Welcome international healthcare professionals

This site is no longer supported and will not be updated with new content. You are welcome to browse and download all content already included in the site. Please note you will have to register your email address to access the site.

You are here

Impact of LIF (leukemia inhibitory factor) expression in malignant melanoma

Experimental and Molecular Pathology, 2, 95, pages 156 - 165


Leukemia inhibitory factor (LIF) signaling regulates cellular processes to maintain the self-renewal and pluripotency of embryonic stem (ES) cells. Independent of these capabilities, LIF was also identified to be responsible for cancer development and progression. However, its detailed cellular function in cancer remains unclear thus far. We found LIF to be expressed in melanoma cell lines of primary and metastatic origin and in melanoma tissue. We further elucidated stimuli that are responsible for the high expression levels of LIF. Interestingly, hypoxia, specifically through HIF-1α, is involved in regulating LIF.

Furthermore, our data showed that the signaling of LIF was not mediated by the classically described pathway via STAT3, but rather through BMP4 and BMP7. We hypothesize that the co-expression of LIF and BMP is necessary for a de-differentiated cancer phenotype. Ancillary to BMP4 and BMP7, classical stem cell proteins, e.g., SOX2, NANOG, OCT3/4 and GBX2, are regulated by LIF. We therefore speculate that LIF can induce a typical “cancer stem cell”-like behavior, as the appropriate genes are regulated by LIF. Particularly, the expression of these genes has been proposed as a driving force for tumorigenesis and the initiation of metastasis. Notably, LIF has an important role not only for ES cells but also for cancer development. Melanoblast-related cells (MBrcs), which resemble the neural crest precursor cells of melanocytes, expressed LIF in minor amounts compared to normal human melanocytes. These data, along with the data that LIF is upregulated in melanoma cell lines compared to melanocytes, strongly indicate that LIF is important for the stabilization of the melanoma phenotype.

To elucidate the role of LIF in cellular melanoma behavior, we analyzed proliferation, attachment, migration and colony formation after silencing LIF by siRNA, and found all four characteristics restricted. In summary, we can show that LIF is an important factor in melanoma progression.



  • LIF is strongly expressed in malignant melanoma
  • LIF expression is modulated via hypoxia and HIF1a
  • LIF induces expression of BMPs and stem cell factors in melanoma
  • Proliferation, attachment and migration are supported by LIF in melanoma

Abbreviations: Angptl4 - Angiopoietin-Like 4, AP-1 - Activator Protein 1, DFX - Desferrioxamine, DP - 2,2 Dipyridyl, GAPDH - Glyceraldehydes-3-phosphate dehydrogenase, HIF-1 - Hypoxia inducible factor, NF-κB - Nuclear factor kappa B, LIF - Leukemia inhibitory factor, PI3K - Phosphotidylinositol 3-kinase, KLF4 - Kruppel-Like factor 4, SOX2 - SRY sex determining region Y-box 2, TBX3 - T-box transcription factor, OCT3/4 - Octamer-binding transcription factors 3 and 4, GBX2 - Gastrulation brain homeobox 2.

Keywords: Melanoma, LIF, Stem cell marker.


Malignant melanoma is the cause of ~ 90% of the mortality rate of all skin carcinomas and is notorious for its resistance to therapy. One option for treatment is the eradication of melanoma stem cells. However, the reliable identification of these melanoma stem cells is complicated by the lack of clearly defined markers to distinguish them from the general tumor cell population. Therefore, the stem cell debate is ongoing in the melanoma research field. Notably, leukemia inhibitory factor (LIF) was identified as a crucial factor for mediating the self-renewal of mouse embryonic stem cells (ESCs). LIF was shown to be a growth factor for ESCs, maintaining their pluripotency ( Casanova et al., 2011 ). In the review by Nakai-Futatsugi and Niva (2013), LIF signaling was described as regulating transcription factors that maintain the self-renewability and pluripotency of ESCs ( Nakai-Futatsugi and Niwa, 2013 ). The authors proposed a network model that consists of transcription factors such as KLF4, SOX2, TBX3, NANOG and OCT3/4, forming a parallel pathway downstream of LIF signaling ( Niwa et al., 2009 ). Unlike mouse ES cells, LIF cannot promote the self-renewal of human or monkey ES cells; it has been shown that neither LIF nor STAT is directly involved in these processes in humans ( Humphrey et al., 2004 ).

LIF is a member of the neuropoietic class of cytokines, including interleukin 6 (IL-6), interleukin 11 (IL-11), ciliary neutrophic factor (CNTF), cardiotrophin (CT-1), and oncostatin M (OSM) ( Turnley and Bartlett, 2000 ), and it has a broad and diverse spectrum of functions and activities in the endocrine, neural, hepatic, stromal, immune, and inflammatory systems. Furthermore, LIF is involved in blastocyst implantation and fetus development by providing an immune-tolerant environment in the deciduas ( Gearing et al., 1993 ). LIF was originally isolated as a factor that induced macrophage differentiation from murine myeloid leukemic cells and suppressed their proliferation in vitro ( Gearing et al., 1987 ). Independently of these capabilities, the molecule was also shown to be responsible for cancer development and progression. LIF can act as a growth-stimulatory or growth-inhibitory factor; it can sustain proliferation or differentiation depending on the tissue types and cell activities. Tumor cell lines known to express LIF include lymphoma, breast carcinoma, colon adenocarcinoma, pancreatic adenocarcinoma, lung adenocarcinoma, bladder carcinoma, epidermoid carcinoma, squamous carcinoma, and melanoma (Alexander et al, 1994, Auernhammer and Melmed, 2000, Gearing et al, 1993, Kamohara et al, 1994, Paglia et al, 1995, and Szepietowski et al, 2001). The detailed cellular function of LIF in these cancers remains unclear thus far. In melanoma, LIF increases the attachment to the extracellular matrix through an increase in αvβ1 and intracellular adhesion molecule-1 (Heymann et al, 1995a and Heymann et al, 1995b). LIF has consensus binding sequences in its promoter region for nuclear factor (NF)-κB, activator protein-1 (AP-1), and NF-IL-6 ( Knight, 2001 ), which are all important transcription factors of the epidermis. LIF can signal via the RAS/mitogen-activated protein (MAP) kinase pathway, the phosphatidyl-inositol-3 (PI3) kinase/AKT pathway, the AP-1 pathway, and the Janus kinase (JAK)-1/signal transducer and activators of transcription (STAT) pathway (Auernhammer and Melmed, 2000, Gearing et al, 1993, Oh et al, 1998, and Thoma et al, 1994). LIF has effects in normal skin and skin growth and is expressed in normal human keratinocytes (McKenzie and Szepietowski, 2004 and Paglia et al, 1996). The protein structure of LIF is also variable. It has a molecular weight of 37 to 62 kDa depending on its degree of glycosylation ( Simpson et al., 1988 ). LIF exists in at least three isoforms, i.e., a soluble form called LIF-D; a nuclear form, called LIF-T; and a third form localized in the extracellular matrix, termed LIF-M ( Rathjen et al., 1990 ).

