Intermittent hypoxia and cancer: Undesirable bed partners?

Respiratory Physiology & Neurobiology xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Respiratory Physiology & Neurobiology journal homepage: www.elsevier.com/locate/resphysiol

Intermittent hypoxia and cancer: Undesirable bed partners? ⁎

Isaac Almendrosa,b, , David Gozalc a

Unitat de Biofísica i Bioenginyeria, Facultat de Medicina i Ciències de la Salut, Universitat de Barcelona, Spain Centro de Investigación Biomédica en Red de Enfermedades Respiratorias, 28029 Madrid, Spain Section of Pediatric Sleep Medicine, Department of Pediatrics, Pritzker School of Medicine, Biological Sciences Division, The University of Chicago, Chicago, IL, United States b c

A R T I C L E I N F O

A B S T R A C T

Keywords: Intermittent hypoxia Cancer Oxidative stress Inflammation

The deleterious effects of intermittent hypoxia (IH) on cancer biology have been primarily evaluated in the context of the aberrant circulation observed in solid tumors which results in recurrent intra-tumoral episodic hypoxia. From those studies, IH has been linked to an accelerated tumor progression, metastasis and resistance to therapies. More recently, the role of IH in cancer has also been studied in the context of obstructive sleep apnea (OSA), since IH is a hallmark characteristic of this condition. Such recent studies are undoubtedly adding more information regarding the role of IH on tumor malignancy. In terms of the IH patterns associated with OSA, this altered oxygenation paradigm has been recently proposed as a determinant factor in fostering cancer incidence and progression from both in vitro and in vivo experimental models. Here, we summarize all the available evidence to date linking IH effects on several types of cancer.

1. Intermittent hypoxia in cancer Hypoxia is a frequent occurrence in most solid tumors, and the presence of low intra-tumoral oxygen tensions has been ascribed as a major modifying factor underlying cancer progression, metastasis, angiogenesis, chemoresistance and resistance to irradiation (Muz et al., 2015). In recent years, increasing interest has emerged around intermittent hypoxia (IH) in cancer. Recently, in vivo tumor oxygenation assessments in humans have revealed that some tumoral regions can experience cyclic hypoxia (Matsumoto et al., 2010). However, these cycles of hypoxia-re-oxygenation in solid tumors do not manifest predictable regularity and repetitive patterns, and can occur with periodicities from minutes to days. The cyclic hypoxia phenomenon in cancer, and particularly in rapidly growing tumors, has been attributed to changes in solid tumor perfusion, development of newer vascular networks and also, intermittent and aberrant blood circulation (Matsumoto et al., 2010). Taking into account these exploratory evidences, a substantial body of work and investigative effort has been put forth to study the contribution of tumoral hypoxia in the enhancement of some of the malignant properties in several types of cancer (for reviews see refs: DeBerardinis and Chandel, 2016; Dehne et al., 2017; Forster et al., 2017; Manoochehri et al., 2016; Qiu et al., 2017; TarradoCastellarnau et al., 2016; Toth and Warfel, 2017; Wong et al., 2016). In parallel with such discoveries, and based on the conceptual framework that a disease characterized by IH could modify the clinical

⁎

trajectories of some cancers, efforts have been conducted in recent years to evaluate the role of IH in cancer progression in the context of obstructive sleep apnea (OSA) (Almendros et al., 2012a). In this set of ever expanding studies, the IH patterns applied have been selected to include specific characteristics in order to mimic this sleep disorder (Dewan et al., 2015). In particular, the severity of hypoxia is less severe as far as PaO2, but includes a much higher frequency of cycles of hypoxia and re-oxygenation, similar to the ranges observed in actual patients with moderate to severe OSA. In addition, the application of IH in OSA-related cancer studies is delivered both to living mice and directly to the cells. Therefore, in vivo IH will not only exert local effects in the tumor cells per se, but also can elicit changes in the host immune response, and pro-inflammatory and angiogenic molecules may be released from different tissues and organs contributing to the oncogenic processes. Most of studies that have thus far focused on OSA-like IH and cancer have revealed increases in the tumor malignant properties and inferential increases in the resistance to cancer therapies. Although the relationship between OSA and cancer can be also modulated by sleep fragmentation as recently illustrated by Hakim et al. (2014), here, we will exclusively review the findings obtained after IH exposures, and describe the several types of cancer that have been investigated to date under IH challenges.

