Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. 2003 Nov 25; 100(24): 13988–13993.
Published online 2003 Nov 17. doi:  10.1073/pnas.2335921100
PMCID: PMC283533
Cell Biology

Atypical methylation of the interleukin-8 gene correlates strongly with the metastatic potential of breast carcinoma cells


Previously, we have shown that a strong correlation exists between the metastatic potential of breast carcinoma cell lines and their ectopic expression of IL-8. The undifferentiated, highly metastatic cell lines with high metastatic potential produce much more IL-8 than their differentiated lower metastatic counterparts. After eliminating the possibility that transcription factor activity was responsible for differences in IL-8 release, we examined the IL-8 gene for possible epigenetic modifications. Here, we report an aberrant methylation pattern that may be responsible for the differences in IL-8 release between the high and low metastatic cell lines. We determined that none of the deoxycytidylate-phosphate-deoxyguanylate (CpG) sites in the reported IL-8 promoter were methylated in either cell type. Much further upstream in the IL-8 gene, two CpG sites were identified that are differentially methylated. These two sites were fully methylated in the high metastatic cell lines, which produce large quantities of IL-8 and remain unmethylated in the low metastatic cell lines where the IL-8 gene is relatively silent. The DNA methylation results presented here differ from the common epigenetic paradigm in which methylation of promoter CpG islands silences gene expression, suggesting that there are additional epigenetic control mechanisms that as yet have not been fully appreciated or explored.

Interleukin-8 (IL-8) is a small basic protein that was purified as a neutrophil chemoattractant (1, 2). It has been reported that, besides being a strong neutrophil chemoattractant, IL-8 possesses potent angiogenic properties. These properties are likely to contribute to the metastatic potential of tumor cells ectopically releasing substantial quantities of IL-8 (3).

Ectopic production and release of IL-8 enhances the metastatic potential of several tumor types. The metastatic potentials of clones derived from a melanoma tumor cell line increase as their levels of ectopic IL-8 increase (46). Passive immunization of severe combined immune deficient (SCID) mice against human IL-8 attenuates the growth and angiogenic response of the human lung tumor line A549, which produces high levels of IL-8. This antibody had no observable effect on the growth of the A549 cells in vitro, indicating that IL-8 has a paracrine role in tumor growth and angiogenesis in vivo that correlates with IL-8 expression levels (7). Recombinant techniques have also demonstrated that increasing IL-8 expression in bladder carcinoma cells enhances their metastatic potential whereas decreasing IL-8 expression attenuates their metastatic potential (8). The levels of mRNA for 13 cytokines expressed in normal breast and breast tumor tissues were quantified and compared in a large clinical study. Of these mRNA, the only one that varied significantly between normal breast tissue and tumor tissues was IL-8 mRNA (P = 0.0017) (9). The IL-8 mRNA levels were significantly higher in tumor tissues, correlating with the tumor status.

In cell culture, we have observed that highly metastatic breast carcinoma cell lines MDA MB 231 (MDA-231) and MDA MB 435 (MDA-435) produce and release prodigious amounts of ectopic IL-8 whereas their low metastatic counterparts MCF-7 and T47D produce little if any (10). We refer to the undifferentiated MDA-231 and MDA-435 cell lines as metastatic because of their high propensity to spontaneously metastasize in xenographic models of tumor growth. We refer to the differentiated MCF-7 and T47D as nonmetastatic because of their low propensity to spontaneously metastasize in these models. The metastatic cells not only produce substantially higher quantities of IL-8 constitutively, but their responses to natural inducers of IL-8, such as IL-1β and tumor necrosis factor (TNF)-α, are robust compared with the modest responses of the nonmetastatic cells (10, 11).

