Introduction
Breast cancer mortality is caused foremost by the spread of cancer cells within the host in a process called metastasis [
1]. Before tumor cells can metastasize, the tumor will need to invade, seek access to the lymphatic or vascular system and colonize the metastatic site [
2,
3]. Insights in this process will aid in the prevention of cancer metastasis and help improve prognosis.
An important characteristic of most solid tumors is the presence of hypoxic regions [
4‐
6]. Absent or inadequate vasculature within the tumor causes disruption of the supply of blood and consequentially an impaired delivery of oxygen and nutrients and an impaired removal of carbon dioxide and waste products. Several studies found low oxygen tension in tumors to be an adverse prognostic marker in different tumor types [
7‐
10]. In addition, endogenous hypoxia-related markers, such as carbonic anhydrase-IX, were also shown to negatively influence patient outcome in breast cancer [
11,
12]. Furthermore, hypoxic tumors were found to correlate with metastatic occurrences: patients with hypoxic primary tumors developed more metastases than patients with less hypoxic tumors [
7,
13‐
15]. Mechanistically, numerous factors have been identified that are induced by hypoxia and that can promote metastasis (reviewed in [
16‐
20]). The common denominator of most, if not all, of these factors is that they are either directly or indirectly influenced by the action of the family of master transcription regulators during hypoxic conditions: the hypoxia-inducible factor (HIF)-family [
18].
Recently, a separate pathway from the HIFs was described, which is able to regulate gene expression during hypoxia, namely the unfolded protein response or UPR [
21‐
24]. Within this response three distinct arms have been classified: the PKR-like endoplasmic reticulum kinase (PERK)/activating transcription factor 4 (ATF4)-arm, the inositol-requiring protein 1 (IRE1)-arm and the activating transcription factor 6 (ATF6)-arm. These pathways are activated during endoplasmic reticulum stress conditions and enable cell survival by regulating apoptosis, angiogenesis and autophagy [
22‐
25]. Thus far, the UPR has not been directly implicated in hypoxia-induced metastasis. However, recently lysosomal-associated membrane protein 3 (LAMP3, also known as DC-LAMP, TSC-403 or CD208) was identified as a factor induced by hypoxia as part of the PERK/ATF4 arm of the UPR [
26,
27]. In addition, we found that LAMP3 has prognostic relevance in breast cancer [
28]. Two homologs of LAMP3, LAMP1 and LAMP2, have been associated with cancer metastasis previously [
29,
30]. LAMP3 itself was also found to be involved in metastasis: overexpression of LAMP3 in a cervical xenograft model showed an increased metastatic potential [
31]. In what way LAMP3 is involved in breast cancer metastasis and which role hypoxia may have in this process is unknown. Therefore, we set out to determine whether the UPR can influence migration and invasion of breast cancer cells via LAMP3 under hypoxic conditions.
Materials and methods
Cell culture and hypoxic incubations
All cell lines used were obtained from LGC Promochem (Teddington, UK) and maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% (vol/vol) fetal bovine serum (FBS), 20 mM Hepes, 1 × nonessential amino acid, 2 mM L-glutamine and 10 U/ml penicillin, 10 μg/ml streptomycin (all from PAA Laboratories, Cölbe, Germany) at 37°C with 5% CO2. Hypoxic conditions were induced with a H35 Hypoxystation (Don Whitley Scientific Ltd, Shipley, UK).
Cell migration using modified Boyden Chambers
Membranes with 8 μm pores (Greiner Bio-one, Alphen a/d Rijn, The Netherlands) were used in a 24-wells format. A total of 40,000 cells, serum-starved overnight where indicated, were added to the upper compartment. The lower compartment was filled with standard cell culture medium. Cells were allowed to migrate for 24 hours, after which chambers were fixed for 10 minutes in cold 70% ethanol. Membranes were stained with 0.5% (w/v) crystal violet (Sigma-Aldrich, St. Louis, MO, USA) for 30 minutes and subsequently washed thoroughly with tap water. Cells that had not migrated to the lower compartment were removed with a cotton swab. Migrated cells were quantified by solubilizing bound crystal violet in 1% (wt/vol) sodium dodecyl sulfate (SDS, Sigma-Aldrich) for 1 hour at 37°C. Absorbance was measured at 595 nm.
