Lipocalin 2: A Multifaceted Modulator of Human Cancer

Abstract

Lipocalin 2 (Lcn2), a member of the lipocalin family that carries small lipophilic ligands, has gained recent attention as both a potential biomarker and a modulator of human cancers. Here we describe recent findings of the functions of Lcn2 in breast cancer and the potential mechanisms that underlie its actions. Lcn2 has been shown to induce the epithelial to mesenchymal transition (EMT) in breast cancer cells and to promote breast tumor invasion. Estrogen receptor α may participate in the pathway that leads to Lcn2-induced EMT. Preliminary evidence also suggests that Lcn2 may be used as a potential non-invasive urinary biomarker of breast cancer. Elevated levels of Lcn2 have also been reported in other human cancers. The potential roles of Lcn2 in some other epithelial tumors as well as leukemia are also reviewed and discussed here.

Introduction

Lipocalin 2 (Lcn2) is a member of the lipocalin protein family. Lipocalins are typically small (160–180 amino acids) extracellular proteins which share limited similarity in primary sequences but do share a highly conserved tertiary structure, 8-stranded anti-parallel β-barrel. Lipocalins usually bind to small lipophilic ligands within the hydrophobic cavity of the β-barrel and transport them inside a cell via receptor-mediated endocytosis. The list of lipocalin ligands includes retinoic acid, progesterone and prostaglandin.¹ As reflected by their ligands, lipocalins perform diversified functions and are involved in many important processes, including enzymatic synthesis, immunomodulation, angiogenesis and regulation of metabolic homeostasis.^2–4 Some lipocalins are biochemical markers of human diseases including orosomucoid (α₁-acid glycoprotein), Protein HC (α₁-microglobulin), apolipoprotein D, retinol-binding protein, complement C8 and prostaglandin D synthase.⁵

Lcn2 (also referred to as NGAL, neutrophil gelatinase-assicated lipocalin) is a ~25 kDa protein composed of 178 amino acids. It was first identified as part of the matrix metalloproteinase-9 (MMP-9) complex in human neutrophils.^{6, 7} The role(s) of Lcn2 in the MMP-9/Lcn2 complex remained unknown until our group demonstrated that the binding of Lcn2 to MMP-9 protected the enzyme from autodegradation thereby preserving its activity.⁸ The protective role of Lcn2 towards MMP-9 suggests that it may be involved in extracellular matrix remodeling.

The first identified ligands for Lcn2 are bacterial catecholate-type ferric siderophores, such as enterobactin, and Lcn2 has been identified as a potent bacteriostatic agent.⁹ Lcn2 inhibits bacterial growth by sequestering the iron-binding bacterial siderophores and blocking the bacterial access to iron. In keeping with these findings, mice deficient in Lcn2 showed increased susceptibility to bacterial infections.^10–12 The localization of Lcn2 expression also suggests a role in the innate immune responses. In humans, high levels of Lcn2 are found in tissues that are prone to the insult of microorganisms, such as trachea, lung and stomach,¹³ as well as at sites of inflammation.^{14, 15}

Lcn2 may also bind to a mammalian counterpart of a bacterial ferric siderophore and participate in mammalian iron metabolism. It has been shown to induce the differentiation of mesenchymal progenitor cells into epithelial tubules during kidney development by way of delivering iron inside the cells.¹⁶ Furthermore, iron-loaded Lcn2 rescued the kidney from ischemia-reperfusion injury whereas blockade of the iron-binding activity abrogated the protective function of Lcn2.¹⁷ Iron trafficking may also mediate other Lcn2-related functions. Apoptosis of hematopoietic cells induced by IL-3 deprivation was found to be mediated by Lcn2 whose transcription is activated in the absence of IL-3.¹⁸ It was later demonstrated that Lcn2 induces apoptosis by affecting the intracellular iron content.¹⁹ In contrast to the iron-loaded Lcn2 which delivers iron inside the cell and increase the iron content, empty Lcn2, when taken in by the cell, binds to iron intracellularly and depletes the iron pool, which further induces the pro-apoptotic protein Bim. Effects on iron content may also mediate the apoptotic activity of Lcn2 in other systems.²⁰

