LAMP3 is a newly described hypoxia regulated gene of potential interest in hypoxia-induced therapy resistance and metastasis. The prognostic value of LAMP3 in breast cancer was investigated.
LAMP3 is a newly described hypoxia regulated gene of potential interest in hypoxia-induced therapy resistance and metastasis. The prognostic value of LAMP3 in breast cancer was investigated.
Expression levels of LAMP3 in breast cancer cell lines and patient tissues were determined by real-time polymerase chain reaction and in a tissue microarray by immunohistochemistry. Immunofluorescent staining was used to evaluate the distribution of LAMP3 in tumor xenografts relative to pimonidazole. Kaplan-Meier analysis as well as multivariate Cox regression survival analyses were performed.
LAMP3 was variably expressed in breast cancer cell lines and induced in an oxygen concentration-dependent manner. LAMP3 protein expression colocalized with hypoxic areas in breast cancer xenografts. LAMP3 mRNA was higher in breast tumors from patients with node-positive (P = .019) and/or steroid hormone receptor-negative tumors (P < .001). Breast cancer patients with high LAMP3 mRNA levels had more locoregional recurrences (P = .032 log-rank). This was limited to patients treated with lumpectomy and radiotherapy as primary treatment (n = 53, P = .009). No association with metastasis-free survival was found. In multivariate Cox regression analysis, LAMP3 remained as a statistically independent prognostic factor for locoregional recurrence (hazard ratio, 2.76; 95% confidence interval, 1.01-7.5; P = .048) after correction for menopausal status, histologic grade, tumor size, nodal status, therapy, and steroid hormone receptor status. LAMP3 protein in breast cancer tissue proved also to be of prognostic relevance.
Evidence was provided for an association of LAMP3 with tumor cell hypoxia in breast cancer xenografts. In the current breast cancer cohorts, LAMP3 had independent prognostic value. Cancer 2011;117:3670–3681. © 2011 American Cancer Society.
Lysosome-associated membrane protein 3 (LAMP3) was first identified in 1998 independently by 2 research groups. De Saint-Vis et al named it DC-LAMP, a marker of mature dendritic cells1 with a possible role in migration of dendritic cells from the periphery into lymph vessels,2 whereas Ozaki et al described it as TSC403 with specific expression in the lung and overexpression in carcinomas of the breast, amongst others.3 Significant similarities of TSC403 with LAMP1 and LAMP2 were found.3 LAMPs are highly glycosylated type 1 integral membrane proteins, which reside in lysosomal membranes.1, 4 Their function is still largely unknown. In noncancerous cells, LAMP1 and LAMP2 are rarely expressed on the cell surface.5 However, in cancer cells their presence in the plasma membrane is frequently observed, with an increased incidence in more metastatic cell lines.6, 7 Experiments with overexpression of LAMP3 indicated similar characteristics for this protein: a strong association between LAMP3 and the promotion of metastatic potential was described both in vitro and in vivo.2 This suggests that LAMP3 expression might be related to metastasis. Indeed, high LAMP3 mRNA levels correlated significantly with poor prognosis in a small cohort of cervical cancers.2
For several decades, hypoxia (oxygen deprivation) has been recognized as a hallmark of solid tumors. A low blood and oxygen supply to cancers complicates their (local) treatment. Hypoxia can decrease the effectiveness of both radiotherapy and chemotherapy and induce treatment resistance.8 Especially for radiotherapy, hypoxia poses problems, as oxygen is required for radiation to cause DNA damage.9 In addition, several studies have reported that less well-oxygenated tumors show a worse prognosis and a higher incidence of metastases compared with better oxygenated tumors.10-12 Hypoxia has therefore been linked to a more aggressive and metastatic tumor phenotype. Several hypoxia-regulated genes have been described that influence metastasis of tumors, including factors that enhance dissociation and migration of tumor cells.13 Recently, LAMP3 was shown to be regulated by hypoxia in a panel of tumor cells, via the unfolded protein response (UPR).14 The UPR is a mechanism of adaptation to endoplasmic reticulum stress and has been demonstrated to contribute to hypoxic adaptation in tumors through multiple mechanisms.15 This adaptation may facilitate the survival of treatment-resistant hypoxic cells, leading to a poor prognosis of patients.16 Whether LAMP3 is associated with treatment resistance or metastasis in solid tumors is unknown.