Our present study supplements other studies showing that LIF expression is found in diverse melanoma cell lines of primary and metastatic origins ( Maruta et al., 2009 ). We therefore searched for external stimuli or regulators that led to the over-expression of LIF in melanoma cells. We examined the effects of the transfection of LIF-targeted siRNA on cell proliferation, attachment, migration and colony formation and also found new target genes of LIF signaling. In summary, LIF over-expression is a factor in promoting the progression of melanoma. However, it still remains unclear whether LIF is responsible for keeping melanoma cells in a pluripotent or un-differentiated state.

Materials and methods

Cell culture

The human melanoma cell lines Mel Wei, Mel Juso and Mel Ho have their origin in primary melanoma. Mel Im, Mel Ju, SKMel28 (ATCC® HTB-72™), SKMel3 (ATCC® HTB-69™) and HTZ19d were derived from metastases of malignant melanomas. Cells were maintained in DMEM supplemented with penicillin (400 U/ml), streptomycin (50 μg/ml), l-glutamine (300 μg/ml) and 10% fetal calf serum (FCS) and split at a 1:5 ratio every three days. Melanocytes were cultivated from the human neonatal foreskin tissue of Caucasian donors. The melanocytes grew in melanocyte serum-free media from PromoCell (M2 media without PMA (phorbol-myristate-acetate)) for three passages.

The designation melanoblast-related cells refers to the aforementioned melanocytes cultivated and de-differentiated in special melanoblast growth media for 5 or 14 days. The cultivation was described in detail by Bosserhoff et al. (2011) . The cells first grew as melanocytes and were then divided into two sub-groups. One sub-group (designated as melanocytes) continued growing in PromoCell M2 medium for 5 or 14 days. For the other sub-group, the growing conditions were changed to MCBD medium for melanoblasts. The melanoblast-related cells originate from the identical melanocytic cell pool and were designated as “MC:MB” cells in the manuscript of Cook et al. (2003) . In our manuscript, melanocytes were induced to de-differentiate to melanoblast-related cells (MBrcs).


Desferrioxamine (DFX, 250 μM) and 2,2′dipyridyl (DP, 50 μM) treatment were performed for 18 h. Recombinant LIF was used at a concentration of 100 ng/ml (Santa Cruz Biotechnology, Inc. [sc4988], Heidelberg, Germany) for 24 h.

RNA isolation, RT-PCR and quantitative real-time PCR

The total RNAs of cells were isolated using the e.Z.N.A. MicroElute Total RNA Kit (peqlab Biotechnologie GmbH, Erlangen, Germany) as described by the manufacturers. The purity and concentration were measured in a NanoDrop (peqlab Biotechnologie). cDNA was generated by Reverse Transcription (RT). The RT reaction was performed in a 20 μl reaction volume containing 500 ng of total RNA, 4 μl of 5 × first-strand buffer, 2 μl of 0.1 M dithiothreitol (both Invitrogen Corporation, Darmstadt, Germany), 1 μl of dN6 primer (10 mM) (Roche Applied Science, Mannheim, Germany) and 1 μl of dNTPs (10 mM) (Amersham Pharmacia Biotech, Pittsburgh, PA, USA). The reaction mix was incubated for 5 min at 70 °C, and 1 μl of Superscript II reverse transcriptase (Invitrogen Corporation) was subsequently added. RNA was transcribed for 1 h at 37 °C. Finally, reverse transcriptase was inactivated at 70 °C for 10 min, and RNA was degraded by digestion with 1 μl RNase A (Roche Applied Science) at 37 °C for 30 min. Quantitative RT-PCR was carried out with the Lightcycler480 system from Roche. A volume of 1 μl cDNA template, 0.5 μl of forward and reverse primers (20 mM) for each gene, 10 μl of SYBR-Green Premix (Roche applied Science) and 8 μl water were combined to a total volume of 20 μl. PCR primers were obtained from Sigma-Aldrich Corporation. The following PCR program was used: 95 °C for 10 min (initial denaturation); 4.4 °C · s− 1 temperature transition rate up to 95 °C for 10 s; 60 °C for 10 s; 72 °C for 20 s, 80 °C acquisition mode single, repeated for 45 times (amplification). The PCR product was evaluated by melting-curve analysis and PCR product analysis on 1.8% agarose gels. Each sample was analyzed in duplicate. The expression ratios of the analyzed genes were normalized to the expression level of the housekeeping gene β-actin. Oligonucleotide primers used in the PCR were as follows: HIF-1α for: 5′-CAC AGG CCA CAT TCA CGT A-3′; HIF-1α rev: 5′-ATC CAG GCT GTG TCG ACT G-3′; AngPTL4 for: 5′-CAG GGT ACC TAA GAG GAT GAG CGG TGC-3′; AngPTL4 rev: 5′-CTG CTC GAG CTG CAG GAG TCC GTG C-3′; LIF for: 5′-TCT TGG CGG CAG GAG TTG TG-3′; LIF rev: 5′-CTT CTC CGT GCC GTT GGC GT-3′; BMP4 for: 5′-GAT CCC GTC CAA GCT ATC-3′; BMP4 rev: 5′-TCC ATG ATT CTT GAC AGC C-3′; BMP7 for: 5′-GCC AGC CTG CAA GAT AGC CAT TTC C-3′; BMP7 rev: 5′-GAG CAC CTG ATT AAA CGC TGA TCC GG-3′; CyclinD1 for: 5′-GCCTGTGATGCTGGGCACTTCATCTG-3′; CyclinD1 rev: 5′-TTTGGTTCGGCAGCTTGCTAGGTGAC-3′; SOX2 for: 5′-TGCTGCCTCTTTAAGACTAGGAC-3′; SOX2 rev: 5′-CCTGGGGCTCAAACTTCTC-3′; NANOG for: 5′-GCCCTGATTCTTCCACCAGT-3′; NANOG rev: 5′-TGCGTCACACCATTGCTATTC-3′; OCT3/4 for: 5′-AAGGAGAAGCTGGAGCAAAA-3′; OCT3/4 rev: 5′-CCACATCGGCCTGTGTATATC-3′; GBX2 for: 5′-GCTCACAAGGAGGAAGACCC-3′; GBX2 rev: 5′-TTGGAATTGGCATTGCCTGC-3′.