Interestingly, in contrast to its pro-apoptotic activity, Lcn2 has also been reported to be a survival factor. Lcn2 protected thyroid carcinoma cells from apoptosis induced by serum deprivation and silencing of Lcn2 reduced thyroid tumor growth.²¹ Similarly, overexpression of Lcn2 reduced, whereas silencing of Lcn2 increased, the apoptosis of lung adenocarcinoma A549 cells in the presence of a PDK1 (phosphoinositide-dependent kinase 1) inhibitor.²² Modulation of the intracellular iron content has also been suggested as the means by which Lcn2 promotes survival.²¹

Lipocalin 2 and breast cancer

Lcn2 has been associated with breast cancer progression. In a study analyzing the plasma proteome, Lcn2 levels were increased in breast tumor-bearing transgenic mice compared to normal littermates and its levels further increased with tumor progression.²³ In human studies, Lcn2 expression was found to be associated with estrogen receptor-negative status in breast cancer cell lines and in breast cancer tissues.^{24, 25} The correlation was observed at the protein level as well.²⁶ Lcn2 levels in breast cancer tissue also strongly correlate with characteristics that are associated with poor prognosis, including poor histologic grade, lymph node metastasis and high proliferation.²⁶ Not surprisingly, it has also been shown to be an independent prognostic marker for decreased survival.²⁶

Results from our laboratory document that Lcn2 protein levels are elevated in Stage I-III breast cancer samples compared to normal breast tissues (Fig.1, also see ref. 27). Importantly, we have also found increased Lcn2 levels in the urine samples from metastatic breast cancer patients compared to samples from healthy controls, suggesting that Lcn2 may be a potential non-invasive biomarker for advanced breast cancer. We also initiated a pilot study of longitudinal sampling and analysis of the urine from four breast cancer patients who were diagnosed as having breast cancer that had progressed to a new stage at the times of sampling (Table 1). In the four patients monitored, urinary Lcn2 levels increased with progression of the disease (Figure 2). These data suggest that urinary Lcn2 maybe useful in predicting disease status and warrants further investigation.

Fig. 1.

Representative microscopic images of Lcn2 staining in normal breast tissue (A) and invasive breast cancer tissue (B). Brown staining indicates presence of Lcn2. Scale bar, 20 µm.

Table 1.

Summary of breast cancer patient diagnoses in the longitudinal study

Patient	Age	Sample #	Collection date	Diagnosis
M676	47	1	6/10/2002	DCIS
		2	5/9/2005	additional DCIS
M764	48	1	8/9/2002	DCIS
		2	5/19/2003	new lesion on mammogram
M1077	61	1	11/17/2003	normal
		2	3/1/2004	DCIS
		3	8/13/2004	DCIS
M1195	46	1	6/7/2004	invasive breast cancer (Stage I)
		2	n/a	n/a
		3	6/27/2005	inflammatory breast cancer (Stage III)

Note: DCIS, ductal carcinoma in situ; n/a, information not available

Fig. 2.

Longitudinal urinary Lcn2 levels (normalized to total protein concentration) as analyzed by ELISA of four patients with progressed breast cancer.

If Lcn2 increases with breast cancer progression, is it a simple correlation or is Lcn2 an active player that contributes to the process? Data from our group supports the latter. We have found that Lcn2 induces the epithelial to mesenchymal transition (EMT) in human breast cancer cells.²⁷ EMT is one of the key mechanisms underlying tumor invasion and metastasis.²⁸ EMT induces a more mobile phenotype in tumor cells, thereby facilitating their invasion into the local extracellular matrix, intravasation/extravasation of blood vessels and invasion and settlement at secondary organ sites. There are several key transcription factors involved in this important mechanism, including Snail, Slug, Twist and SIP1. The activation of these factors can repress E-cadherin transcription and the epithelial phenotype and induce EMT.