In this study, we investigated the correlation between hypoxia and LAMP3 expression in breast cancer xenografts and we assessed the prognostic value of LAMP3 mRNA and protein expression in breast cancer patients.
MDA-MB-231 and SKBR-3 breast tumor cells were obtained from LGC Promochem (London, UK) and cultured at 37°C with 5% CO2 in DMEM (Lonza, Biowhittaker, Walkersville, MD) supplemented with 10% (v/v) heat-inactivated fetal bovine serum (FBS), 10 U/ml penicillin, 10 μg/ml streptomycin, 2 mM L-glutamine, 20 mM Hepes (all from Invitrogen, Carlsbad, CA), and 1× nonessential amino acids (NEAA, Promocell, Heidelberg, Germany). Cells were harvested and reseeded at 100% confluence. cDNA from a panel of breast cancer cell lines (n = 16) was a gift from Dr. M. Schutte (Department of Medical Oncology, Josephine Nefkens Institute, Erasmus University Medical Center, Rotterdam, the Netherlands).17, 18
For culturing under hypoxic conditions, MDA-MB-231 cells were seeded at 300,000 cells per 10-cm dish and were allowed to recover overnight. Hypoxia was induced using a H35 Hypoxystation (Don Whitley Scientific, Ltd, Shipley, UK), set at 0.5%, 0.2%, or 0.1% oxygen. After incubation of up to 72 hours, cells were harvested.
RNA from hypoxic cultured cells was isolated using Norgen's total RNA purification kit (Norgen Biotek Corp, Thorold, Canada), according to the manufacturer's instructions with addition of an on-column DNase treatment. Isolated RNA was stored at −80°C until further processing.
cDNA synthesis was performed by reverse transcribing 1 μg of RNA using the Reverse Transcription System (Promega, Madison, WI) according to the manufacturer's instructions. The primers used for LAMP3 qPCR were FW 5′-TGAAAACAACCGATGTCCAA-3′ and RV 5′-TCAGACGAGCACTCATCCAC-3′.
mRNA expression was analyzed on a CFX96 real-time PCR detection system (Bio-Rad Laboratories Inc, Richmond, CA) using SYBR Green (Applied Biosystems, Foster City, CA). Hypoxanthine-guanine phosphoribosyl transferase (HPRT) was used as the reference gene, analyzed with a predeveloped assay (Applied Biosytems).
Cells were harvested in RIPA buffer (1% NP-40, 0.5% sodium desoxycholate and 0.1% sodium dodecylsulfate) to which phosphatase and protease inhibitors (Roche, Indianapolis, IN) were added. Protein concentrations were determined using the Pierce BCA assay (Thermo Fisher Scientific, Rockford, IL), after which 30 μg of protein was fractionated on 12% Criterion XT Bis-Tris gels (Bio-Rad). After running, samples were transferred to PVDF membranes (Millipore Immobilon, Millipore, Bedford, MA) for 2 hours at 4°C. After blocking with 4% nonfat dry milk (Bio-Rad), membranes were incubated overnight at 4°C with rabbit-anti-human-LAMP3 IgG (AP1827a, 0.25 mg/ml, Abgent, San Diego, CA) in a dilution of 1:100 or rabbit-anti-β-actin IgG (#4970, Cell Signaling Technology, Beverly, MA) in a dilution of 1:1000, both diluted in 5% (w/v) bovine serum albumin (Sigma-Aldrich, St. Louis, MO) in tris buffered saline (TBS) with 0.1% Tween-20 (w/v) (Sigma). Next, membranes were incubated for 1 hour at room temperature with horseradish peroxidase (HRP) -labeled goat-anti-rabbit immunoglobulin G (IgG; 1:2000, #7074, Cell Signaling). Proteins were detected using chemiluminescent peroxidase substrate (Sigma) and imaged on a ChemiDoc XRS+ imaging system (Bio-Rad).