Transfection experiments

Cells were plated 2 × 105 cells/well into 6-well plates and transfected with 0.5 μg plasmid DNA using the Lipofectamine Plus method (Invitrogen Corporation, Darmstadt, Germany) according to the manufacturer's instructions. The expression vector for dominant negative HIF-1α was generated by introducing a stop codon (TGA) after amino acid 380 in the wild-type mouse HIF-1α coding sequence. The resulting truncated protein thus lacks the oxygen-dependent degradation domain and all transactivation domains. Therefore, it competes with wild-type HIF-1α for the dimerization partner ARNT and may bind to HIF-responsive elements, but it does not transactivate the respective target genes. TAM67 (dominant-negative AP-1), having a deleted N-terminal transactivation domain, inhibits AP-1 activation by dimerizing with Jun and Fos family proteins, producing a complex with low activity ( Brown et al., 1994 ). The STAT5 expression vector STAT5a‐Flag wild‐type has been kindly provided by T. Kitamura ( Kawashima et al., 2006 ). The control pcDNA3 vector was from Invitrogen Corporation.

siRNAs and transfection procedures

HIF-1α siRNA (HIF-1α sense 5′ CUGAUGACCAGCAACUUGAdTdT) and LIF siRNA (sense 5′ CCCAACAACCTGGACAAGCTAdTdT) were synthesized by Qiagen (Cologne, Germany). The transfection reagent RNAiMAX (Life Technologies Corporation, Frankfurt a. M., Germany) was used according to the manufacturer's instructions.

Western blotting

Cells (3 × 106) were lysed in 200 μl RIPA-buffer (50 mM Tris/HCl pH 7.5, 150 mM NaCl, 1% NP-40, 0.1% SDS, 0.5% sodium deoxycholate and protease inhibitors) and incubated for 15 min at 4 °C. Insoluble fragments were removed by centrifugation at 13,000 rpm for 10 min, and supernatant lysates were immediately shock-frozen and stored at − 80 °C. Lysates (40 μg protein per lane) were resolved on 10% SDS-PAGE gels and blotted onto PVDF membranes. Membrane blocking was achieved by incubation with 3% BSA/TBS/0.1% Tween for 1 h. The following primary antibodies were used: anti-LIF (1 in 200 dilution; Abcam), anti-STAT3 (1 in 3000 dilution, Cell Signaling by NEB, Frankfurt a.M., Germany), anti-phospho-STAT3 (1 in 1000 dilution, Cell Signaling), anti-STAT5a, anti-phospho-STAT5a (both 1 in 1000 dilution, New England BioLabs, Frankfurt a.M., Germany), anti-GAPDH (1 in 1000 dilution, Cell Signaling) and anti-β-actin (1 in 5000 dilution, Sigma-Aldrich Corporation, Steinheim, Germany). Incubation with primary antibodies was for 16 h. After three washes with TBS, membranes were incubated for 1 h with an alkaline phosphatase-coupled secondary anti-mouse, anti-goat or anti-rabbit IgG in TBS/0.1% Tween (1 in 3000 dilution, NEB, Frankfurt a.M., Germany). Immunoreactive proteins were visualized by NBT/BCIP (Sigma-Aldrich Corporation, Steinheim, Germany) staining.

Boyden chamber

Invasion and migration assays were performed in Boyden chambers containing polycarbonate filters with 8 μm pore sizes (Costar, Bodenheim, Germany). For migration assays, filters were coated with a commercially available reconstituted basement membrane (Matrigel, diluted 1:3 in H2O; Becton Dickinson, Heidelberg, Germany). The lower compartment was filled with fibroblast-conditioned medium used as a chemoattractant plus recombinant (rc) LIF. NHEMs were harvested by trypsinization for 1 min, resuspended in M2 medium without FCS at a density of 3 × 104 cells/ml (migration) or 2 × 105 cells/ml (invasion) and placed in the upper compartment of the chamber. After incubation at 37 °C for 4 h, filters were removed. Cells adhering to the lower surface were fixed, stained and counted.

Measurement of attachment and proliferation by real-time-cell-analysis (RTCA)

The xCELLigence System (Roche Diagnostics GmbH, Mannheim, Germany) is an innovative method based on the measurement of electrical impedance to measure cell attachment, proliferation, apoptosis, migration and invasion in real time. E-plates (attachment, proliferation) were used, and basic protocols recommended by the manufacturer were followed. The chamber was equilibrated with 100 μl DMEM/10% FCS for 30 min at room temperature. After recording background impedance, cells suspended in DMEM were added to the chambers (4 × 103 cells/well for attachment and proliferation). Thereafter, impedance was measured continuously for approximately 96 h. Impedance is represented by the relative and dimensionless parameter named cell index (CI). CI values = Zi − Z0 / 15[Ohm]; Z0 = impedance at the start of the experiment, and Zi = impedance at individual time points during the experiment. The normalized cell index (NCI) was calculated as the cell index CIti at a given time point (ti) divided by the cell index CInml_time at the normalization time point (nml_time).