Corresponding author at: Unitat de Biofísica i Bioenginyeria, Facultat de Medicina i Ciències de la Salut, Casanova 143, 08036 Barcelona, Spain. E-mail address: [email protected] (I. Almendros).

http://dx.doi.org/10.1016/j.resp.2017.08.008 Received 19 June 2017; Received in revised form 27 July 2017; Accepted 10 August 2017 1569-9048/ © 2017 Elsevier B.V. All rights reserved.

Please cite this article as: Almendros, I., Respiratory Physiology & Neurobiology (2017), http://dx.doi.org/10.1016/j.resp.2017.08.008

Respiratory Physiology & Neurobiology xxx (xxxx) xxx–xxx

I. Almendros, D. Gozal

Table 1 Studies focused on effects of intermittent hypoxia and melanoma. Abbreviations: OSA: Obstructive sleep apnea; VEGF: Vascular endothelial growth factor; HIF: Hypoxia-inducible factor. Study

Cell line

Intermittent hypoxia pattern

Field of study

Main effects of intermittent hypoxia

Almendros et al. (2012a)

B16F10 (mouse melanoma)

20 s 6% O2 40 s 21% O2 6 h/day–14 days

OSA

- Accelerates tumor growth and more extended necrotic areas.

Almendros et al. (2012b)

B16F10 (mouse melanoma)

20 s 6% O2 40 s 21% O2 14 days

OSA

- Augments circulating VEGF and tumor vascularization. - Obesity has not synergistic effects with IH.

Almendros et al. (2013a)

B16F10 (mouse melanoma)

20 s 6% O2 40 s 21% O2 28 days

OSA

- Increases lung metastasis and mortality.

Almendros et al. (2013b)

B16F10 (mouse melanoma)

30 min 2% O2

OSA

- Melanoma cells presented increased proliferation in co-culture with macrophages.

OSA

- Increases lung metastases, tumor vascularization and extended necrotic areas.

30 min 21% O2 4 days Eubank et al. (2013)

B16F10 (mouse melanoma)

80 s 5% O2 160 s 21% O2 13 days

Li et al. (2016)

B16F10 (mouse melanoma)

Non specified

OSA

- Increases lung metastases which are partially blocked by antioxidant tempol.

Martinive et al. (2006)

B16F10 (mouse melanoma)

1 h 7% O2 30 min 21% O2 (x3 cycles)

Cancer

- Reduces radiotherapy-induced apoptosis.

Perini et al. (2016)

B16F10 (mouse melanoma)

30 s 7% O2 30 s 21% O2 8 h/day–14 days

OSA

Increases markers of melanoma aggressiveness.

Rofstad et al. (2010)

A-07 (human melanoma)

In vivo: 10 min 8% O2 10 min 21% O2 (x12 cycles) In vitro: 30 min 8% O2 30 min 21% O2

Cancer

- Increases blood perfusion, vascularization and lung metastasis. - Upregulates HIF and VEGF expression.

- Upregulates of VEGF in vitro. - Pre-exposure of A-07 cells to IH in vitro did not increased lung colonization potential.

(x6 cycles)

2. Effects of intermittent hypoxia on different types of cancer

Furthermore, Martinive et al. showed that murine melanoma cells exposed to IH (three cycles of 1 h 7% O2 alternating with 30 min 21% O2) showed reduced magnitude of radiotherapy-induced apoptosis of melanoma cells (Martinive et al., 2006). In the context of OSA, Almendros et al. published 3 separate studies employing melanoma as an experimental model to study tumor growth (Almendros et al., 2012a), metastasis (Almendros et al., 2013a) and potential interactions with obesity (Almendros et al., 2012b). These studies corroborated that high frequency IH, such as occur in OSA, promotes similar effects on melanoma as previously reported by Rofstad and co-workers (Rofstad et al., 2010). Using a mouse melanoma model (B16F10), Almendros et al. exposed tumor bearing mice to IH (20 s 6% O2 followed by 40 s 21% O2), 6 h per day. After 14 days, tumors of mice exposed to IH exhibited a two-fold increase in weight compared to those exposed to intermittent room air (Almendros et al., 2012a). However, despite the original assumptions obesity did not exert synergistic effects with IH, suggesting that OSA and obesity are concurrent enhancers of tumor growth but do not do not promote each other to further tumor malignant properties among obese patients (Almendros et al., 2012b). Interestingly, a clinical study showed that the relationship OSA-cancer was more robust in non-obese patients, suggesting that obesity may mask the harmful effects of OSA on cancer (Nieto et al., 2012). In addition, in lean mice exposed to IH increased tumor vascularization and VEGF levels were uncovered. In a subsequent study, IH for 28 days increased metastatic lung dissemination from the primary inoculum of melanoma tumors (Almendros et al., 2013a). In this model, melanoma cells were more likely to leave the site of the primary tumor, intravasate to the circulation, resist anoikis, evade