Xenografts of breast cancer cells were grown on nu/nu BALB/c athymic mice, kept in a specific pathogen-free unit at the central animal facility of the Radboud University Nijmegen. All animal procedures were approved by the local ethics committee. In short, a cell suspension of MDA-MB-231 or SKBR-3 cells was diluted in Matrigel (BD Bioscience, Bedford, MA) 2:1 and injected subcutaneously (sc) into the flank of themice. At an average tumor size of 8 mm to 9 mm, mice were injected intraperitoneally (ip) with the hypoxic cell marker pimonidazole (1-[(2-hydroxy-3-piperidinyl)propyl]-2-nitroimidazole hydrochloride, Natural Pharmacia International Inc, Burlington, MA). One hour postinjection, mice were sacrificed, and tumors were harvested and immediately stored in liquid nitrogen until further processing.
Frozen 5-μm tumor sections were fixed in acetone for 10 minutes at 4°C, after which they were rehydrated for 30 minutes in phosphate-buffered saline (PBS) pH 7.4 (Klinipath, Duiven, the Netherlands) and subjected to staining. All antibodies were dissolved in primary antibody diluent (PAD, AbD Serotec, Oxford, UK). Between all incubations, sections were rinsed 3 times in PBS. Sections were incubated overnight at 4°C with goat-anti-human-LAMP3 IgG (AF4087, 0.2 mg/ml, R&D Systems Inc, Minneapolis, MN) used in a dilution of 1:100. Subsequently, sections were treated with Cy3-conjugated donkey-anti-goat IgG (705-166-147, 1.5 mg/ml, Jackson ImmunoResearch Laboratories Inc, West Grove, PA) in a 1:600 dilution for 30 minutes at 37°C as the secondary antibody. Vessels were visualized by incubating sections with undiluted 9F1 supernatant, a rat monoclonal antibody to mouse endothelium (Department of Pathology, Radboud University Nijmegen Medical Centre, RUNMC) for 45 minutes at 37°C. Then sections were incubated with rabbit-anti-pimonidazole (J.A. Raleigh) diluted 1:1000 for 30 minutes at 37°C. Finally, sections were incubated with Alexa-Fluor-647-conjugated chicken-anti-rat IgG (A21472, 2 mg/ml, Invitrogen) diluted 1:100 and Alexa-Fluor-488-conjugated donkey-anti-rabbit IgG (A21206, 2 mg/ml, Invitrogen) diluted 1:600 for 60 minutes at 37°C. Nuclei were visualized using Hoechst 33342 (1 mg/ml) 1:3000 for 5 minutes. Slides were mounted in Fluorostab (ICN Pharmaceuticals, Inc, Zoetermeer, the Netherlands).
For the TMA construction, formalin-fixed paraffin-embedded (FFPE) tumor sections were marked on hematoxylin and eosin-stained slides of the primary tumors by a pathologist. TMAs were constructed with a tissue arrayer using a 2-mm-diameter punch (Beecher Instruments, Silver Spring, MD). Staining for LAMP3 was performed as described in the next section and scored blinded for outcome semiquantitatively as being either absent/low or present/high.
FFPE sections were dewaxed in Histosafe (Adamas Instruments BV, Leersum, the Netherlands) and rehydrated in graded alcohols. Antigen retrieval was performed by boiling sections in 10 mM citrate buffer pH 6.0 (Dako, Glostrup, Denmark) for 30 minutes after which endogenous peroxidase was blocked for 10 minutes using 3% H2O2 in methanol. Sections were incubated for 30 minutes with 5% normal donkey serum (Jackson) in PAD and subsequently incubated overnight at 4°C with goat-anti-human-LAMP3 (R&D Systems), diluted 1:100 in PAD. Next, sections were incubated for 60 minutes with Biotin-conjugated donkey-anti-goat IgG (705-066-147, 1.5 mg/ml, Jackson) diluted 1:200 in PBS, after which they were incubated for 30 minutes with avidin-biotin complex reagent (Vector Laboratories, Inc, Burlingame, CA). After rinsing with deionized water, sections were incubated with diaminobenzidine (Invitrogen) for 10 minutes. Next, sections were rinsed with tap water and counterstained with hematoxylin (Klinipath) for 1 minute, after which they were dehydrated and mounted in mounting medium (Klinipath).