We observed that normally non-aggressive MCF-7 human breast cancer cells that were induced to overexpress Lcn2 switched from typical epithelial morphology to a mesenchymal-like morphology. The epithelial marker E-cadherin, whose presence indicates strong cell-cell interaction, was repressed, whereas mesenchymal markers, such as vimentin and fibronectin, were induced by exogenous Lcn2 expression. Consistent with the changes in morphology and marker expression profile, cell migration and invasion as analyzed with in vitro assays were significantly increased in Lcn2-overexpressing cells.

Furthermore, when MCF-7 cells overexpressing Lcn2 were orthotopically injected into the mammary fat pads of nude mice, they exhibited characteristics that are commonly associated with aggressive and invasive breast tumors. In contrast to controls, tumors formed by Lcn2-expressing cells were poorly differentiated and grew in a disorganized manner. Secondly, increased local invasion into the neighboring tissues and metastasis into lymph nodes were also observed with Lcn-2 expressing tumors.

Not only is Lcn2 sufficient to transform a non-aggressive breast cancer cell line into an aggressive one, it also contributes to the migratory behavior of aggressive breast cancer cells.²⁷ MDA-MB-231 cells express markedly higher Lcn2 levels compared to non-aggressive MCF-7 cells. These cells also display the mesenchymal characteristics with the scattering morphology and strong expression of mesenchymal markers. When Lcn2 was repressed in these cells via siRNA, cell migration was significantly reduced, accompanied by a decreased expression of the mesenchymal marker vimentin. Cells also changed into a clustering morphology. Knockdown of Lcn2, however, did not fully switch the MDA-MB-231 cells to an epithelial phenotype since E-cadherin expression was not affected, suggesting that other mechanisms may also take part in maintaining the mesenchymal phenotype of these cells.

Taken together, results from our in vitro, in vivo and human studies, together with results from other correlative studies of Lcn2 in breast cancer, strongly suggest that Lcn2 promotes breast cancer progression. Furthermore, our findings in human breast cancer cells are corroborated by another study using murine breast cancer cells. When it was overexpressed in 4T1 murine mammary cancer cells, Lcn2 significantly increased cell migration and invasion in vitro and enhanced lung metastasis in vivo.²⁹

We next studied the mechanisms that mediate Lcn2-induced EMT. Interestingly, we observed reduced estrogen receptor α (ERα) levels with Lcn-2 overexpression, which was accompanied by reduced response to estrogen treatment. This reduction in ERα expression was detected at both the transcript and protein levels. This suggests that Lcn-2 may be able to manipulate the expression of this hormone receptor, which is a significant finding considering the critical role of ER in breast cancer pathology and treatment. We also found evidence that ERα may play a part in Lcn2-induced EMT. Signaling through ERα has been shown to be necessary for the maintenance of E-cadherin expression and the epithelial phenotype in breast cancer cells by repressing the transcription factor Snail.³⁰ Absence of ERα leads to a loss of E-cadherin and the epithelial phenotype. In Lcn2-expressing cells we found an inverse correlation between ERα and Slug, one of the key EMT transcription factors and a close family member of Snail. Inducing ERα expression in Lcn2-expressing cells partially reversed the EMT, with increased E-cadherin and reduced vimentin and fibronectin, and repressed Slug, suggesting that Lcn2-induced EMT is mediated, at least in part, through the ERα/Slug axis.

Lcn2 can form a complex with MMP-9 and protect this enzyme from degradation.⁸ Since MMP-9 has been implicated in cancer progression, we asked whether it might be contributing to Lcn2-induced EMT? We found that monomeric Lcn2 is the predominant form of this protein secreted by Lcn2-expressing cells as analyzed via immunoblot (unpublished data). Only using a more sensitive zymographic assay can small amounts of MMP-9/Lcn2 complex be detected in the conditioned media from Lcn2-expressing cells.³¹ Although the small amount of active MMP-9 might contribute to the increased migration and invasion observed with the Lcn2-expressing cells, we believe that, in this system, EMT is primarily induced through MMP-9-independent pathways.