Determining the cellular mechanisms responsible for ectopic IL-8 production by tumor cells is likely to increase our understanding of IL-8-induced metastatic cascades. There are at least three potential cellular mechanisms that may contribute to the differential regulation of IL-8 in breast carcinoma lines. First, metastatic cells may express active forms of the transcription factors required for a robust expression of IL-8, which are not expressed or expressed at much lower levels in the nonmetastatic cells. Second, the nonmetastatic cells may preferentially express a repressor that silences IL-8 gene transcription. It is known that the basal IL-8 promoter, between –136 and +43, contains four cis-regulatory elements (2). Three positive acting cis-regulatory elements, which are located at –82 to –70, –94 to –84, and –126 to –120 are reported to bind NF-κB, CCAAT/enhancer-binding protein (C/EBP), and activator protein 1 (AP1), respectively (2). The fourth cis-regulatory element, which is located at –90 to –83, is reported to bind Oct-1 and repress IL-8 transcription (2). A third potential mechanism for regulating the differential IL-8 expression is the methylation of deoxycytidylate-phosphate-deoxyguanylate (CpG) sequences within the promoter region, resulting in epigenetic regulation of gene transcription. The methylation of CpG islands within the promoter region usually leads to silencing of gene expression (12, 13). In this study, we demonstrate a differential methylation pattern of the IL-8 gene ≈1.2 kb upstream of the published promoter. However, this methylation, unlike the common epigenetic paradigm, shows a positive correlation between the methylation of two upstream CpG sites and gene expression. The findings in this study are more analogous to those found for the glycoprotein hormone α (GPH-α) gene. It has been shown that methylation of CpG sites in an Alu sequence in the 5′-flanking region of the GPH-α gene correlates with an increase in the expression of GPH-α (13, 14).

Materials and Methods

Cell Lines and Culture Conditions. The MDA-231, MDA-435, MCF-7, and T47D cell lines (American Type Culture Collection, Manassas, VA) were cultured with antibiotic-free DMEM (Mediatech, Herndon, VA) plus 10% FBS (BioWhittaker) in sterile tissue culture flasks and incubated at 37°C/6% CO2. The cell lines were certified to be mycoplasm-free. Cells were subcultured by trypsinizing in 5 mg/ml trypsin (Sigma) and 0.5 mM EDTA in Hanks' balanced salt solution without Ca2+ or Mg2+ in a laminar flow hood during their logarithmic phase of growth.

ELISA for Human IL-8. The breast carcinoma cell lines were seeded in six-well plates containing 2 ml of complete medium per well. At 80% confluency, the medium was exchanged for 2 ml of fresh complete medium with or without 1 ng/ml IL-1β (a gift from J. Cone, Otsuka America Pharmaceutical, Rockville, MD). After a 24-h treatment, 1.5 ml of medium was collected from each well, clarified of cells and cellular organelles and stored at –20°C, and the number of cells per well was determined. The IL-8 ELISA was performed on the conditioned medium according to the manufacturer's instructions (OptEIA Human IL-8 set, Pharmingen), and the results are given in pg/ml per 24 h per 106 cells.

RT-PCR. Populations of MCF-7, T47D, MDA-231, and MDA-435 were cultured in the presence or absence of 1 ng/ml IL-1β for 24 h. Total RNA was isolated from the populations by using TRIzol reagent (Invitrogen). mRNA was isolated from precipitated total RNA samples by using the Poly(A)Pure mRNA isolation kit (Ambion, Austin, TX). RT-PCR was performed by using the Enhanced Avian HS RT-PCR Kit (Sigma) following the recommended one-step RT-PCR procedure to produce cDNA for each sample. The cDNA were purified by using the QIAquick PCR Purification Kit (Qiagen, Valencia, CA). IL-8 and GAPDH cDNA were amplified by PCR using 1 and 10 ng of template cDNA for each sample and the following primer sets: IL-8, TCTGCAGCTCTGTGTGAAGG (forward) and ACTTCTCCACAACCCTCTGC (reverse); GAPDH, AACCCATCACCATCTTCCAGGAGC (forward) and CACAGTCTTCTGAGTGGCAGTGAT (reverse). Amplification was performed as follows: denaturation at 95°C for 30 s, annealing at 58°C for 20 s, and extension at 72°C for 30 s for 30 cycles. The PCR products (15 μl of each IL-8 PCR product and 2 μl of each GAPDH PCR product) were separated on a 2% agarose gel.