All images were acquired using IP-lab imaging software (Scanalytics Inc, Fairfax, VA) in combination with a Leica DM 6000 (fluorescence) microscope.
Coded tumor tissues were used in accordance with the Code of Conduct of the Federation of Medical Scientific Societies in the Netherlands (Code for Proper Secondary Use of Human Tissue in the Netherlands, http://www.fmwv.nl). The study adhered to all relevant institutional and national guidelines, and was reported according to REMARK guidelines.19
A series of 183 patients with unilateral, resectable breast cancers who had undergone resection of their primary tumors between November 1987 and December 1997 were selected based on the availability of RNA in the tumor bank of the Department of Laboratory Medicine of the RUNMC.20 This bank contains material from breast cancer patients of 7 different hospitals of the Comprehensive Cancer Centre East in the Netherlands. The inclusion and exclusion criteria have been described earlier.20 Postoperative radiotherapy was given to the breast after an incomplete resection or after breast-conserving treatment, or parasternal when the tumor was medially localized. Axillary irradiation was given in the case of positive lymph nodes.
A TMA was constructed from FFPE tissues as described above from another cohort of 61 breast cancer patients who had undergone lumpectomy and radiotherapy as primary treatment of their tumors between 1991 and 1996 at the Rijnstate Hospital, Arnhem, the Netherlands. These patients had no involved axillary lymph nodes and received no systemic adjuvant therapy, as was the practice at the time. Other clinicopathologic characteristics (steroid hormone receptor status, tumor size, and histologic grade) were essentially similar to those of the cohort from which RNA was extracted.
Statistical analyses were carried out using SPSS 16.0 software (SPSS Benelux BV, Gorinchem, the Netherlands). Normality of data distribution was tested by Kolmogorov-Smirnov testing. An analysis of variance (ANOVA) with post-hoc Tukey's honestly significant difference (HSD) testing was used to assess the effect of hypoxia on LAMP3 mRNA levels. Nonparametric Mann-Whitney U tests or Kruskall-Wallis tests were used to assess differences in LAMP3 mRNA levels between categories of patients. Locoregional control (defined as the time from surgery until diagnosis of locoregional recurrent disease), disease-free survival (DFS) time (defined as the time from surgery until diagnosis of recurrent disease), metastasis-free survival (defined as the time between date of surgery and diagnosis of a distant metastasis), and overall survival (OS) time (defined as the time between date of surgery and death by any cause) were used as follow-up end points. Survival curves were generated using the Kaplan-Meier method. Equality of survival distributions was tested using log-rank testing and Cox univariate and multivariate regression analyses. Variables were selected for the multivariate survival analyses by backward stepwise selection, with removal testing based on the probability of the likelihood-ratio statistic, at a P > .10. Two-sided P-values <.05 were considered to be statistically significant.
We first measured the expression of LAMP3 mRNA in a panel of 16 different breast cancer cell lines under standard (normoxic) culturing conditions (Fig. 1). Three cell lines (SUM44PE, MPE600, and ZR75-30) displayed no measurable LAMP3 mRNA expression. In the other breast cell lines, levels varied >100-fold between SKBR-3 and MDA-MB-361, with OCUB-F cells exhibiting even higher LAMP3 levels. No association between LAMP3 expression and cancer subtype or ER, PR, or HER2 expression was found (Fig. 1).