Growth in soft agar

For the measurement of attachment-independent growth, 1.2 × 104 cells per well were seeded into six well plates in DMEM/10% FCS in 0.4% agar on top of a 0.8% agar bed prepared in water. The cultures were incubated for 10 days at 37 °C, 5% CO2. After 10 days, the number and the size of colonies were recorded under an inverted microscope.

Statistical analysis

All experiments were performed at least 3 times. Unless otherwise stated, results are given as the mean ± SD. Comparison between groups was made using Student's paired t-test. All calculations were performed using the GraphPad Prism software (GraphPad Software Inc., San Diego, USA). A p value of < 0.05 was considered significant.


Expression of LIF in melanoma cell lines

Eight human melanoma cell lines ( Fig. 1 A) were evaluated for the expression of LIF using qRT-PCR and compared to human primary melanocytes (NHEMs). A higher expression level of LIF was found in all melanoma cell lines when normalized to melanocyte expression. We tested these cell lines by Western blot ( Fig. 1 B) and confirmed LIF expression in all but one (Mel Ho) of them, which reflects a good agreement between the real-time data and Western blot analysis, even if the mRNA and protein status of each cell line is not completely congruent. Immunohistochemical staining against LIF showed a weak expression of LIF in nevi but a strong expression in melanoma metastases; representative images are shown for three tissue samples ( Fig. 1 C). The data were supported by immunohistochemistry data at proteinatlas.org, where 22 melanoma tissues were analyzed; all showed a moderate to strong LIF staining in the cytoplasm and at the cell membrane.


Fig. 1 Expression of LIF in melanoma cell lines and tissues. (A) qRT-PCR for the expression level of LIF mRNA in normal human epidermal melanocytes (NHEMs), primary melanoma and melanoma metastases. The results are standardized to β-actin. The expression in NHEMs was set as 1. (B) Westen blot analysis of the protein level of LIF in melanoma cell lines and NHEMs. GAPDH was used as loading control. (C) Three human nevi and melanoma metastases were sectioned and immunohistologically stained with anti-LIF antibody. Strong LIF expression was found in the cytoplasm of melanoma metastases.

Regulation of LIF

We were interested in the stimuli that are responsible for the high expression rates of LIF in melanoma, and we analyzed several possibilities. First, we analyzed whether STAT5a or AP-1 are regulators of LIF expression, as has been described in the literature (Knight, 2001 and Salas et al, 2011). We used one primary cell line (Mel Juso) and three metastatic cell lines (HTZ19d, Mel Im, and Mel Ju) for these experiments. Investigating STAT5a phosphorylation and activity status in the melanoma cell lines compared to NHEMs, we surprisingly found nearly no phosphorylated STAT5a ( Fig. 2 A). To further validate this finding, we transfected a STAT5a expression construct into the melanoma cells, but found no connection between STAT5a and the expression of LIF ( Fig. 2 B). Additionally, the transfection of TAM67, a dominant negative AP-1 expression vector, does not hint at a connection between LIF and AP-1 ( Fig. 2 C).


Fig. 2 Involvement of STAT5a in LIF regulation. (A) Western blot analysis of the protein level of STAT5a in four melanoma cell lines and NHEMs. The phosphorylation of STAT5a was also detected by Western blot analysis. GAPDH was used as loading control. (B) qRT-PCR for the expression level of LIF mRNA after the transient transfection of a STAT5a expression construct in four melanoma cell lines, HTZ19d, Mel Im, Mel Ju and Mel Juso. The results are standardized to β-actin. pcDNA-transfected control cells were set as 1. (C) qRT-PCR for the expression level of LIF mRNA after the transient transfection of a TAM67 dominant negative AP-1 expression construct in four melanoma cell lines. The results are standardized to β-actin. pcDNA-transfected control cells were set as 1.

Interestingly, hypoxia-inducing reagents DP and DFX induced a detectable impact on LIF expression ( Fig. 3 A). The activity of both chemicals was confirmed by the qRT-PCR analysis of the known hypoxia-inducible gene ANGPTL4, which was up-regulated in the four melanoma cell lines analyzed ( Fig. 3 B). To confirm the involvement of HIF in this regulation, we transfected the melanoma cell lines with a dominant negative HIF construct. Again, after this transfection, we found that Angptl4 was down-regulated ( Fig. 3 C), and we also detected a strong and significant reduction of LIF mRNA expression ( Fig. 3 D). Furthermore, to focus on the participation of HIF-1α, we used specific siRNA against this HIF subunit and confirmed knock-down by qRT-PCR ( Fig. 3 E). In Fig. 3 F, the regulation of LIF by HIF-1α was revealed by qRT-PCR. In summary, hypoxia and, more precisely, HIF-1α are new players responsible for LIF over-expression in melanoma cells.


Fig. 3 HIF-1α as a regulator of LIF. (A; B) qRT-PCR for the expression level of LIF and ANGPTL4 mRNA after the incubation of four melanoma cell lines with chemical inducers of HIF: desferrioxamine (DFX) and 2, 2 dipyridyl (DP). Control incubated cells were set as 1. The results are standardized to β-actin. (C; D) qRT-PCR for the expression level of ANGPTL4 and LIF mRNA after the transient transfection of a dominant negative HIF expression construct. pcDNA control transfected cells were set as 1. The results are standardized to β-actin. (E; F) qRT-PCR for the expression level of HIF-1α and LIF mRNA after the transient transfection of specific siRNA against HIF-1α. siRNA control transfected cells were set as 1. The results are standardized to β-actin.