Transient Transfections. The IL-8 promoter was cloned from breast carcinoma cell genomic DNA by using nested PCR. The first primer set (forward, 5′-CCCAGGCATTATTTTATCCTCAAGTC-3′; reverse, 5′-ACACACAGTGAGAATGGTTCCTTCC-3′) amplifies the area between –1470 and +90 of the IL-8 gene transcription start site. The product from this reaction was used as the template to create the promoter inserts of varying sizes. The promoter inserts were created to contain restriction enzyme sites for XhoI and HindIII (New England Biolabs) (underlined in primer sequence below) to allow insertion into the pGL3 vector (Promega), which is a luciferase reporter vector. An extended version of the IL-8 promoter, –1453 to +47, was created by using the following: forward, 5′-GACTCTCGAGCCTCAAGTCTTAGGTTGGTTG-3′; reverse, 5′-AGAAGCTTGTGTGCTCTGCTGTCTCTG-3′. A short version of the IL-8 promoter, –560 to +47, was created by using the following primer sequence: forward, 5′-GTGCTCGAGCCCCTTCACTCTGTTAACTAGC-3′; and the same reverse primer. The promoter inserts were digested with XhoI and HindIII, gel purified, and ligated into similarly treated pGL3 vector by using T4 DNA ligase (Promega) and expanded in Escherichia coli. The purified constructs were transfected into the metastatic MDA-231 and the nonmetastatic MCF-7 breast carcinoma cell lines by using Lipofectamine 2000 (Invitrogen). Cells were plated in 24-well plates and transfected 24 h later at 40–60% confluency. They were cotransfected with the β-galactosidase vector, pCMV-βgal, to correct for differences in transfection efficiency. Transiently transfected cells were incubated at normal growth conditions for 24 h and assayed for transgene expression by using Luciferase Reporter Assay (Promega) and TopCount-NXT (Packard). Briefly, the cells were rinsed and lysed in the plate by using the cell culture lysis reagent. The lysate was vortexed and chilled, and the cell debris was pelleted by centrifugation. Luciferase expression was determined by adding 100 μl of Luciferase Assay Reagent (Promega) to 20 μl of cell lysate. Each sample was assayed for luminescence for 10 s. β-Galactosidase activity was determined by combining 50 μl of cell lysate from the luciferase assay system with 100 μl of water and 150 μl of 2× β-gal buffer (200 mM sodium phosphate, pH 7.3/2 mM MgCl2/100 mM 2-mercaptoethanol/1.33 mg/ml o-nitrophenyl-β-galactopyranoside). The reactions were incubated at 37°C for 30 min, and the reaction was stopped by adding 1 ml of 1 M Tris-free base. The results are expressed as a function of the luciferase values divided by the absorbance of the β-galactosidase samples at 420 nm for each replicate.

Stable Transfections. The IL-8 promoter for stable transfection was cloned from the original PCR product used to produce the inserts for transient transfection. Promoter inserts were then created to contain restriction enzyme sites for XhoI and BamHI (New England Biolabs) (underlined in primer sequence) to allow insertion into the pd2EGFP-1 vector (Clontech), which is a GFP reporter vector. A long version of the IL-8 promoter, –1453 to +46, was created by using the following primer sequence: forward, 5′-GACTCTCGAGCCTCAAGTCTTAGGTTGGTTG-3′; and reverse, 5′-CTGAGGATCCGTGTGCTCTGCTGTCTCTG-3′. The promoter insert was digested with XhoI and BamHI, gel purified, and ligated into similarly treated pd2EGFP-1 vector by using T4 DNA ligase and expanded in E. coli. The purified construct was linearized with BglII (New England Biolabs), which cuts upstream of the insert, and transfected into the metastatic MDA-231 and the nonmetastatic MCF-7 breast carcinoma cell lines by using Lipofectamine 2000. Cells were plated in six-well plates and transfected 24 h later at 40–60% confluency, according to the manufacturer's instructions. After an additional 24 h, the cells were trypsinized and replated to 10-cm dishes. At 48 h posttransfection, G418 selection media were added to the plates. MDA-231 was selected with 1,200 μg/ml G418 and MCF-7 with 400 μg/ml G418.