In MDA-MB-231 cells, which show moderate basal LAMP3 mRNA expression levels, LAMP3 was strongly induced under hypoxic conditions in an oxygen-dependent manner (Fig. 2A). The highest induction (ie, 4-fold) was observed after 48 hours of incubation under severe hypoxic conditions (0.1% O2, P = .003). An incubation of 72 hours at 0.1% O2 led to a reduction in LAMP3 expression compared with 48 hours. At 0.2% O2 there was a 2-fold increase overall as a maximum, whereas at 0.5% O2 no statistically significant change was seen. Thus, the level of LAMP3 mRNA induction is dependent on the level of hypoxia, with severe hypoxia leading to a stronger induction of LAMP3 mRNA than moderate hypoxia. This dependency is consistent with the known oxygen dependency of UPR activation.21 Protein expression was found to be induced in a similar manner (Fig. 2B).
Next, xenografts of SKBR-3 (low endogenous expression) and MDA-MB-231 breast cancer cells were established in immunodeficient mice. After subcutaneous injection of the hypoxic cell marker pimonidazole, mice were sacrificed and xenografted tumors were stained for LAMP3 and hypoxia. In agreement with our in vitro expression data, we found only limited staining of the SKBR-3 tumors for LAMP3 in normoxic areas (Fig. 3A-C). Simultaneous staining for LAMP3 and pimonidazole revealed a regional colocalization of intense LAMP3 staining with areas of severe hypoxia, distant from perfused blood vessels (Fig. 3D). A hematoxylin and eosin staining showed the epithelial cell origin of this breast cancer xenograft and LAMP3 expressing cells (Fig. 3E). The MDA-MB-231 xenografted tumors were well oxygenated and did not show hypoxic areas or LAMP3 expression (results not shown).
LAMP3/HPRT mRNA values were assessed in 183 human invasive breast cancer tissues and were found to be log-normally distributed with a >100-fold variation in values, similar to the range in the breast cancer cell lines. There was no association between LAMP3 and menopausal status, type of surgery, histologic grade, or tumor size (pT, Table 1). However, LAMP3 was significantly higher in node-positive patients (P = .019, Kruskal-Wallis, Fig. 4A), steroid hormone receptor-negative tumors (P < .001, Mann-Whitney U, Fig. 4B), patients receiving radiotherapy as part of their primary treatment (P = .022), and patients receiving adjuvant systemic chemotherapy (P = .050). Thus, LAMP3 is associated with poor prognosis characteristics (node positive, steroid hormone receptor negative) and therefore with more aggressive treatment (radiotherapy, chemotherapy).
Patients were dichotomized based on LAMP3 levels in their primary tumor and evaluated for response parameters. Patients with higher levels of LAMP3 (n = 99) had more locoregional recurrences than those with low LAMP3 mRNA levels (n = 84), P = .032 log-rank (Fig. 5A). In Cox regression analysis, this amounted to a hazard ratio (HR) of 2.85 (95% CI, 1.05-7.8). There was no association with distant metastasis-free or overall survival. In exploratory subgroup analyses, we found that the association of LAMP3 with locoregional control was limited to patients who had received radiotherapy as part of their primary treatment. In patients who had not received radiotherapy, LAMP3 showed no association with locoregional control (n = 49, P = .762, Fig. 5B), whereas in patients who had received radiotherapy, a significant association was found (n = 134, P = .034, Fig. 5C). Within the patient cohort that had received radiotherapy, no association of LAMP3 with locoregional control was found if this was part of a modified radical mastectomy procedure (n = 81, P = .615, Fig. 5D). LAMP3 was significantly associated with locoregional control but only in those patients who received radiotherapy as part of a breast-conserving lumpectomy (n = 53, P = .009, Fig. 5E).
In multivariate Cox regression analysis, only LAMP3 remained as an independent prognostic factor for locoregional recurrence (HR = 2.76, 95% CI, 1.01-7.5, P = .048) after correction for menopausal status, histologic grade, tumor size, nodal status, therapy, and steroid hormone receptor status (Table 2).
