retrieval was performed by boiling for 20 minutes in a 0.01 M sodium citrate (pH 6) solution, and endogenous biotin blocked using Avidin/Biotin Blocking Kit (Vector Labs, SP-2001). Slides were incubated overnight in the primary antibody for MK (1 : 250 dilution). Following incubation with the appropriate biotin-labeled secondary antibodies, the labeled antigens were visualized by streptavidin-biotin (Vectastain Elite ABC kit; Vector Laboratories) followed by ImmunoPure Metal Enhanced DAB Substrate (Thermo Scientific) and counterstained with hematoxylin (Richard-Allen Scientific).
Cytoplasmic and nuclear staining were scored on a scale of none (0), low (1+), medium (2+), and high (3+) on a blinded basis. When both prechemotherapy and postchemotherapy biopsy specimens were available, expression of MK was compared to determine if MK expression changes after treatment in tumor cells which remain viable despite chemotherapy. Expression of MK was also compared between patients by age at diagnosis (<1 year versus >1 year), stage, N-MYC amplification status (amplified versus nonamplified), histology (favorable versus unfavorable), and survival status.
Data are expressed as means ± SE. Differences in measured variables between the experimental and control groups were assessed using Student’s
Earlier studies revealed that MK was able to exert a survival function in a variety of cellular systems and against various stimuli, suggesting that it could be associated with drug resistance [
Co-culture effect on cellular response to doxorubicin. Wild type human neuroblastoma (SK-N-SH WT) cells co-cultured with SK-N-SH WT cells (WT/WT) were compared to co-cultures of SK-N-SH WT cells with doxorubicin resistant cells (SK-N-SH DoxR) (WT/DoxR). Co-cultures were treated with and without doxorubicin at 10−7 or 10−6 M for 48 hours. Surviving cells were quantified using trypan blue staining. The cell survival ratio represents the number of live WT cells in the drug-treated co-culture divided by the number live cells in the untreated (control), *
These findings demonstrate that ligands secreted by doxorubicin resistant human neuroblastoma cells impart a cytoprotective effect on drug-sensitive cells, providing further support for the humoral mediation of resistance to chemotherapy. In order to confirm that MK was the molecule responsible for this protective effect, we created a MK over expressing cell line (SK-N-SH HMK) as previously described [
Effect of midkine overexpression on wild type cell survival. A midkine overexpressing SK-N-SH cell line (SK-N-SH HMK) was created as previously described [
The
Previous co-culture experiments using GFP-transfected wild type SK-N-SH cells (GFP-WT) co-cultured with either SK-N-SH WT cells or DoxR cells confirmed the cytoprotective effect of SK-N-SH DoxR cells suggesting humoral protection. In order to confirm that MK was the molecule responsible for this effect, these co-culture experiments were repeated using the MK overexpressing cell line (SK-N-SH HMK). Co-cultures were grown using SK-N-SH GFP-WT cells in co-culture with either GFP-WT cells, doxorubicin resistant cells (SK-N-SH DoxR), or human MK overexpressing SK-N-SH cells (SK-N-SH HMK). Cells were grown in co-culture for 48 hours then incubated with (+) or without (−) doxorubicin (10−7 M) for 48 hours. Inserted SK-N-SH WT cells were then stained using Hoechest 33342 to determine cell viability. As seen in Figures
Effect of midkine overexpression on cellular response to doxorubicin in co-culture conditions. (a) GFP-transfected wildtype SK-N-SH cells (GFP-WT) were co-cultured with wild type (WT), doxorubicin resistant (DoxR), or human midkine overexpressing SK-N-SH cells (HMK), incubated with (+) or without (−) doxorubicin at 10−7 M for 48 hours. Photographs were taken using fluorescence microscopy. (b) The viability percentage of fluorescent cells (GFP-WT) after each treatment was determined through Hoechst 33342 staining and cell counting. Data represent an average of three independent determinations +/− SE.
To further investigate the cytoprotective function of MK, we set out to determine if inhibition of MK secretion reversed or reduced the cytoprotective action exerted by doxorubicin resistant neuroblastoma cells upon wild type, drug-sensitive cells. siRNA to MK was used to knock down MK expression in doxorubicin resistant cells. ELISA (Figure
Effect of midkine siRNA on cytoprotection. (a) ELISA assay and (b) western blot were used to confirm decreased expression of midkine in doxorubicin resistant SK-N-SH cells (DoxR) treated with midkine. Culture medium from SK-N-SH wild type (WT), DoxR, DoxR cells treated with scramble sequence siRNA (DoxR-scramble), and DoxR cells treated with siRNA to midkine (DoxR-si-midkine) was harvested after growth for 96 hours. (c) SK-N-SH WT cells were grown in co-culture with WT, DoxR, and DoxR-scramble or DoxR-si-midkine cells. Co-cultures were incubated for 48 hours with or without doxorubicin at 10−7 M and 10−6 M. SK-N-SH WT cell survival was then quantified through cell counting after staining with trypan blue. Data represents the average of 4 experiments +/− SE. (d) OSA WT cells were grown in co-culture with OSA WT, DoxR, and DoxR-scramble or DoxR-si-midkine cells. Co-cultures were incubated for 48 hours with or without doxorubicin at 10−7 M and 10−6 M. OSA WT cell survival was then quantified through cell counting after staining with trypan blue. Data represents the average of 4 experiments +/− SE.