Fluorescence-Activated Cell Sorting (FACS). Analysis of GFP activity in the transfected cell lines was performed by FACS. The University of Minnesota Cancer Center FACS Core Facility contains BD Biosciences FACSCalibur machines for bench top analysis. GFP expression levels from cell populations transfected with the extended promoter construct were compared with that of the vectoronly-transfected populations to gauge GFP intensity and the percentage of each population that is expressing GFP.

Sodium Bisulfite Sequencing. Genomic DNA was harvested from metastatic and nonmetastatic breast carcinoma cells by using the GenElute Mammalian Genomic DNA Miniprep Kit (Sigma). Sodium bisulfite sequencing was performed based on previously published methods (15, 16). Genomic DNA (≈2 μg) from each cell line was denatured by adding NaOH to 0.3 M and incubating at 50°C for 15 min. DNA was then incorporated into agarose by mixing 50 μl of 2% agarose with 25 μl of the denatured DNA solution, and agarose beads were formed by dropping 10-μl aliquots of the mixture into chilled mineral oil. Freshly prepared bisulfite reaction mix (2.8 M sodium metabisulfite/0.57 mM hydroquinone/0.5 M NaOH) was added (0.5 ml per bead), and the beads were incubated at 50°C in the dark for 16 h. Beads were washed six times in 1 ml of TE (10 mM Tris, pH 8/1 mM EDTA) for 10 min. Adding 0.5 ml of 0.2 M NaOH and incubating at 37°C for 30 min drove the reaction to completion. The beads were then washed three more times with TE and amplified via nested PCR by using primers specific for the modified DNA, which also contain restriction sites for cloning. The final PCR products were digested by the restriction enzymes HindIII and SacI (New England Biolabs), gel purified, and ligated into similarly treated pUC19 vector by using T4 DNA Ligase (Promega). E. coli were transformed with the resulting vector/insert constructs. Ten clones per cell line were grown in overnight cultures, and plasmids were isolated by QiaPrep Spin Miniprep (Qiagen) and sequenced. Cytosine residues are seen in the sequencing results only if 5-methylcytosine is present.

Results and Discussion

The metastatic cell lines MDA-231 and MDA-435 constitutively produced much higher levels of IL-8 and were more responsive to IL-8-inducing cytokines than the nonmetastatic cell lines MCF-7 and T47D (Table 1). Comparing the constitutive levels of IL-8 released by the metastatic MDA-231 cells with that released by the nonmetastatic MCF-7 cells indicates that the metastatic cells released a minimum of 15-fold more IL-8 than the nonmetastatic cells (MDA-231 released 215 pg/ml per 24 h per 106 cells vs. 14 pg/ml per 24 h per 106 cells for MCF-7). The T47D cells did not release measurable quantities of IL-8 constitutively. The differential expression of IL-8 by these cell lines increased on treatment with the inflammatory cytokine IL-1β, which is a potent inducer of IL-8. The increased IL-8 expression in response to 1 ng/ml IL-1β was ≈225-fold for the MDA-231 cells and ≈13-fold for the MCF-7 cells. The differential level of IL-8 expression between these cell lines increased to ≈250-fold after exposure to IL-1β. The stimulated metastatic MDA-435 cells, which constitutively expressed very high levels of IL-8, released ≈135,000 pg/ml per 24 h per 106 cells whereas the IL-1β-stimulated nonmetastatic T47D cells release 96 pg/ml per 24 h per 106 cells after IL-1β exposure.