2.1. Melanoma Melanoma is one of the most studied models of cancer when examining the harmful effects of IH on tumor malignancy (Table 1). In the first published study focused on tumor hypoxia, Rofstad et al. investigated whether exposure of melanoma cells to cyclic hypoxia could change its metastatic potential (Rofstad et al., 2010). These investigators employed a humanized model of mice (BALB/c nu/nu) bearing A-07 human melanoma xenografts. The day after tumor cell inoculation, the mice were exposed to 12 cycles of 20 min that alternated 8% O2 with room air until the tumor xenografts reached a volume of 100 mm3. Interestingly, the primary tumors of mice exposed to cyclic hypoxia presented higher blood perfusion and vascularization and also resulted in an increased number of lung metastases. These investigators also proposed that the increased vascularization observed in IH-exposed mice could be mediated by the upregulation of vascular endothelial growth factor (VEGF), and primarily mediated by hypoxia-inducible factors (HIFs). The authors concluded that the increased vascularization and blood perfusion that emerged under cyclic hypoxia could facilitate tumor cell intravasation into circulation explaining the increased lung metastasis under these conditions. In addition to the in vivo experiments, Rofstad et al. observed that melanoma cells exposed to cyclic hypoxia in vitro (6 cycles of 30 min of 8% O2 alternating with 30 min at 21% O2) increased VEGF production when compared to continuous normoxia. However, exposure of A-07 melanoma cells to cyclic hypoxia prior to inoculation failed to increase the number of metastases in mice.