Functional relevance of LIF

LIF expression has been implicated to have relevance for stem cell characteristics and plays an important role in many different cancers, such as lymphoma, melanoma, breast, colon, pancreatic, lung, bladder, epidermoid and squamous cell carcinoma. The detailed cellular function of LIF in these cancers remains unclear thus far. To analyze the functional role of LIF, we silenced it in three metastatic melanoma cell lines by transient siRNA transfection. Successful knock-down was achieved and confirmed at the mRNA and protein levels ( Figs. 4 A and B). First, we analyzed the consequences of this knock-down on the expression level of several embryonic stem cell transcription factors; SOX2, NANOG, OCT3/4 and GBX2 were regulated by LIF ( Fig. 4 C). Interestingly, we found lower LIF protein levels in the de-differentiated MBrcs than in NHEMs at day 5 and at day 14 of cultivation (see Materials and methods section) (Supplementary Fig. 1A). To study LIF's functional role, cell proliferation, attachment, migration and anchorage-independent growth were analyzed after LIF silencing. The cells with diminished LIF expression showed a decline in the proliferation rate detected by RTCA measurement over 96 h ( Fig. 4 D). The proliferation rate was approximately 3 times lower than in the siRNA control (ctrl) transfected cells. Additionally, cyclinD1 was down-regulated after LIF silencing, which is in accord with the proliferative behavior of the cells ( Fig. 4 E). Furthermore, the LIF siRNA-transfected cells showed reduced attachment in the RTCA system after 1 h of measurement ( Fig. 4 F). The capacity to migrate in the Boyden chamber assay was also significantly reduced in the tested cell lines after knock-down of LIF. The cells with reduced LIF expression had approximately 40% of the migration capacity of the control siRNA transfectants after 48 h ( Fig. 4 G). The ability to form colonies in soft agar was also reduced. For this assay, we analyzed HTZ19d and Mel Ju. HTZ19d showed a significantly diminished colony diameter in LIF siRNA-transfected cells compared to control-transfected cells. Mel Ju showed this same effect but not to a significant extent (Supplementary Fig. 2A). We further incubated NHEMs with recombinant (rc) LIF. The endogenous expression of LIF was not affected by recombinant LIF (Supplementary Fig. 3A). We focused on migration and invasion in a Boyden chamber system to assess the functional relevance of LIF in primary cells. In this assay, the treatment of NHEMs with recombinant LIF led to a stimulation of NHEMs to migrate and invade the artificial basement membrane simulated by a polycarbonate filter with pores ( Figs. 4 H and I).


Fig. 4 Consequences of LIF silencing in melanoma and treatment of NHEMs with recombinant LIF. (A; B) qRT-PCR and Western blot analysis for the expression level of LIF after siRNA treatment. The knock-down was shown for three melanoma metastases. The results are standardized to β-actin for qRT-PCR, and GAPDH is the loading control for the Western blots. (C) qRT-PCR for the expression level of the genes SOX2, NANOG, OCT3/4 and GBX2 after the silencing of LIF. siRNA control transfected cells were set as 1. The results are standardized to β-actin. (D) The analysis of the proliferation rate of three melanoma cell lines after LIF silencing. (E) qRT-PCR analysis for the expression level of Cyclin D1 mRNA after the silencing of LIF with specific siRNA. The results are standardized to β-actin. The expression in NHEMs was set as 1. (F, G) The attachment and migration potential of LIF siRNA transfected melanoma cell lines were analyzed using the xCELLigence System ( Materials and methods section). (H, I) Normal human epidermal melanocytes (NHEMs) were treated with 100 ng recombinant LIF (rc LIF) for 24 h, and the cells were used in migration (H) and invasion (I) assays.

LIF targets

The commonly described signaling pathways that are regulated by LIF are controlled by STAT3 ( Gearing et al., 1993 ). Furthermore, a correlation with integrin αvβ1 was found in melanoma ( Heymann et al., 1995a ). We analyzed these molecules in our melanoma cell lines but could not find a strong phosphorylation of STAT3 in the melanoma cell lines used (Supplementary Fig. 4A). Therefore, our conclusion is that the high expression rates of LIF do not affect STAT3 signaling in our cell lines. Furthermore, integrin αv subunit, integrin β1 subunit and integrin β3 subunit were not regulated by LIF, based on their mRNA levels (data not shown). During the search for alternative signaling pathways of LIF, we surprisingly found that BMP4 and BMP7 were regulated by LIF. Diminished LIF expression led to the reduced expression of both BMP molecules, as determined at the mRNA level in four melanoma cell lines ( Figs. 5 A and B).


Fig. 5 LIF regulates BMP expression. (A; B) qRT-PCR for the expression level of BMP4 (A) and BMP7 (B) mRNA after LIF siRNA transfection. siRNA control transfected cells were set as 1. The results are standardized to β-actin.


LIF (leukemia inhibitory factor) is expressed at elevated levels in malignant cells from a wide variety of tissues. Paglia and colleagues showed the expression of LIF in melanoma cell lines ( Paglia et al., 1995 ). Our data confirm the expression of LIF in melanoma at the mRNA and protein levels. We also showed LIF expression in human melanoma tissue.

However, the mechanisms for enhanced LIF expression in melanoma are as yet unknown. Therefore, we analyzed possible stimuli for LIF induction and excluded STAT5a and AP-1. Interestingly, we found hypoxia as one reason for up-regulated LIF expression. It was known that melanoma cells and melanocytes have an unusual antioxidative system ( Wittgen and van Kempen, 2007 ). The melanocytes' primary function is delivering melanin in melanosomes to keratinocytes, resulting in protection against the harmful effects of UV radiation. Within the melanocytes, the synthesis of melanin results in the generation of hydrogen peroxide, and if inappropriately processed, the generation of hydroxyl radicals and other reactive oxygen species (ROS), which are part of an especially important hypoxia-inducible system in melanoma cells ( Meyskens et al., 2001 ). The adaption to reduced oxygen levels by hypoxia is usually regulated by HIF (hypoxia inducible factor) molecules that accumulate in the nucleus. Additionally, HIF activity is increased in malignant melanoma cells under normoxic conditions, in contrast to other tumor types ( Kuphal et al., 2010 ). Here, we experimentally increased HIF expression by the chemical inducers desferrioxamine (DFX) and 2, 2 dipyridyl (DP) and achieved a further induction of LIF expression in melanoma. HIF-1α is the master regulator of the cellular and systemic adaptation to hypoxia and is recognized as a key factor in cancer development ( Kuphal et al., 2010 ). Therefore, we focused on the HIF-1α subunit and showed that HIF-1α is specifically involved in LIF induction.