RNA isolation, cDNA synthesis and quantitative polymerase chain reactions (qPCR)
RNA was isolated with Norgen's total RNA purification kit (Norgen Biotek Corp., Thorold, Canada) and stored at -80°C until further processing. cDNA was synthesized using the iScript cDNA synthesis kit (Bio-Rad Laboratories Inc., Richmond, CA, USA) with 1 μg RNA as input. The following primers were used for the qPCR: PERK FW: 5'-CTGATTTTGAGCCAATTC-3' and RV: 5'- CCGGTACTCGCGTCGCTG-3', ATF4 FW: 5'-CCTTCACCTTCTTACAACCT-3' and RV: 5'-GTAGTCTGGCTTCCTATCTC-3', LAMP3 FW: 5'-TGAAAACAACCGATGTCCAA-3' and RV: 5'-TCAGACGAGCACTCATCCAC-3'.
qPCR was carried out on a CFX96 real-time PCR detection system (Bio-Rad) with SYBR Green (Applied Biosystems, Foster City, CA, USA). As a reference gene, hypoxanthine-guanine phosphoribosyltransferase (HPRT) in a pre-developed assay (Applied Biosystems) was used.
Transient transfection
Cells were transfected transiently for PERK, ATF4 or LAMP3 using mission siRNAs (Sigma-Aldrich):
PERK (NM_004836), SASI_Hs01_00096845 and SASI_Hs01_00096846
ATF4 (NM_001675), SASI_Hs02_00332313 and SASI_Hs01_00175197
LAMP3 (NM_014398), SASI_Hs01_00214233 and SASI_Hs02_00345584
Transfections were performed using Saint-Red (Synvolux Therapeutics, Groningen, The Netherlands) according to the manufacturer's instructions.
Generation of stable MDA-MB-231 shLamp3 pools
A U6 promoter-driven short hairpin RNA (shRNA) expression vector targeting LAMP3 and a non-targeting control vector (PLKO1_shLAMP3 (TRCN0000148784, number 842) and PLKO.1 control, respectively) were purchased from Sigma-Aldrich. Briefly, pseudotyped lentiviral particles were produced in HEK293FT cells using the ViraPower lentiviral expression system according to the manufacturer's instructions (Invitrogen, Carlsbad, CA, USA). MDA-MB-231 cells were infected at a low passage number after which a pool of transfected cells was derived by puromycin (4 μg/ml) selection for approximately 10 days.
For the colony-forming assays 500 transiently transfected cells were plated in T25 cell culture flasks (Greiner Bio-one) and allowed to adhere overnight. Cells were incubated under hypoxic conditions (1% O2) for 24 hours after which they were transferred to the normoxic incubator and given time to form colonies. Once colonies in the normoxic controls comprised of at least 50 cells, flasks were fixated in 70% ethanol for 10 minutes at 4°C and stained with 0.5% crystal violet for 30 minutes. Colonies of at least 50 cells were scored manually.
Gap closure assays
Monolayer gap closure assays (formerly known as scratch assays) were conducted using silicone cell culture inserts (Ibidi, Martinsried, Germany) attached to culture plates. In short, 30,000 cells were seeded in the inserts and allowed to recover overnight to form a confluent monolayer. Inserts were removed with tweezers, after which cells were rinsed thrice with Hank's buffered saline solution (HBSS, PAA Laboratories) to remove detached cells. Culture medium was re-added and closure of the gap was followed for 24 hours using live imaging (see below). Gap closure was quantified using ImageJ (National Institutes of Health, Bethesda, MD, USA).
Spheroid culture
Multicellular tumor spheroids were prepared from conventional monolayer cultures using an adapted liquid overlay technique. In short, V-shaped 96-wells plates (Greiner Bio-one) were coated with 0.5% (wt/vol) poly-HEMA (Sigma-Aldrich). A total of 1,000 cells in 100 μl of standard culture medium with 2.5% (vol/vol) Matrigel (BD Biosciences, San Jose, CA, USA) were added to each well, after which cells were spun down for 10 minutes at 1,000 × g. Within 24 hours spheroids were formed.