2

Respiratory Physiology & Neurobiology xxx (xxxx) xxx–xxx

I. Almendros, D. Gozal

regulated, such as interleukin (IL)-6 as well as of CXCL1 (C-X-C motif ligand 1) (KC) and macrophage inflammatory protein 2 (MIP-2), along with increased leukocyte content in the tumor in those tumors from mice exposed to cyclic hypoxia. More recently, the effects of high frequency IH on lung cancer was investigated in the context of OSA-cancer. These studies have been carried out only on adenocarcinoma, and no data are so far available on other types of lung cancer. In the initial study, Almendros et al. preexposed mice to IH (90 s 8% O2 followed by 90 s 21% O2) or room air (21% O2) for two weeks (Almendros et al., 2014). Then, 105 lung murine adenocarcinoma TC1 cells were implanted subcutaneously in the right flank, while the mice continued their corresponding IH and RA exposures. Three weeks later, the tumors were excised and weighted. Similar to the previous melanoma findings, the size and weight of tumors of those mice exposed to IH were ∼2 fold greater than those in RA conditions (Almendros et al., 2014). However, in contrast to melanoma, this type of solid epithelial tumor allowed for assessment of tumor invasion toward adjacent healthy tissues. In this context, a significant higher number of mice presented evidence of invasion to surrounding tissues in the IH group (Almendros et al., 2014). The authors also assessed the potential contribution of the immune system and other tumor stromal cells in the adverse cancer outcomes observed under IH. First, in an in vitro model, murine macrophages (RAW 264.7) were pre-exposed to either IH (30 min 2% O2, 30 min 21% O2) or room air (21% O2). After 48 h, melanoma or lung cancer cells were seeded in single culture or in co-culture with macrophages for an additional 48 h while maintaining IH or RA conditions, respectively (Almendros et al., 2013b). The findings confirmed that co-culture of tumor cells with macrophages increased their proliferation. However and perhaps more interestingly, the application of IH on tumor cells in a single culture did not confer higher proliferative rates, but the co-culture with macrophages revealed that these immune cells under IH enhance the proliferation of tumor cells (Almendros et al., 2013b). The in vivo model of lung adenocarcinoma showed that IH promotes a shift toward a protumoral phenotype in tumor-associated macrophages (TAMs) (Almendros et al., 2014). The phenotypic changes were assessed by measuring some relevant surface markers by proteomics and flow cytometry. In particular, Almendros et al. identified 85 proteins that are de-regulated under IH rwhen compared to control TAMs. Interestingly, IFIT1, IFIT3, TAP1, and TAP2, which are all proteins that are involved in the interferon response and in antigen processing and presentation. In parallel, cell surface expression of CD86 and CD40 (M1 markers) and TFRC and CD206 (M2 markers) in TAMs were also measured. Accordingly with the proteomic analyses, down-regulation of M1 markers and no effects on M2 marker expression emerged in IH-exposed tumors. Thus, from both technical approaches, a reduction in the surface expression of genes characteristic of the classical activation pathway in macrophages occurred, suggesting that IH attenuates the pro-inflammatory phenotype in TAMs. Furthermore, the pro-tumoral macrophage phenotype was confirmed by additional functional in vitro assays. In particular, TAMs isolated from tumors of mice exposed to IH promoted lung cancer cell proliferation, migration, invasion, and extravasation across an endothelial monolayer compared to TAMs exposed to normoxia (Almendros et al., 2014). Changes induced by IH on the recruitment of host immune cells to the tumor were also assessed (Almendros et al., 2015). Specifically, adipose tissue macrophages (ATMs) from the ipsilateral surrounding adipose tissue (AT) and contralateral AT to the tumor were quantified. In addition, markers of resident and bone-marrow derived monocytes were also employed to assess the potential source of TAMs (Almendros et al., 2015). The authors found that the AT surrounding the tumor appears to serve as a major reservoir and source of resident macrophages to the tumor (∼10x higher than bone-marrow derived monocytes). Most importantly, the changes on macrophage phenotype strongly depended in the presence or absence of tumor. In this context, IH induced a pro-inflammatory phenotype in the contralateral non-

immune recognition and physical stress and eventually home in and proliferate in distant organs. The increased metastatic capacity of melanoma cells under IH was also accompanied by a higher mortality rate in the mice exposed to IH. Similar results were obtained when melanoma cells were directly injected into the lateral tail vein (Almendros et al., 2013a). Eubank and collaborators showed remarkably similar findings, whereby mice were pre-exposed for three days to IH (80 s 5% O2 followed by 160 s 21% O2) or intermittent room air prior to tumor implantation and then continued for another 10 days (Eubank et al., 2013). At the end of these experiments, although no differences in tumor growth were apparent, they found a ∼ 8-fold increase in pMel17 mRNA in the circulation and a 1.4-fold increase in the lungs in those mice exposed to IH, thereby indicating enhanced tumor cell intravasation and pulmonary metastases under IH (Eubank et al., 2013). Further histopathological analysis showed larger necrotic areas and vascularization under IH exposures. IH was able to promote changes in the expression of some matrix metalloproteinases which are known to participate in the degradation of the extracellular matrix during invasion processes. In a recent study, Li et al. described that the higher number of melanoma metastasis in mice exposed to IH could be mediated in part by IH-induced oxidative stress and enhanced activation of nuclear factor kappa B (NFκ-B) transcription (Li et al., 2016). Perini and colleagues revealed that exposures to IH (30 s 7% O2 followed by 30 s normoxia) in mice increased the expression of some markers that have been characteristically related to melanoma tumor aggressiveness, including caspase-1, Ki-67, PCNA, and Melan-A when compared to the control normoxic group (Perini et al., 2016). Although the authors found an overall 50% increase in tumor size in those mice exposed to IH, it did not reach to statistical significance probably due to the small number of mice employed in their study. In general, most of the studies carried out in melanoma tumor models showed that chronic IH increases tumor vascularization, thereby providing some explanation as tothe increased tumor growth and metastasis observed under this condition. However data on the molecular pathways remainscarce. IH in melanoma tumors resulted in the activation of HIF, which could me mediated directly by hypoxia or by a concurrent increase in NFκ-B transcription (Rius et al., 2008). The increased HIF and its downstream molecule VEGF have been proposed as the main regulators of vascularization, and consequently the IH-induced melanoma aggressiveness. However, new mechanistic studies are still sorely required to explore the source of VEGF in those tumors, and to ascertain the potential role of HIF, possibly by employing transgenic mice or inhibitors. 2.2. Lung adenocarcinoma An increasing number of studies have examined the effect of IH on lung tumor behavior over the last decade (Table 2). In 2010, Liu et al. investigated the expression of some key hypoxia-associated genes in response to chronic exposures to IH (Liu et al., 2010). In this study, two different models (chronic and low frequency intermittent hypoxia) were employed along with two lung cancer cell lines (A549 and H446). The exposures to IH consisted in 20 repeated cycles of 24 h in hypoxia (0.1% O2) followed by 72 h under normoxia (20% O2). After IH treatments, both lung cancer cell lines were more proliferative and presented higher resistance to radiotherapy. Also, chronic exposures to IH induced de-regulation of several hypoxia-inducible genes and other genes related to apoptotic processes. Tellier et al. investigated whether cancer-related cyclic hypoxia could enhance tumor-promoting Inflammation (Tellier et al., 2015). These investigators implanted i.m. 106 Lewis lung carcinoma (LLC) cells in the left flank. Once the tumors reached ∼10 mm in diameter, the mice were subjected to either room air (4.5 h 21% O2) or cyclic hypoxia (3 cycles of 1 h 7% O2 interrupted by 30 min of room air). After exposures, they found that cycling hypoxia promoted an increase of cyclo-oxygenase 2 (COX-2) and VEGF expression. In addition, other pro-inflammatory cytokines were also up3