Several important questions still remain to be answered. One question is whether extracellular Lcn2 represses ERα and induces EMT through a cell-surface receptor. To date, two receptors have been reported for Lcn2, solute carrier family 22 member 17 (SLC22A17, also refer to as BOCT, brain-type organic cation transporter) and megalin. SLC22A17 was first discovered in leukocytic cell lines as a cell-surface receptor that can mediate the apoptotic effect of Lcn2.^{19, 32} Whether SLC22A17 specifically binds to Lcn2 and mediates its function in cells other than leukocytes remains to be elucidated. Although it has been shown to bind to Lcn2 with high affinity and mediate Lcn2 intake,³³ megalin is not a Lcn2-specific receptor. It will be interesting to learn whether one of these two receptors, or other yet unidentified ones, mediate the actions of this extracellular protein in breast cancer. Lcn2 may also exert some of its actions intracellularly even before it is secreted. We have shown that treatment with recombinant Lcn2 can repress E-cadherin but is unable to induce mesenchymal markers in MCF-7 cells,²⁷ suggesting that extracellular Lcn2 only exerts part of its EMT-inducing actions. It will be important to investigate the intracellular binding-partner(s) of Lcn2 as well.

Another remaining question is whether the iron-transporting activity is involved in Lcn2-induced EMT. As discussed in the Introduction, Lcn2 can induce the differentiation of kidney tubules from mesenchymal progenitor cells by transporting iron inside the cells.¹⁶ Iron-empty Lcn2 is also able to shuttle iron outside of the cells.¹⁹ Whether changes in intracellular iron could play a role in Lnc2-induced EMT remains to be investigated.

It is noteworthy that Lcn2’s actions in breast cancer may also be modulated by the presence of oncogenes. In one study using 4T1 murine mammary cancer cells, it was observed that the introduction of constitutively active H-ras into these cells induced EMT and overexpression of Lcn2 in the ras-transformed cells reversed the EMT and inhibited invasion and metastasis.³⁴ It appears as if Lcn2 is associated with different, even opposite effects, in the presence of activated Ras compared to in the absence of Ras, the latter being the case in MCF-7 human breast cancer cells²⁷ as well as in untransformed 4T1 murine cells.²⁹

Lipocalin 2 in other epithelial tumors

Increased Lcn2 levels have been described in cancers of a variety of organs. The functions of Lcn2 in these cancers are not as well characterized. In spite of the elevated levels, it remains unknown as to whether Lcn2 modulates the progression of these cancers. In some cancers, the association between Lcn2 and tumor invasion/metastasis have also been noted.

Elevated Lcn2 expression has been reported in pancreatic cancer cell lines as well as in pancreatic cancer specimens.^35–39 Lcn2 protein has also been identified in the pancreatic juice proteome in patients with pancreatic adenocarcinoma.⁴⁰ A recent study further documented the association between Lcn2 and pancreatic cancer differentiation, with stronger Lcn2 staining in well- or moderately-differentiated samples than poorly-differentiated samples, suggesting Lcn2 as a potential biomarker of early disease.⁴¹ The observations that Lcn2 overexpression inhibits adhesion, invasion and angiogenesis of pancreatic cancer cell lines suggest that Lcn2 may be a suppressor of pancreatic cancer progression and may provide a potential explanation for relatively low levels of Lcn2 in advanced pancreatic cancer.⁴²

Lcn2 was identified as one of the most upregulated genes in ovarian serous papillary carcinomas compared to normal ovarian epithelium.⁴³ Results from two more recent studies further support the potential of Lcn2 as a marker for early-stage ovarian cancer. Highest Lcn2 levels were detected in the tissue and serum samples from patients with benign or low-grade disease.⁴⁴ The second study found a significant correlation between Lcn2 and differentiation as well, with the highest tissue or serum Lcn2 levels in low-grade well-differentiated cancer.⁴⁵ While an inverse correlation was documented between Lcn2 expression and epidermal growth factor (EGF)-induced EMT in ovarian cancer cells,⁴⁴ more thorough studies are warranted to define the functions of Lcn2 in ovarian cancer development and progression.