Pimonidazole staining
After formation, MDA-MB-231 spheroids were incubated for 1 hour with 200 μM of pimonidazole (1-[(2-hydroxy-3-piperidinyl)propyl]-2-nitroimidazole hydrochloride, Natural Pharmacia International Inc., Burlington, MA, USA). Next, spheroids were fixed in 4% paraformaldehyde (Merck Chemicals, Darmstadt, Germany), and embedded in paraffin. Staining was performed on 5-μm sections as previously described [
28], using the following antibodies: mouse-anti-pimonidazole (Natural Pharmacia International Inc.) diluted 1:800 and biotin-conjugated donkey anti-mouse IgG (715-066-150, Jackson ImmunoResearch Laboratories Inc., West Grove, PA, USA) diluted 1:200.
Cell labeling with cell tracker
To label cells with CellTracker Green (Life Technologies, Carlsbad, CA, USA) or CellTracker Orange (Life Technologies), 0.15 × 106 cells/ml were seeded in a T75 cell culture flask (Greiner Bio-one). Cells were allowed to adhere, after which CellTracker was dissolved in DMSO and added at 10 μM to the cell culture medium. Flasks were incubated at 37°C for 45 minutes. Next, cell culture medium was replenished and cells were allowed to recover for 30 minutes, after which they were ready for further experimentation.
Spheroid invasion assays in collagen
Twenty-four hours after formation, spheroids were embedded into collagen. Spheroids were collected in a 15 ml tube and allowed to descend to the bottom. Cell culture medium was renewed once. Next, the spheroids were combined with 4 mg/ml of rat-tail collagen type I (BD Biosciences), according to the manufacturer's instructions. The collagen was allowed to polymerize for 10 minutes at room temperature, after which the mixture was pipetted carefully into a 12-wells plate (Greiner Bio-one). The collagen disk was incubated at 37°C until fully solidified, after which standard cell culture medium was added. Spheroids were allowed to invade the collagen for several days, as indicated in the legends to the figures.
F-actin staining
After invasion of spheroids into collagen, collagen disks were fixed in 4% paraformaldehyde for 30 minutes at room temperature. Next, disks were washed with PBS, after which they were incubated with Alexa 488-conjugated Phalloidin (1:100; Life Technologies) for 3 hours at room temperature. After washing once more with PBS, images were acquired.
Live imaging and microscopy
Live imaging of cells was performed using the JULI (Just Unbelievable Live Imaging) system from PAA. All other microscopic images were obtained using a Leica DM 6000 fluorescence microscope in combination with IPLab imaging software (Scanalytics Inc., Fairfax, VA, USA).
Data analysis and statistics
Unless stated otherwise, all data are presented as mean values ± standard deviation. Statistical analysis was carried out using Student's t tests or one-way analysis of variance (ANOVA) with Tukey's post hoc test where appropriate unless stated otherwise. Two-sided P values < 0.05 were considered statistically significant. Asterisks indicate statistical significance: *** is P < 0.001, ** is P < 0.01 and * is P < 0.05.
Discussion
In this study, evidence is provided that UPR-induced LAMP3 can influence hypoxia-mediated cell migration of breast cancer cells. We provide evidence that apart from the established involvement of the HIF-pathway in the induction of cancer cell spread, the UPR is a second manner in which hypoxia is implicated in breast cancer cell migration.
A characteristic of the UPR is its maximum induction under conditions of severe hypoxia (< 0.2% O
2) or anoxia [
32]. Indeed, LAMP3 as a UPR-induced factor was found to have its peak of induction under these conditions [
26,
28]. However, cell migration and invasion is often studied under conditions of more moderate hypoxia, around 1% oxygen [
33‐
38], conditions which maximally induce HIF-1α expression [
23]. The current study shows that knockdown of PERK, ATF4 or LAMP3 in combination with hypoxic exposure to 1% O
2 led to a reduction in cell migration. If the UPR and its associated factors are maximally induced by anoxia, why are the largest effects observed at moderate levels of hypoxia? Several studies have shown that the UPR can indeed be induced by moderate hypoxia as well, but with different kinetics [
21,
39,
40]. In this study assays were performed at 1% O
2 as stronger hypoxic conditions did not lead to a stimulation of cell migration. Possibly, at severe hypoxic conditions, cells apply the UPR more for cell survival [
25] than migration.