Table 1.
IL-8 release by breast carcinoma cell lines

To ensure that the differences in IL-8 protein levels were not merely attributable to translational differences, RT-PCR was performed to determine whether levels of IL-8 mRNA correlated with secreted protein levels. Double-stranded cDNAs were produced from each cell line (untreated and stimulated with IL-1β) and probed by using specific primers for either IL-8 or GAPDH mRNA. We found the relative quantity of PCR product to be consistent with the variation in IL-8 protein levels released by these cells under these conditions. Nonmetastatic MCF-7 cells constitutively released little if any IL-8 protein, and a faint IL-8 mRNA-specific PCR product was seen by using 10 ng of cDNA input (Fig. 1A). MDA-231 cells showed easily detectable IL-8 cDNA, and the MDA-435 cells showed even greater expression (Fig. 1 A). This result is consistent with the levels of IL-8 protein released by these cells (Table 1). RT-PCR data from IL-1β-induced cells (Fig. 1C) show that IL-8-specific PCR product derived from the metastatic cells was more highly induced than that from the nonmetastatic lines. IL-1β had a greater affect on the level of IL-8-specific PCR product from MDA-231 cells than from the MDA-435 cells, similar to its affect on IL-8 protein induction in these cells. The GAPDH control product showed consistency in these PCR reactions and gel loading. The levels of GAPDH PCR product remained relatively constant and were not affected by IL-1β stimulation.

Fig. 1.
Relative IL-8 mRNA expression by breast carcinoma cells. cDNAs were made from mRNA isolated from the four breast carcinoma cell lines with and without 1 ng/ml IL-1β and examined for presence of IL-8 (A and C) and GAPDH (B and D) cDNA by PCR. One ...

The large variation in the IL-8 PCR product levels between the metastatic and nonmetastatic cells suggests that IL-8 production levels may be attributable to differential activity of the IL-8 promoter. Examination of the IL-8 promoter indicates that there are four transcription factor consensus sequences within the vicinity of the TATA box (2). To determine whether the presence or absence of these active transcription factors is responsible for the observed differential levels of IL-8 expression, an IL-8 promoter spanning the sequence between –560 and +47 was inserted into a luciferase reporter construct. If either the presence or absence of these factors is responsible for the distinctly different levels of IL-8 expression between these classes of cells, then levels of luciferase activity would be expected to parallel the differences in IL-8 expression. MDA-231 and MCF-7 breast carcinoma cells were chosen for transfection because they are both well characterized breast adenocarcinoma lines, the MDA-231 being an undifferentiated metastatic cell line and the MCF-7 being a differentiated nonmetastatic line. These cell lines were transiently transfected with the IL-8 promoter–reporter construct and a β-galactosidase vector, the latter to correct for differences in transfection efficiency. The data indicate that this promoter–reporter construct was active in both cell types, and the differences in luciferase expression did not parallel the differences in IL-8 expression (Fig. 2). MCF-7 cells, which produce much less IL-8, expressed higher levels of luciferase than the metastatic cells when transiently transfected with this IL-8 promoter sequence. These data show that both of these cell lines have transcription factors necessary to activate the IL-8 promoter–reporter construct. They also argue that the differential expression of IL-8 between the metastatic and nonmetastatic cells is not attributable to mere differences in the levels of active transcription factors that bind to the region –560 to +47 of the IL-8 gene. Thus, other mechanisms must account for the differences in protein expression between these cells.

Fig. 2.
Luciferase expression by IL-8 promoter–reporter constructs in breast carcinoma cell lines. Cells were plated in 24-well tissue culture plates in DMEM containing 10% FBS. After 24 h, the cells were transfected with the stated length constructs ...