Respiratory Physiology & Neurobiology xxx (xxxx) xxx–xxx

I. Almendros, D. Gozal

Table 2 Works studying intermittent hypoxia-induced alterations on lung carcinoma. Abbreviations: OSA: Obstructive sleep apnea; VEGF: Vascular endothelial growth factor; HIF: Hypoxiainducible factor; TAMs: Tumor-associated macrophages; PTGS2: Prostaglandin synthase 2; GzmB: Granzyme B; IFN- γ: Interferon gamma; CTLs: Cytotoxic T Lymphocytes; COX-2: Cyclooxygenase-2; PGE-2; Prostaglandin E2; IL-6: Interleukin 6. Study

Cell line

Intermittent hypoxia pattern

Field of study

Main effects of intermittent hypoxia

Akbarpour et al. (2017)

TC1 (mouse adenocarcinoma)

90 s 8% O2 90 s 21% O2 12 h/day 42 days

OSA

- CTLs showed reductions of GzmB, perforin and IFN- γ. - Increased stemness markers in TC1 tumor cells.

Almendros et al. (2014)

TC1 (mouse adenocarcinoma)

90 s 8% O2 90 s 21% O2 12 h/day

OSA

- Increased tumor growth and invasiveness. - Enhanced polarization toward M2 phenotype in TAMs - TAMs promoted higher tumor cell proliteration, migration, invasion and endothelial disruption.

Almendros et al. (2015)

TC1 (mouse adenocarcinoma)

OSA

- Polarization of TAMs depends on the presence/absence of tumor microenvironment. - Promotes higher macrophage recruitment from adipose tissues.

OSA

- Increases exosome secretion into circulation. - Exosomes facilitate lung tumor proliferation, invasion, migration and extravasation.

OSA

- Blockade of COX-2 attenuated IH-induced tumor growth and avoid M2 polarization in TAMs. - Tumor cells and macrophages are sources of IH-induced PGE2.

42 days 90 s 8% O2 90 s 21% O2 12 h/day 42 days Almendros et al. (2016)

TC1 (mouse adenocarcinoma)

90 s 8% O2 90 s 21% O2 12 h/day 42 days

Campillo et al. (2017)

LLC (mouse adenocarcinoma)

20 s 6% O2 40 s 21% O2 6 h/day 35 days

Cortese et al. (2015)

TC1 (mouse adenocarcinoma)

90 s 8% O2 90 s 21% O2 12 h/day 42 days

OSA

- Increased plasma circulating cell free DNA. - Induces epigenetic alterations.