SK-N-SH WT cells (Figure
We have shown that induced expression of MK in wild type SK-N-SH cells confers a resistance to doxorubicin. We sought to determine if MK plays a self-protective effect in doxorubicin resistant cells themselves. We used MK siRNA to knock down MK expression in doxorubicin resistant (SK-N-SH DoxR) cells. Cells were either untreated or treated with doxorubicin (10−7 and 10−6 for 24 hr) (Figure
Effect of midkine siRNA on doxorubicin resistant cellular response to doxorubicin. siRNA was used to knock down midkine expression in doxorubicin resistant SK-N-SH cells (DoxR) to determine if loss of midkine expression results in restoration of drug sensitivity in the DoxR cells. Wild type (WT), DoxR, DoxR cells treated with scramble sequence RNA (DoxR-scramble) and SK-N-SH DoxR cells treated with siRNA to midkine (DoxR-si-midkine) were cultured with or without doxorubicin at 10−7 M and 10−6 M for 24 hours. Cell survival was assayed using trypan blue and cell counting. Data represents the average of 4 experiments +/− SE.
To prove that drug-resistant cells confer a humoral mediated cytoprotective effect on drug-sensitive cells
Effect of midkine cytoprotection on growth of wild type human neuroblastoma tumors
MK expression has previously been implicated as a poor prognostic indicator in oral squamous cell carcinoma [
Correlation of midkine staining in human tissues with prognostic indicators.
Nuclear score | Cytoplasmic score | |||||||
---|---|---|---|---|---|---|---|---|
Prechemo |
|
Postchemo |
|
Prechemo |
|
Postchemo |
| |
Stage | 0.46 | 0.16 | 0.23 | 0.22 | ||||
I |
|
|
|
0 | ||||
II |
|
|
|
|
||||
III |
|
|
|
|
||||
IV |
|
|
|
|
||||
N-MYC Status | 0.86 | 0.41 | 0.48 | 0.26 | ||||
Nonamplified |
|
|
|
|
||||
Amplified |
|
|
|
|
||||
Histology | 0.49 | 0.34 | 0.77 | 0.51 | ||||
Favorable |
|
|
|
|
||||
Unfavorable |
|
|
|
|
||||
Age | 0.17 | 0.20 | 0.22 | 0.90 | ||||
<1 year |
|
|
|
|
||||
>1 year |
|
|
|
|
||||
Survival status | 0.64 | 0.79 | 0.33 | 1 | ||||
Survivor |
|
|
|
|
||||
Nonsurvivor |
|
|
|
|
Midkine expression prechemotherapy compared to postchemotherapy.
Nuclear Score | Cytoplasmic score | |||||
---|---|---|---|---|---|---|
Pre-chemo | Post-chemo |
|
Pre-chemo | Post-chemo |
| |
All biopsies* |
|
|
0.15 |
|
|
0.005 |
Matched biopsies** |
|
|
0.74 |
|
|
<0.001 |
**Comparison based only on the 10 patients with matched pre- and postchemotherapy biopsies available for analysis.
Immunohistochemistry for midkine in human neuroblastoma tissue samples. Prechemotherapy and postchemotherapy patient neuroblastoma biopsy samples were collected and stained for midkine. Midkine staining was scored using a 0 (negative), 1+ (weak staining), 2+ (moderate staining), and 3+ (strong staining) scale on a blinded basis. Midkine staining was then correlated with tumor stage, N-MYC amplification status, histology (favorable or unfavorable), and survival. Biopsy scores were also evaluated for a change in midkine staining prechemotherapy to postchemotherapy. Shown here is a sample neuroblastoma histology slide stained for midkine.
In 1988, the Muramatsu group first described midkine as a 13–15 KD heparin binding polypeptide present in embryonal carcinoma cells which acts to enhance neuronal cell survival and stimulate neurite extrusion [
Our laboratory has demonstrated that an intercellular crosstalk occurs between cells that have acquired a drug resistance phenotype and neighboring drug-sensitive tumor cells [
Our
Midkine expression is usually low in normal human adult tissues. Wilms’ tumor, neuroblastoma, esophageal, pancreatic, lung, and breast cancers have all been shown to have enriched MK expression [
The exact mechanism by which midkine mediates drug resistance remains poorly understood, but it is postulated that midkine has anti-apoptotic (or antiautophagy) [
Perhaps the most intriguing questions regarding the anti-apoptotic activity of midkine and other heparin binding growth factors reside in determining potential sites of action in the cell. Because midkine is secreted from drug-resistant cells, the neighboring drug-sensitive cells must express a receptor for midkine as well as the related signaling pathway. By virtue of its apparent lack of specificity, midkine may bind nonspecifically to and activate various receptors, leading to activation of more than one survival pathway and thus increasing its chances to mediate cytoprotection. At least four receptors, namely, N-syndecan [
In summary, the evidence presented in this study confirms the existence of an intercellular cytoprotective signal within drug-resistant tumors which allows phenotypically drug-resistant tumor cells to protect nearby otherwise drug sensitive cells. This humoral effect appears to be mediated primarily by midkine, allowing protection of neighboring cells from drug-induced killing. Future studies identifying the pertinent midkine receptors and downstream signaling pathways may identify novel tumor markers or therapeutic targets.
The authors do not have a direct financial relationship with the commercial entities mentioned in the paper which may lead to conflict of interests.
Fei Chu and Jessica A. Naiditch contributed equally to this work.
Support for this work was provided by the John W. Anderson Foundation, the Children’s Surgical Foundation of the Ann & Robert H. Lurie Children’s Hospital of Chicago, the Super Jake Foundation, and the Mark Staehely Pediatric Cancer Foundation. This paper is dedicated to the memory of Dr. Bernard L. Mirkin.