A possible mechanism to account for this apparent discrepancy is that the nonmetastatic cells may express a repressor activity that recognizes a sequence not present in the above promoter construct. A putative repressor preferentially expressed in the nonmetastatic cells could attenuate IL-8 expression. Consequently, an extended promoter–luciferase reporter construct was assembled by using the sequence between –1453 and +47. The activity of this extended promoter was compared with that of the –560 to +47 (short) promoter–reporter construct in transient transfections into MDA-231 and MCF-7 cells. The extended promoter was much less active than the shorter promoter in both cell types (Fig. 2). In the MCF-7 cells, the extended promoter had ≈5% of the activity of the shorter promoter. These results are consistent with there being a negative regulatory sequence located in the 5′ portion of the extended promoter to which a putative repressor could bind, attenuating IL-8 expression. However, because the extended promoter had greatly reduced activity in both cell types, a negative regulatory sequence alone is not sufficient to explain the preferential expression of IL-8 by the undifferentiated metastatic cells.

To determine whether the activity of the extended promoter, when incorporated in the chromatin of breast carcinoma cells, more closely parallels the activity expressed by the native IL-8 promoter, a promoter–reporter construct was created for stable transfection. The extended promoter, –1453 to +46, was ligated into the pd2EGFP-1 vector, which has a selectable marker and a short half-life GFP as its reporter. The GFP reporter allowed us to monitor conveniently the reporter expression level within the same population over time. Both MDA-231 and MCF-7 cells were transfected with this extended promoter construct. The transfected MDA-231 cells fluoresced green 3 weeks after transfection whereas the corresponding MCF-7 cells did not fluoresce even after 12 weeks. The quantity of GFP expressed by populations of MCF-7 and MDA-231 cells 9 weeks posttransfection is demonstrated by FACS analysis (Fig. 3). No detectable GFP fluorescence above that of the vector only control (Fig. 3A) was expressed by MCF-7 cells transfected with this construct-(Fig. 3B). However, the MDA-231 cells transfected with this construct express much greater GFP fluorescence (Fig. 3D) than the low background level expressed by the MDA-231 cells transfected with the vector alone (Fig. 3C). These results are distinctly different from those obtained when the promoter construct was transiently transfected into these cell lines. In transient transfection, the MCF-7 cells produced a greater reporter response than the MDA-231 cells whereas this promoter did not seem to be active in the stably transfected MCF-7 cells. Results from the stably transfected promoter–reporter constructs more closely parallel IL-8 protein expression levels from these cells than do the results from the transient transfections. The MDA-231 cells released substantially more IL-8 and expressed more GFP fluorescence from the promoter–reporter constructs than the MCF-7 cells.

Fig. 3.
GFP expression in stable transfectants. MDA-231 and MCF-7 cells were stably transfected with the IL-8 promoter–GFP reporter construct and carried under G418 selection for 9 weeks. Histograms of fluorescence intensity in GFP channels are shown ...

To gain further insight into the cellular mechanism responsible for the differential expression of IL-8 between these tumor cell types, their IL-8 genes were analyzed for the presence of epigenetic regulatory components. Epigenetic processes play key regulatory roles in development, differentiation, and tumor formation (17). An example of epigenetic modifications found in tumor cells is the silencing of tumor suppressor genes (18, 19). The common paradigm for epigenetic regulation of gene expression is methylation of the cytosine in CpG islands in promoters, which silences gene expression. Consequently, we examined the methylation patterns of the CpG sites in the IL-8 promoter to determine whether there is a differential pattern of methylation that correlates with IL-8 expression. Four CpG sites are located within the vicinity of the reported promoter (20). These sites are located at –7, –83, –158, and –168. Although the four CpG sequences lack the numbers to constitute a typical “CpG island,” they present potential targets for methylation and the epigenetic suppression of IL-8 promoter activity. The methylation status of these four sites, in two metastatic (MDA-231 and MDA-435) and two nonmetastatic breast carcinoma cell lines (MCF-7 and T47D), was determined by using the sodium bisulfite sequencing technique. Ten clones from each line were sequenced. None of these four CpG sites was found to be methylated in any of the cell lines. Two additional CpG sequences located upstream of the TATA box at –1241 and –1311 were sequenced. The data indicated these two CpG sites were preferentially methylated in both of the metastatic cell lines, which expressed high levels of IL-8. However, they were unmethylated in the nonmetastatic cells that express little or no IL-8 (Table 2).