Lcn2 protein levels have also been shown to be increased in colorectal cancer tissues.^{14, 46–48} A gene expression profiling study further found Lcn2 transcript levels to be elevated in samples at all four Dukes’ stages (A–D) compared to normal colorectal specimens, with higher levels at stage A and D.⁴⁹ However, the functions of Lcn2 during colorectal cancer progression remain controversial. In one study, ectopic expression of Lcn2 in KM12SM cells, the metastatic version of KM12C human colon cancer cells, led to reduced Matrigel invasion in vitro and decreased experimental liver metastasis in vivo when these cells were injected into the spleen, suggesting that Lcn2 may serve as a suppressor of colon cancer metastasis.⁵⁰ In contrast, it has also been reported that the overexpression of Lcn2 in KM12C cells reduced cell-cell adhesion and increased cell invasion when these cells were cultured on collagen IV-coated surface. Consistently, silencing of Lcn2 in two other colon cancer cell lines reduced cell invasion.⁴⁸ This same study further demonstrated that Lcn2 may regulate cell migratory behavior by affecting the subcellular localization of Rac1 and E-cadherin, which appeared to be iron-dependent. The discrepancies between these studies could be due to the differences in the culture conditions (e.g. substratum) and the differences in the characteristics of the cell lines used (e.g. oncogene activation status). Interestingly, expression of both K-ras and N-ras was found to be increased in KM12SM cells compared to KM12C cells.⁵¹

Lcn2 levels were elevated in esophageal squamous cell carcinoma tissue samples compared to normal mucosa samples.⁵² In a gene profiling study of esophageal adenocarcinoma premalignant conditions, Lcn2 was found to be upregulated in Barrett esophagus (BE)-associated high-grade dysplasia compared to nondysplastic samples.⁵³ High-grade dysplasia is considered as a transitional stage in the progression from BE to esophageal adenocarcinoma.⁵⁴

Strong Lcn2 expression was also observed in human thyroid carcinomas.²¹ Knockdown of its expression in thyroid cancer cells impaired the growth capability of these cells both in soft agar and in mice, suggesting that Lcn2 may promote thyroid cancer survival. It has also been shown that Lcn2 promotes thyroid cancer cell survival by transporting iron inside the cells.²¹

Lipocalin 2 in leukemia

Lcn2 was found to be induced in an expression profiling study of chronic myeloid leukemia (CML).⁵⁵ Its expression in blood have been found to correlate with the expression of BCR-ABL, the fusion oncogene found in most CML patients, and its protein levels in serum have been found to be significantly increased in CML patients at diagnosis.⁵⁶ Further mechanistic studies indicate that Lcn2 may mediate BCR-ABL-induced tumorigenesis. The presence of the BCR-ABL oncoprotein induces the expression of Lcn2 in tumor cells, which in turn induces apoptosis in normal hematopoietic cells, therefore facilitating the expansion and invasion of tumor cells in bone marrow and spleen.^{57, 58} BCR-ABL-transformed cells themselves, despite elevated Lcn2 levels, are immune to Lcn2-induced apoptosis due to repressed Lcn2 receptor expression.¹⁹

Conclusion

Evidence from studies of various cancers indicate that Lcn2 may actively influence cancer progression by affecting two critical processes, one being apoptosis/survival and the other being migration/invasion. Whether Lcn2 modulates these processes through the same or distinct mechanisms remains unknown. Lcn2 may also exert either pro- or anti-cancer actions at different stages of a certain cancer or in different types of cancers. Moreover, Lcn2 has been shown to be a potential diagnostic and prognostic biomarker for a variety of human cancer. Further elucidation of the mechanisms that underlie the functions of Lcn2 will considerably contribute to the use of Lcn2 as both a diagnostic/prognostic biomarker and therapeutic target in human cancer.