We were further interested in signaling downstream of LIF and found the transcription factors SOX2, NANOG, OCT3/4 and GBX2 regulated by LIF. In addition to their functions as classical stem cell proteins, these molecules were also described as important transcription factors in cancer (Gao and Isaacs, 1996 and Ling et al, 2012). Therefore, we classify LIF as a typical oncogene of melanoma with an auxiliary function as a growth factor for the maintenance of ESC pluripotency. This conclusion was underscored by the lower expression rates of LIF in de-differentiated MBrcs, which are neural crest precursors of NHEMs. LIF appears to be an important cytokine for normal keratinocyte growth and wound healing, and it is reported to be involved in the proliferation of skin tumors, e.g., squamous cell carcinoma (SSC) (Hu et al, 2000, Paglia et al, 1996, and Szepietowski et al, 2001). The involvement of LIF in regulating proliferative processes was confirmed by our data showing reduced proliferation measured by the RTCA method and reduced cyclinD1 expression after LIF siRNA treatment of three melanoma cell lines. Furthermore, it was reported that LIF increases attachment through an increase in integrin αvβ1 and intercellular adhesion molecule-1 expression in keratinocytes (Heymann et al, 1995a and Heymann et al, 1995b). The attachment of melanoma cells was also affected by LIF, but we could not detect a regulation of integrin αv, integrin β1 or integrin β3. We further analyzed migration and colony formation capacity under the influence of LIF silencing and found a diminished migration rate and smaller colonies in soft agar. Using NHEMs in Boyden chamber assays, treatment with recombinant LIF led to the elevated migration and invasion rates of these primary cells. We conclude that LIF has an influence on melanoma at early stages.

Searching for new target molecules of LIF signaling, we found (bone morphogenetic proteins) BMP4 and BMP7 regulated by LIF. Both molecules could be part of the reported phenomenon that LIF is involved in melanoma-induced bone metastasis or bone destruction ( Maruta et al., 2009 ). BMP4 and BMP7 were reported to be over-expressed in melanoma ( Rothhammer et al., 2005 ). However, any correlation between LIF, BMP4, BMP7 and bone metastasis is speculative because the mechanism responsible for bone destruction in melanoma is still poorly understood.

Furthermore, LIF and BMPs are involved in regulating the pluripotency of embryonic stem cells (ESCs). At first glance, it was puzzling to find that while mouse ES cells required LIF and serum in their culture medium for the maintenance of pluripotency and self-renewal, human ES cells did not need LIF, but instead must be grown with activin and FGF factors to maintain their un-differentiated state (Greber et al, 2008 and Xu et al, 2001). However, newer findings demonstrated that human ES cells express a relatively low level of LIF signaling components (LIFR, JAK and STAT) and a high level of suppressor of cytokine signaling (SOCS), which together keep the LIF expression at a lower level ( Wei et al., 2005 ). Frequently, LIF cooperates with additional molecules (in the past these were provided in serum) to influence cellular self-renewal. One example is Pramel7 (preferentially expressed antigen in melanoma-like 7), a novel factor crucial for the maintenance of LIF-dependent self-renewal ( Casanova et al., 2011 ). A further described cooperation was demonstrated for BMP4 and LIF. Under serum-free culture conditions, LIF alone stimulated the neural differentiation of ES cells. However, the addition of BMP4 was able to maintain the un-differentiated state of mouse ES cells, even in the absence of serum ( Ying et al., 2003 ). Therefore, we speculate that it is necessary for melanoma cells to express LIF and BMP4 (alternatively BMP7) in a co-stimulatory manner to keep melanoma cells un-differentiated.

Finally, we could not find the classically described signaling pathways mediated by LIF in melanoma. In fact, we analyzed the regulation of phospho-STAT3, phospho-p65, phospho-AKT and phospho-ERK signaling for their dependence on LIF. It was reported that STAT3 primarily and the additionally mentioned kinases are targets of LIF (Auernhammer and Melmed, 2000, Gearing et al, 1993, Oh et al, 1998, and Thoma et al, 1994). Interestingly, STAT3 is not detectably phosphorylated or active in our analyzed melanoma cell lines. We know that these findings are not in accord with the literature ( Niu et al., 2002 ), but we excluded STAT3 as the main mediator for LIF signaling in melanoma. Additionally, LIF knock-down did not influence the phosphorylation of p65, AKT or ERK, which we analyzed by Western blot analysis using antibodies against the phosphorylated forms of the proteins (data not shown).

In summary, we found that LIF was primarily regulated by hypoxia, specifically through the molecule HIF-1α. Furthermore, LIF signaling regulates transcription factors such as SOX2, NANOG, OCT3/4 and GBX2, which maintain the un-differentiated state and pluripotency of ESCs and most likely also have important oncogenic functions in melanoma cells. Additionally, cancer characteristics, such as proliferation, attachment and migration, are positively stimulated by LIF expression, and we can define BMP4 and BMP7 as new target molecules of LIF signaling.

Conflict of interest statement

The authors declare no conflict of interest.


We thank Anne Rascle (Institute of Immunology, University of Regensburg) and Christina Warnecke (Department of Nephrology and Hypertension, University Clinic Erlangen, Germany) for their experimental support.

Appendix A. Supplementary data


Download file

Supplementary Fig. 1: Expression of LIF in MBrcs and NHEMs. (A) Melanoblast-related cells (MBrcs) were cultivated for 5 and for 14 days via the de-differentiation of NHEMs in a special melanoblast-medium ( Materials and methods section). Western blot analysis showed the expression level of LIF compared to NHEMs and MBrcs of the same cell fraction. β-actin was used as a loading control. Supplementary Fig. 2: Consequences LIF silencing in melanoma. (A) The analysis of colony formation after LIF silencing in HTZ19d and Mel Ju using specific siRNA against LIF. Supplementary Fig. 3: Treatment of NHEMs with recombinant (rc) LIF. (A) qRT-PCR for the expression level of LIF mRNA in normal human epidermal melanocytes (NHEMs) after treatment with recombinant (rc) LIF. The results are standardized to β-actin. Supplementary Fig. 4: STAT3 expression in melanoma cell lines. (A) Western blot analysis of the protein level of STAT3 in seven melanoma cell lines and NHEMs. The phosphorylation of STAT3 was also detected by Western blot analysis. β-actin was used as a loading control.