|HR||95% CI||P||HR||95% CI||P|
|Post vs pre||0.827||0.320-2.134||0.964||0.360-2.579||.942|
|III vs I/II||1.433||0.553-3.718||1.622||0.605-4.352|
|pT2 vs pT1||1.905||0.630-5.768||1.989||0.640-6.185|
|pT3 vs pT1||0.974||0.108-8.746||1.125||0.118-10.74|
|pT4 vs pT1||2.783||0.309-25.07||6.663||0.646-68.72|
|1-3 vs 0||1.531||0.621-3.771||1.414||0.104-19.283|
|≥4 vs 0||0.906||0.194-4.227||0.523||0.027-10.197|
|Mastectomy vs lumpectomy||0.873||0.360-2.117||0.761||0.314-1.845|
|Yes vs no||1.381||0.465-4.108||0.826||0.215-3.172|
|Endocrine vs none||1.173||0.465-2.956||1.384||0.546-3.506|
|Chemo vs none||1.311||0.353-4.868||1.196||0.319-4.477|
|Both vs none||—a||—|
|Positive vs negative||0.726||0.293-1.798||1.371||0.364-5.155|
|Positive vs negative||0.736||0.310-1.749||0.833||0.329-2.107|
|High vs low||2.853||1.044-7.792||2.758||1.010-7.534|
We set out to validate the data on LAMP3 mRNA as being prognostic in breast cancer patients who were treated with lumpectomy and radiotherapy on the protein level. For this, LAMP3 protein was semiquantitatively scored in TMA of a series of FFPE sections of breast cancer patients that were only treated with lumpectomy and radiotherapy. These patients received no adjuvant systemic therapy, making it possible to distinguish a prognostic from a predictive value of LAMP3 in these patients. Of the 61 patients, 21 (34%) had low or absent levels of LAMP3 protein (Fig. 6A), and 40 (66%) had high levels (Fig. 6B) of LAMP3 protein in their tumor. Despite the clear difference in the number of events between these subgroups (1 versus 8 in the LAMP3 low and high cohort, respectively), the power was too low to obtain a significant difference in locoregional control (P = .121). These 2 categories of patients did, however, differ significantly in both disease-free survival (DFS) (P = .024, Fig. 6C) and overall survival (OS) time (P = .019, Fig. 6D), confirming the prognostic value of LAMP3 mRNA levels in this patient group on the protein level.
In this study, the occurrence and prognostic relevance of LAMP3 in breast cancer was evaluated. We present evidence indicating that LAMP3 mRNA levels are elevated in aggressive ER-negative breast tumors and in tumors that have already spread to the axillary lymph nodes. Patients with high LAMP3 mRNA levels also have more locoregional recurrences than those with low levels, a phenomenon that remains after correction for menopausal status, tumor size, histologic grade, nodal status, steroid hormone receptor status, adjuvant therapy, and radiotherapy. The prognostic value of LAMP3 for locoregional control is limited to patients treated with breast-conserving lumpectomy and radiotherapy. The prognostic value of LAMP3 mRNA is also seen for LAMP3 protein as assessed by immunohistochemistry.
In 2004, Treileux et al examined the potential prognostic role of LAMP3 in breast cancer tissue, using a semiquantitative immunohistochemical analysis.22 They found that the presence of mature LAMP3+ dendritic cells (DC) within clusters of lymphocytes at the margin of the tumor correlated with lymph node involvement and tumor grade, although no association with prognosis was found. Our immunohistochemical data indicate that tumor hypoxia is also an important regulator of LAMP3 in the epithelial cancer cells. LAMP3 expression in breast cancer xenografts colocalizes with tumor cell hypoxia as revealed by pimonidazole labeling. Together with our immunohistochemical data on actual breast tumors, this indicates that lymphocyte infiltration, which reportedly has different effects on prognosis in breast cancer depending on ER status,23 is unlikely to be the sole cause of the association of LAMP3 mRNA with prognosis we report here. Rather, the tumor epithelial cells themselves aberrantly express LAMP3, with consequences for locoregional control.