Table 2.
Methylation of CpG sites in breast carcinoma cell lines

Previously, it was shown that nonmetastatic breast carcinoma cells that survive exposure to chemotherapeutic agents express much more IL-8 than their untreated parental cells. Along with this increase in IL-8 expression, the survivors exhibited a dramatic increase in their metastatic potential (11). To test the association between IL-8 expression and methylation of the gene, we compared the levels of IL-8 expression and the methylation pattern of the two 5′-upstream CpG sites in a population of MCF-7 cells that survived three 72-h exposures to docetaxel with those of the untreated control MCF-7 population (Table 3). The data indicate that, along with an increase in the constitutive and inducible IL-8 expression levels in the population of survivors, there is a dramatic, concomitant increase in methylation of the IL-8 gene at these two 5′-upstream CpG sites. Thus, there is a strong correlation between methylation of the two upstream CpG sites and the expression of IL-8.

Table 3.
Changes in IL-8 expression and CpG methylation in MCF-7 docetaxel survivors

In the standard paradigm, extensive methylation of cytosines in CpG islands correlates with gene silencing, as opposed to our current findings in which methylation correlates with enhanced IL-8 expression. Not only are these two isolated CpG sequences not islands, but they are >1.2 kb upstream of the TATA box, and methylation of the cytosine in these sites occurs in cells actively producing high levels of IL-8. Our results indicate that methylation of these two upstream CpG sites in these breast carcinoma lines correlates with an increase in their IL-8 expression. It further suggests that methylation of these sites is required for a cell to express high levels of IL-8 when induced by IL-1β. Together, the data argue that there is an epigenetic mechanism operative in regulating IL-8 expression in the breast carcinoma cells that, as yet, is not fully appreciated or understood. Several possible mechanisms could explain how methylation of these two 5′ CpG sites could enhance IL-8 expression. The methylated cytosines could inhibit the binding of a repressor within the vicinity of the two 5′ CpG sites. This inhibition may be either direct or indirect if the methylated cytosine serves as a binding site for the assembly of a protein complex that interferes with the binding of the putative repressor. Interestingly, methylation of the 5′ flanking region of the GPH-α gene is believed to block the binding of a repressor that inhibits GPH-α transcription (13, 14). Alternatively, the methylated sites may serve as a center around which a protein complex forms to enhance the remodeling of the chromatin in the vicinity of the IL-8 gene and up-regulate its expression.

We expect that this aberrant methylation may contribute to ectopic IL-8 expression in other tumor systems as well as contribute to the physiological regulation of IL-8 and other gene products. Understanding the molecular mechanisms responsible for this DNA methylation and its role in ectopic IL-8 expression may present an opportunity for the development of a rational means of attenuating both IL-8 expression and metastatic potential. The methylation status of these two upstream CpG sites in fresh tumor specimens may also be useful as a prognosticator of tumor status.


We thank Sally Palm for helpful suggestions and discussions and Rebecca Willaert for technical help during these experiments. We thank the Department of Laboratory Medicine and Pathology at the University of Minnesota, the Allen–Pardee endowed chair to Professor Leo Furcht, and the Susan G. Komen Foundation Twin Cities Affiliate for supporting this work. We acknowledge the assistance of the Flow Cytometry Core Facility of the University of Minnesota Cancer Center, a comprehensive cancer center designated by the National Cancer Institute, supported in part by National Cancer Institute Grant P30 CA77598.


Abbreviations: GPH-α, glycoprotein hormone α; CpG, deoxycytidylate-phosphate-deoxyguanylate; FACS, fluorescence-activated cell sorting.