  • Alexander et al., 1994 H.R. Alexander, K.G. Billingsley, M.I. Block, D.L. Fraker. D-factor/leukaemia inhibitory factor: evidence for its role as a mediator in acute and chronic inflammatory disease. Cytokine. 1994;6(6):589-596
  • Auernhammer and Melmed, 2000 C.J. Auernhammer, S. Melmed. Leukemia-inhibitory factor-neuroimmune modulator of endocrine function. Endocrine Reviews. 2000;21(3):313-345
  • Bosserhoff et al., 2011 A.K. Bosserhoff, L. Ellmann, S. Kuphal. Melanoblasts in culture as an in vitro system to determine molecular changes in melanoma. Experimental Dermatology. 2011;20(5):435-440
  • Brown et al., 1994 P.H. Brown, T.K. Chen, M.J. Birrer. Mechanism of action of a dominant-negative mutant of c-Jun. Oncogene. 1994;9(3):791-799
  • Casanova et al., 2011 E.A. Casanova, O. Shakhova, S.S. Patel, I.N. Asner, P. Pelczar, F.A. Weber, et al. Pramel7 mediates LIF/STAT3-dependent self-renewal in embryonic stem cells. Stem Cells. 2011;29(3):474-485
  • Cook et al., 2003 A.L. Cook, P.D. Donatien, A.G. Smith, M. Murphy, M.K. Jones, M. Herlyn, et al. Human melanoblasts in culture: expression of BRN2 and synergistic regulation by fibroblast growth factor-2, stem cell factor, and endothelin-3. The Journal of Investigative Dermatology. 2003;121(5):1150-1159
  • Gao and Isaacs, 1996 A.C. Gao, J.T. Isaacs. Expression of homeobox gene-GBX2 in human prostatic cancer cells. Prostate. 1996;29(6):395-398
  • Gearing et al., 1987 D.P. Gearing, N.M. Gough, J.A. King, D.J. Hilton, N.A. Nicola, R.J. Simpson, et al. Molecular cloning and expression of cDNA encoding a murine myeloid leukaemia inhibitory factor (LIF). EMBO Journal. 1987;6(13):3995-4002
  • Gearing et al., 1993 D.P. Gearing, T. Druck, K. Huebner, J. Overhauser, D.J. Gilbert, N.G. Copeland, et al. The leukemia inhibitory factor receptor (LIFR) gene is located within a cluster of cytokine receptor loci on mouse chromosome 15 and human chromosome 5p12–p13. Genomics. 1993;18(1):148-150
  • Greber et al., 2008 B. Greber, H. Lehrach, J. Adjaye. Control of early fate decisions in human ES cells by distinct states of TGFbeta pathway activity. Stem Cells and Development. 2008;17(6):1065-1077
  • Heymann et al., 1995a D. Heymann, J. Harb, S. Ringeard, F. Blanchard, D. Lassort, S. Raher, et al. Upmodulation of alpha v beta 1 integrin expression on human tumor cells by human interleukin for DA cells/leukemia inhibitory factor and oncostatin M: correlation with increased cell adhesion on fibronectin. Journal of Cellular Biochemistry. 1995;58(3):305-314
  • Heymann et al., 1995b D. Heymann, A. Godard, S. Raher, S. Ringeard, D. Lassort, F. Blanchard, et al. Human interleukin for DA cells/leukemia inhibitory factor and oncostatin M enhance membrane expression of intercellular adhesion molecule-1 on melanoma cells but not the shedding of its soluble form. Cytokine. 1995;7(2):111-117
  • Hu et al., 2000 J. Hu, S. Ono, H. Katayama, T. Imai, N. Shimizu, H. Nakagawa. Leukemia inhibitory factor induces epidermal hyperplasia in patients with amyotrophic lateral sclerosis. The Journal of Investigative Dermatology. 2000;115(3):486-492
  • Humphrey et al., 2004 R.K. Humphrey, G.M. Beattie, A.D. Lopez, N. Bucay, C.C. King, M.T. Firpo, et al. Maintenance of pluripotency in human embryonic stem cells is STAT3 independent. Stem Cells. 2004;22(4):522-530
  • Kamohara et al., 1994 H. Kamohara, K. Sakamoto, T. Ishiko, S. Mita, Y. Masuda, T. Abe, et al. Human carcinoma cell lines produce biologically active leukemia inhibitory factor (LIF). Research Communications in Molecular Pathology and Pharmacology. 1994;85(2):131-140
  • Kawashima et al., 2006 T. Kawashima, Y.C. Bao, Y. Nomura, Y. Moon, Y. Tonozuka, Y. Minoshima, et al. Rac1 and a GTPase-activating protein, MgcRacGAP, are required for nuclear translocation of STAT transcription factors. The Journal of Cell Biology. 2006;175(6):937-946
  • Knight, 2001 D. Knight. Leukaemia inhibitory factor (LIF): a cytokine of emerging importance in chronic airway inflammation. Pulmonary Pharmacology & Therapeutics. 2001;14(3):169-176
  • Kuphal et al., 2010 S. Kuphal, A. Winklmeier, C. Warnecke, A.K. Bosserhoff. Constitutive HIF-1 activity in malignant melanoma. European Journal of Cancer. 2010;46(6):1159-1169
  • Ling et al., 2012 G.Q. Ling, D.B. Chen, B.Q. Wang, L.S. Zhang. Expression of the pluripotency markers Oct3/4, Nanog and Sox2 in human breast cancer cell lines. Oncology Letters. 2012;4(6):1264-1268
  • Maruta et al., 2009 S. Maruta, S. Takiguchi, M. Ueyama, Y. Kataoka, Y. Oda, M. Tsuneyoshi, et al. A role for leukemia inhibitory factor in melanoma-induced bone metastasis. Clinical & Experimental Metastasis. 2009;26(2):133-141