LAMPs have previously been implicated in cancer;24-26 especially in relation to the metastatic potential of tumors they may serve important functions.6, 27, 28 LAMPs are characterized by high and complex levels of glycosylation.4, 29 They are the main carriers of sialyl-Lex (sialylated Lewis x antigen), which is amplified in carcinoma cells.30 Via these carbohydrates, tumor cells are able to connect with E-selectins, adhere to endothelial cells, and form metastases, suggesting that LAMPs play a role in the metastasis of certain solid tumors.31-33 Indeed, overexpression of LAMP3 in cervical cancer cells used for xenografts led to an enhancement of formation of metastasis.2 Our observation that mRNA levels are significantly higher in breast cancer patients with positive axillary lymph nodes would indicate that LAMP3 might also serve a role in metastasis of breast cancers. Other studies have demonstrated that tumor hypoxia is also associated with a more malignant phenotype and a higher prevalence of metastasis.10-12
Although a direct link between hypoxia, LAMP3, and metastasis has not yet been explored, hypoxia was found to increase the expression of LAMP-associated sialyl-Lex at the surface of colon cancer cells.34 This leads to increased adhesion to endothelial E-selectin and potentially could provide a mechanism for hypoxia-meditated metastasis. We did not find an association between LAMP3 mRNA expression in breast cancer and distant metastasis-free survival in our breast cancer cohort. However, it is possible that protein localization and/or glycosylation patterns of the protein may be relevant for its association with metastasis.
Interestingly, unlike most other hypoxia targets, LAMP3 induction occurs independently of the master hypoxia transcription factor Hypoxia Inducible Factor-1α (HIF-1α).14 Consistent with previous reports, we found that LAMP3 induction was dependent on the level of hypoxia, with severe hypoxia being most effective. This is consistent with the reported dependency of UPR activation in response to hypoxia. The UPR is induced potently and rapidly under conditions of severe hypoxia but requires oxygen levels well below those required for most HIF-1 target genes.21 As such, LAMP3 may serve as a unique biomarker for the presence of UPR activation under conditions of severe hypoxia in human tumors. This is interesting in the context of our intriguing finding that LAMP3 is specifically related to locoregional control in patients treated with breast-conserving lumpectomy and radiotherapy. We speculate that high levels of LAMP3 and UPR activation may indicate the presence of substantial numbers of radiation-resistant hypoxic cells in these tumors. As such, LAMP3 may be a particularly useful marker of the radiation-resistant fraction of the tumor. There have been intense efforts made at identifying endogenous markers of tumor hypoxia,35, 36 although most of these studies have focused on HIF target genes such as carbonic anhydrase 9 (CA9), glucose transporters (GLUT-1,4), and vascular endothelial growth factor (VEGF). Although clearly important, these hypoxia targets are activated at moderate oxygen concentrations (1% to 2% oxygen), which may not be sufficient to cause radiation resistance. Tumor cells become maximally radiation resistant only below 0.1% oxygen, levels that are well below those required to activate HIF. The UPR, and LAMP3, may thus provide a more accurate surrogate of oxygenation levels associated with clinical resistance to radiotherapy. Our results suggest that the prognostic value of LAMP3 in breast cancer is due to its induction by severe hypoxia, the consequential association with radiotherapy resistance, and lack of locoregional control.
In conclusion, LAMP3 is expressed to highly variable levels within a large panel of breast cancer cell lines and breast cancer patient samples. In addition, LAMP3 is strongly induced by hypoxia both in vitro and in vivo. The correlation between LAMP3 mRNA expression and nodal spread strengthens the concept of a function of LAMP3 in (hypoxia-mediated) metastasis formation in (breast) cancer. Furthermore, the identification of LAMP3 as an independent prognostic factor for locoregional recurrence supports the notion that LAMP3 may be a biomarker for hypoxia-mediated treatment resistance.
We acknowledge the contributors to the breast tumor bank: J.J.T.M. Heuvel, D. van Tienoven, and A.J. Geurts-Moespot (Department of Laboratory Medicine, RUNMC), as well as surgeons, internists, oncologists, and in particular patients. Sandra Heskamp (Department of Nuclear Medicine, RUNMC) is acknowledged for supplying the xenografts used in this study. We thank Dr. M. Schutte for providing cDNAs of the breast cancer cell-line panel.
The authors made no disclosures.