Liu et al. (2010)

A549 and H446 (human adenocarcinoma)

24 h 0.1% O2

Cancer

- Increased proliferation. - Lower radiosensitivity. - Increased invasiveness.

72 h 20% O2 (x20 cycles) Tellier et al. (2015)

LLC (mouse adenocarcinoma)

1 h 7% O2 30 21% O2 (x3 cycles)

Cancer

- Enhanced leukocyte infiltration in tumors. - Upregulation of PTGS2 and IL-6 genes.

collaborators (Akbarpour et al., 2017). In this study, the mice were subjected to IH consisting of alternating cycles of 90 s (6% FiO2 followed by 21% FiO2, 20 cycles/h) for 12 h per day. After 2 weeks, 105 TC1 cells were subcutaneously injected and mice were maintained under antecedent RA and IH conditions for an additional 28 days. At the end of the exposures, tumors under IH were ∼3-fold bigger and were more invasive toward adjacent tissues than those exposed to RA. Evaluation of cytotoxic T lymphocytes (CTLs), which were defined as CD3+ CD8+ GzmB+ T cells, showed reductions of this anti-tumoral immune population under IH exposures. In addition, the characterization of CD8+ tumor infiltrating lymphocytes (TILs) showed reductions in granzyme B (GzmB), perforin, and interferon-γ (IFN-γ) expression levels under IH exposures, all of which are critical for the anti-tumoral cytotoxic activity of these cells. Indeed, the expression of those molecules is essential in the context of an intact immune surveillance system, and has been related with poor cancer prognosis. Furthermore, cytolytic assays of ex-vivo CTLs revealed that these cells effectively killed RA tumor cells, but were much less effective in lysing IH-exposed tumor cells. Authors also determined the effect of IH on cancer stem cells showing that IH exposure induces an increase in some stemness surface markers (Oct4+ and CD44+ CD133+) in TC1 tumor cells (Akbarpour et al., 2017). More recent studies have been focused on the identification of potential mechanisms involved in IH-induced increases in lung cancer malignant properties. Consequently, the epigenetic changes induced by IH in tumor cells were investigated (Cortese et al., 2015). This work

tumor containing AT (versus RA conditions), which is in accordance with previous data indicating that inflamed macrophages in the AT can contribute to the metabolic consequences attributable to OSA (Almendros and Garcia-Rio, 2017; Carreras et al., 2015; Gileles-Hillel et al., 2016; Gileles-Hillel et al., 2017; Gozal et al., 2017). However, macrophages in the ipsilateral AT where the tumor resided showed a pro-tumoral phenotype (M2) which was more pronounced under IH. This switch toward the M2 pro-tumoral phenotype could be due to the release of hypoxia inducible genes from tumor cells and/or the higher presence of immune suppressor cells, such as regulatory T lymphocytes (Tregs) and myeloid derived suppressor cells (MDSCs) in the ipsilateral AT as well as in the tumor stroma (Gao et al., 2014; Tiemessen et al., 2007). Both, Tregs and MDSCs have been postulated as key contributors to tumor progression and macrophage recruitment into the tumor stroma. Accordingly, while the total number of TAMs within the tumors of mice exposed to IH was higher compared to those under RA conditions, the number of macrophages in the ipsilateral AT was reduced, suggesting a more pronounced mobilization of macrophages from the AT surrounding the tumor under IH toward the tumor itself. This work also revealed an increased number of adipose stem cells (ASCs) in the ipsilateral AT. These cells have been ascribed a potential role in the increased extracellular matrix stiffness which could contribute to the increased tumor progression and invasiveness assessed under IH (Chaudhuri et al., 2014). In addition to macrophages, the alterations of effector T lymphocyte activity and cancer cell stemness were also assessed by Akbarpour and 4