1. Matsushima, K., Morishita, K., Yoshimura, T., Lavu, S., Kobayashi, Y., Lew, W., Appella, E., Kung, H. F., Leonard, E. J. & Oppenheim, J. J. (1988) J. Exp. Med. 167, 1883–1893. [PMC free article] [PubMed]
2. Mukaida, N., Shiroo, M. & Matsushima, K. (1989) J. Immunol. 143, 1366–1371. [PubMed]
3. Strieter, R. M., Kunkel, S. L., Elner, V. M., Martonyi, C. L., Koch, A. E., Polverini, P. J. & Elner, S. G. (1992) Am. J. Pathol. 141, 1279–1284. [PMC free article] [PubMed]
4. Singh, R. K., Gutman, M., Radinsky, R., Bucana, C. D. & Fidler, I. J. (1994) Cancer Res. 54, 3242–3247. [PubMed]
5. Singh, R. K., Gutman, M., Reich, R. & Bar-Eli, M. (1995) Cancer Res. 55, 3669–3674. [PubMed]
6. Luca, M., Huang, S., Gershenwald, J. E., Singh, R. K., Reich, R. & Bar-Eli, M. (1997) Am. J. Pathol. 151, 1105–1113. [PMC free article] [PubMed]
7. Arenberg, D. A., Kunkel, S. L., Polverini, P. J., Glass, M., Burdick, M. D. & Strieter, R. M. (1996) J. Clin. Invest. 97, 2792–2802. [PMC free article] [PubMed]
8. Inoue, K., Slaton, J. W., Kim, S. J., Perrotte, P., Eve, B. Y., Bar-Eli, M., Radinsky, R. & Dinney, C. P. (2000) Cancer Res. 60, 2290–2299. [PubMed]
9. Green, A. R., Green, V. L., White, M. C. & Speirs, V. (1997) Int. J. Cancer 72, 937–941. [PubMed]
10. De Larco, J. E., Wuertz, B. R. K., Rosner, K. A., Erickson, S. A., Gamache, D. E., Manivel, J. C. & Furcht, L. T. (2001) Am. J. Pathol. 158, 639–646. [PMC free article] [PubMed]
11. De Larco, J. E., Wuertz, B. R. K., Manivel, J. C. & Furcht, L. T. (2001) Cancer Res. 61, 2857–2861. [PubMed]
12. Jones, P. A. (1999) Trends Genet. 15, 34–37. [PubMed]
13. Cox, G. S., Gutkin, D. W., Haas, M. J. & Cosgrove, D. E. (1998) Biochim. Biophys. Acta 1396, 67–87. [PubMed]
14. Xiong, W., Tapprich, W. E. & Cox, G. S. (2002) J. Biol. Chem. 277, 40235–40246. [PubMed]
15. Frommer, M., McDonald, L. E., Millar, D. S., Collis, C. M., Watt, F., Grigg, G. W., Molloy, P. L. & Paul, C. L. (1992) Proc. Natl. Acad. Sci. USA 89, 1827–1831. [PMC free article] [PubMed]
16. Olek, A., Oswald, J. & Walter, J. (1996) Nucleic Acids Res. 24, 5064–5066. [PMC free article] [PubMed]
17. Robertson, K. D. (2002) Oncogene 21, 5361–5379. [PubMed]
18. Gonzalez-Zulueta, M., Bender, C. M., Yang, A. S., Nguyen, T., Beart, R. W., Van Tornout, J. M. & Jones, P. A. (1995) Cancer Res. 55, 4531–4535. [PubMed]
19. Baylin, S. B., Herman, J. G., Graff, J. R., Vertino, P. M. & Issa, J. P. (1998) Adv. Cancer Res. 72, 141–196. [PubMed]
20. Mukaida, N., Mahe, Y. & Matsushima, K. (1990) J. Biol. Chem. 265, 21128–21133. [PubMed]

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...