  • McKenzie and Szepietowski, 2004 R.C. McKenzie, J. Szepietowski. Cutaneous leukemia inhibitory factor and its potential role in the development of skin tumors. Dermatologic Surgery. 2004;30(2 Pt 2):279-290
  • Meyskens et al., 2001 F.L. Meyskens Jr., P. Farmer, J.P. Fruehauf. Redox regulation in human melanocytes and melanoma. Pigment Cell Research. 2001;14(3):148-154
  • Nakai-Futatsugi and Niwa, 2013 Y. Nakai-Futatsugi, H. Niwa. Transcription factor network in embryonic stem cells: heterogeneity under the stringency. Biological and Pharmaceutical Bulletin. 2013;36(2):166-170
  • Niu et al., 2002 G. Niu, T. Bowman, M. Huang, S. Shivers, D. Reintgen, A. Daud, et al. Roles of activated Src and Stat3 signaling in melanoma tumor cell growth. Oncogene. 2002;21(46):7001-7010
  • Niwa et al., 2009 H. Niwa, K. Ogawa, D. Shimosato, K. Adachi. A parallel circuit of LIF signalling pathways maintains pluripotency of mouse ES cells. Nature. 2009;460(7251):118-122
  • Oh et al., 1998 H. Oh, Y. Fujio, K. Kunisada, H. Hirota, H. Matsui, T. Kishimoto, et al. Activation of phosphatidylinositol 3-kinase through glycoprotein 130 induces protein kinase B and p70 S6 kinase phosphorylation in cardiac myocytes. Journal of Biological Chemistry. 1998;273(16):9703-9710
  • Paglia et al., 1995 D. Paglia, A. Oran, C. Lu, R.S. Kerbel, D.N. Sauder, R.C. McKenzie. Expression of leukemia inhibitory factor and interleukin-11 by human melanoma cell lines: LIF, IL-6, and IL-11 are not coregulated. Journal of Interferon and Cytokine Research. 1995;15(5):455-460
  • Paglia et al., 1996 D. Paglia, S. Kondo, K.M. Ng, D.N. Sauder, R.C. McKenzie. Leukaemia inhibitory factor is expressed by normal human keratinocytes in vitro and in vivo. British Journal of Dermatology. 1996;134(5):817-823
  • Rathjen et al., 1990 P.D. Rathjen, S. Toth, A. Willis, J.K. Heath, A.G. Smith. Differentiation inhibiting activity is produced in matrix-associated and diffusible forms that are generated by alternate promoter usage. Cell. 1990;62(6):1105-1114
  • Rothhammer et al., 2005 T. Rothhammer, I. Poser, F. Soncin, F. Bataille, M. Moser, A.K. Bosserhoff. Bone morphogenic proteins are overexpressed in malignant melanoma and promote cell invasion and migration. Cancer Research. 2005;65(2):448-456
  • Salas et al., 2011 E.M. Salas, M.J. Garcia-Barchino, S. Labiano, M. Shugay, M. Perez-Encinas, C. Quinteiro, et al. LIF, a Novel STAT5-Regulated gene, is aberrantly expressed in myeloproliferative neoplasms. Genes Cancer. 2011;2(5):593-596
  • Simpson et al., 1988 R.J. Simpson, D.J. Hilton, E.C. Nice, M.R. Rubira, D. Metcalf, D.P. Gearing, et al. Structural characterization of a murine myeloid leukaemia inhibitory factor. European Journal of Biochemistry. 1988;175(3):541-547
  • Szepietowski et al., 2001 J.C. Szepietowski, C. Walker, D.B. McKenna, J.A. Hunter, R.C. McKenzie. Leukaemia inhibitory factor and interleukin-8 expression in nonmelanoma skin cancers. Clinical and Experimental Dermatology. 2001;26(1):72-78
  • Thoma et al., 1994 B. Thoma, T.A. Bird, D.J. Friend, D.P. Gearing, S.K. Dower. Oncostatin M and leukemia inhibitory factor trigger overlapping and different signals through partially shared receptor complexes. Journal of Biological Chemistry. 1994;269(8):6215-6222
  • Turnley and Bartlett, 2000 A.M. Turnley, P.F. Bartlett. Cytokines that signal through the leukemia inhibitory factor receptor-beta complex in the nervous system. Journal of Neurochemistry. 2000;74(3):889-899
  • Wei et al., 2005 C.L. Wei, T. Miura, P. Robson, S.K. Lim, X.Q. Xu, M.Y. Lee, et al. Transcriptome profiling of human and murine ESCs identifies divergent paths required to maintain the stem cell state. Stem Cells. 2005;23(2):166-185
  • Wittgen and van Kempen, 2007 H.G. Wittgen, L.C. van Kempen. Reactive oxygen species in melanoma and its therapeutic implications. Melanoma Research. 2007;17(6):400-409
  • Xu et al., 2001 C. Xu, M.S. Inokuma, J. Denham, K. Golds, P. Kundu, J.D. Gold, et al. Feeder-free growth of undifferentiated human embryonic stem cells. Nature Biotechnology. 2001;19(10):971-974
  • Ying et al., 2003 Q.L. Ying, J. Nichols, I. Chambers, A. Smith. BMP induction of Id proteins suppresses differentiation and sustains embryonic stem cell self-renewal in collaboration with STAT3. Cell. 2003;115(3):281-292


Institute of Pathology, Molecular Pathology, University of Regensburg, Regensburg, Germany

lowast Corresponding author at: Institute of Pathology, University Regensburg, Franz-Josef Strauss Allee 11, 93053 Regensburg, Germany. Fax: + 49 941 944 6602.

Grant support: The project was supported by the Deutsche Forschungsgemeinschaft (German Research Foundation) and by the Bayerisches Forschungsnetzwerk für Molekulare Biosysteme (Bavarian Research Network for Molecular Biosystems, BioSysNet).