Respiratory Physiology & Neurobiology xxx (xxxx) xxx–xxx

I. Almendros, D. Gozal

invasion was higher in IH vs normoxia-exposed mice. However, Ce administration did not prevent the increased invasiveness induced by IH. Assessment of M2 and M1 markers in TAMs showed that Ce attenuates the M2 polarization induced by IH. In fact, LLC1 cells increased their proliferation when co-cultured with TAMs obtained from IH-exposed mice, but not when those mice were treated with Ce. Finally, by using high frequency in vitro IH to expose macrophages and tumor cells, evidence that these cells constitute potential sources of PGE2 was confirmed. In general, the studies investigating the effects of IH on lung cancer showed increased lung adenocarcinoma growth and invasiveness in both in vitro and in vivo experimental models. These studies suggest a pivotal role of the immune response involving TAMs, lymphocytes and other immune suppressor cells. From molecular biology analyses, HIF and COX inflammatory pathways seem to participate in the increased tumor malignancy observed under IH conditions. Also, IH can promote stemness and induce epigenetic changes on cancer cells explaining in part an increased resistance or tumor escape from immunosurveillance. However, these studies have only employed lung adenocarcinoma as a model of lung cancer. Thus, the response of other types of lung cancer cells to IH still remains unknown.

employed the same experimental setting as previously described by Almendros et al. (2014) but included both xenografted and non-tumor injected (control) mice. At the end of the experiments and congruent with previous findings, tumors from mice exposed to IH were almost 2fold greater compared to mice exposed to RA conditions. As expected, those mice bearing tumors showed increased plasma circulating cell free DNA (cirDNA) concentrations. More interestingly, application of IH exposures increased plasma cirDNA concentrations in both tumor bearing and control mice. Plasma cirDNA was also significantly increased in mice bearing evidence for invasive tumors compared to those without invasive tumors. In addition, authors detected epigenetic variations in the XenoIH group in gene clusters within chromosomes 7, 13, 14 and X. Specifically, they identified regions of significant loss of signal at 7qB3 and XqF5 loci which are associated to malignant phenotypes in the lung in orthologous regions in human (15q15.3 and Xp22.2 for 7qB3 and XqF5, respectively). In a second mechanistic study, Almendros et al. investigated whether circulating exosomes released under IH conditions could facilitate tumor malignancy (Almendros et al., 2016). Exosomes are small vesicles (30–100 nm in diameter) which are ubiquitously manufactured within the endosomal system of cells, and are secreted into the extracellular space. Exosomes contain information in the form of DNA, lipids, protein, mRNA and also microRNA species, and can transfer these to other cells to modulate their transcriptome and phenotype. Currently, exosomes are considered as an universal mechanism for cell-to-cell communication and have been recently associated with increased tumor progression and metastasis in cancer. In this work, authors exposed adenocarcinoma bearing and non-tumor bearing mice to either IH or RA, as previously reported (Almendros et al., 2014). The amount of circulating exosomes was increased in those mice exposed to IH compared with RA. These exosomes were derived from multiple cell sources, such as platelets, endothelial cells, endothelial progenitor cells, and monocytes. After miRNA expression profiling analyses, those exosomes isolated from mice exposed to IH exhibited alterations in 11 distinct miRNAs compared to RA conditions. Furthermore, integrative pathway analyses revealed 246 transcription factors potentially associated with the differentially affected IH exosomes. Most of these transcription factors were linked with mmu-miR-671-5p, mmu-miR-6418-5p, mmu-miR-609, mmu-miR-882, mmu-miR-92a-3P, mmu-miR-3082-5p and mmu-miR-5113. In addition, some of these transcription factors have been related to networks involved in cancer biology and T-cell inflammatory processes, including forkhead box M1, SMAD family member 1, signal transducer and activator of transcription 6, zinc fingers, metastasis-associated 1, T-cell acute lymphocytic leukemia 1, and tumor growth factor-β-induced factor homeobox 2. At the end of the exposures, exosomes were isolated from blood and were co-cultured with naïve murine lung adenocarcinoma cells (TC1). As a result of co-culture with naïve tumor cells, the exosomes obtained from IH conditions facilitated tumor cell proliferation, migration, invasion and disruption of endothelial cell barrier integrity and tight junctions, indicating that exosome-based intercellular communications may operate both locally and remotely to alter tumor cell phenotype and promote the increased malignant properties associated with IH patterns mimicking sleep apnea. In a recent study, the potential role of cyclooxygenase-2 (COX-2) on IH-induced lung cancer malignancy was assessed in LLC1 tumors (Campillo et al., 2017). COX-2 downstream metabolite prostaglandin E2 (PGE2) can promote tumor cell proliferation, migration and invasion either directly, or indirectly through induction of a tumor-permissive microenvironment characterized by impaired immunosurveillance and increased angiogenesis. In this study, mice were exposed to RA or IH conditions and were treated with celecoxib (Ce), a widely employed COX-2 inhibitor, or placebo. As in previous studies, the application of IH increased the size of the LLC1 tumors (Almendros et al., 2014). In particular, the weight of tumors from mice exposed to IH experienced ∼2-fold increase vs. RA. Under IH conditions, daily administration of Ce promoted a marked reduction in tumor weight. As previously reported, the number of mice presenting