In the transwell assays an intriguing effect of the addition of serum on the induction of cell migration by hypoxia was observed. Serum-starved cells migrated most profoundly at 1% O
2, whereas serum-supplemented cells migrated best at 0.1% O
2. Serum dependency of cancer cell invasion has been observed before [
41,
42]. When MDA-MB-231 cells were serum depleted, no increase in invasion was found for a hypoxic incubation (1.5% O
2) [
42]. As addition of serum led to an increased invasion under hypoxia, it was suggested that serum might contain factors that increase invasion under hypoxic conditions [
42]. In a different study, the effect of hepatocyte growth factor (HGF) on tumor invasion was examined in U2-OS and SiHa cells [
43]. Mild hypoxia (3% O
2) was found to increase invasion by amplifying HGF signaling, thereby sensitizing cells to HGF stimulation [
43]. These and the current data indicate that the role of serum in cancer cell migration and invasion is more than just a chemoattractant and that growth factor signaling has a vast influence on the effects of hypoxia in migration and invasion assays. Trying to survive during hypoxia is of critical importance for cells. Depriving cells of serum may make survival even more difficult. Possibly, without serum and under severe hypoxic conditions (0.1% O
2) cells are not migrating as fast as under moderate hypoxia (1% O
2) because cell survival is more essential. When serum is present, the need for survival may become less critical even at 0.1% O
2, causing the increase in cell migration.
The importance of the tumor microenvironment during cell migration was emphasized when migration assays were performed with cells grown as monolayer and cells initially grown as spheroids. The latter were found to be more migratory than the former. In other words, cells that experienced a simplified microenvironment prior to the assay were more migratory, despite the fact that spheroids had to be disintegrated back to a single-cell suspension. This behavior was also observed previously with murine breast cancer cell lines in transwell invasion assays [
44]. As the spheroids used contain a central hypoxic core, the enhanced ability of spheroid cells to migrate or invade could be a consequence of hypoxia. This, however, remains to be established.
With the evidence that LAMP3 is involved in hypoxia-induced cell migration, it needs to be elucidated which mechanism LAMP3 uses to cause the actual spread of cancer cells. LAMP3 protein under physiological conditions is localized within the lysosomal membrane [
45]. For LAMP1 and LAMP2 it has been previously established that their expression can relocalize to the plasma membrane in cancer cells [
29]. Cell lines with a stronger metastatic capacity showed an enhanced expression of LAMP1 and LAMP2 on the cell membrane [
29]. It is believed that LAMP surface expression provides a cancer cell with means to attach to selectins on endothelial cells and enhance their capacity to form metastases [
30]. A similar mechanism could be responsible for the role of LAMP3 in hypoxia-induced cell migration. LAMP3 was found to have the ability to relocalize to the plasma membrane upon influenza A virus infections in HeLa cells [
46]. Nevertheless, we have not been able to show LAMP3 expression on the cell surface in the MDA-MB-231 cells used, under either normoxic or hypoxic conditions. Immunohistochemically, in MDA-MB-231 cells we have only observed LAMP3 expression in the cytoplasm. An alternative explanation for the function of LAMP3 in cell migration is the link LAMP proteins have with autophagy [
47]. Autophagy has been previously suggested as a possible mechanism responsible for increased survival and increased metastasis of cancer cells [
48,
49]. Analysis of the autophagy marker LC3B expression in a large subset of breast tumors revealed that it is associated with metastasis and a worse outcome [
50]. However, Indelicato
et al. found that chemically induced autophagy results in reduced invasion of MDA-MB-231 cells under both normoxic and hypoxic (1% O
2) conditions, whereas LC3B silenced cells showed a decreased invasion during hypoxic conditions [
37]. Thus, there is evidence that autophagy may function as a mechanism behind hypoxia-induced metastasis, but its precise role is far from clarified.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
Laboratory work and data analysis was performed by AN. The study was designed by AN, JB, FCGJS and PNS. SL created the stable LAMP3 knockdown cell lines. AN, JB, FCGJS and PNS wrote the manuscript. HM, BGW and SL critically revised the manuscript. All authors approved the final version of the manuscript.