2.3. Breast carcinoma The contribution of continuous hypoxia on breast cancer biology has been widely studied. There is now well established evidence showing that hypoxia promotes stemness of breast cancer cells (Semenza, 2017; Xie et al., 2016) as well as enhance extracellular matrix remodeling (Gilkes et al., 2014). However, the number of studies using the IH paradigm is still very scarce and none of them were based on OSA-related IH patterns (Table 3). Gutske et al. investigated on the SUM149PT cell line, a human triple-negative inflammatory breast cancer cell line, the aggressiveness of breast cancer after 45–60 days of IH exposures in vitro (Gutsche et al., 2016). Their results showed that chronic IH will upregulate some proteins related to tumor invasiveness, such as tenascin-C (TNC), matrix metalloproteinase 9 and cyclooxygenase-2 (COX-2). The expression of TNC was also dependent on oxidative stress-induced NF-κB activation (Gutsche et al., 2016). In another study, two human breast cancer cell lines (MDA-MB 231 and BCM2) were subjected to hypoxia (1% O2) and nutrient deprivation for one week, and the surviving cells were then maintained for 1–3 weeks depending on their viability (Louie et al., 2010). In these studies, 10% of the breast cancer cells became non-adherent after the hypoxic cycle. Interestingly, new non-adherent populations of cancer cells appeared after repeated hypoxia-re-oxygenation cycles, and such non-adherent cells were highly tumorigenic when injected in a nude mouse model, when compared to the hypoxia-exposed adherent or to non-exposed cells. Also, the IH-exposed non-adherent cells presented increased cancer stem cell markers and assessment of their colony formation capability showed that these IH-exposed non-adherent cells also have an enhanced ability to self-renew. Verduzco et al. exposed breast carcinoma cells (MCF10A) to 50 cycles of 16 h of 0.2% O2 followed by 8 h of normoxia (Verduzco et al., 2015), and showed that chronic in vitro exposures to IH promoted permanent changes in the expression of p53, E-cadherin, and HIF-1α. Notably, these changes were associated with an increased drug resistance and increased survival under hypoxic conditions (Verduzco et al., 2015). 2.4. Kidney cancer Renal cell carcinoma prognosis has been linked to the upregulation of hypoxia-inducible genes which are associated with neovascularization. In a recent study, Vilaseca et al. studied whether IH mimicking OSA could increase kidney cancer malignancy (Vilaseca et al., 2017) (Table 3). In this work, a subcutaneous kidney cancer was induced in mice by injecting subcutaneously RENCA cells into the left flank and 5

Respiratory Physiology & Neurobiology xxx (xxxx) xxx–xxx

I. Almendros, D. Gozal

Table 3 Effects of intermittent hypoxia in neuroblastoma, glioblastoma, breast, liver and kidney cancer. Abbreviations: OSA: Obstructive sleep apnea; VEGF: Vascular endothelial growth factor; HIF: Hypoxia-inducible factor; NF-κB: Nuclear factor kappa B; TNC: Tenascin-C; MMP-9: Metalloproteinase-9; ROS: Reactive oxygen species; COX-2: Cyclooxygenase-2. Study

Cell line

Intermittent hypoxia pattern

Field of study

Main effects of intermittent hypoxia

Bhaskara et al. (2012)

NB1691 (human neuroblastoma)

24 h 1% O2

Cancer

- Upregulates HIF-1α and VEGF. - Increases stemness associated genes in neuroblastoma cells.

24 h 21% O2 1–10 cycles Bhaskara et al. (2014)

SH-SY5Y (human neuroblastoma)

48 h 1% O2

Cancer

- Increases migration and upregulates osteoclastogenic factors, VEGF and the receptor activator of nuclear factor kappa-B ligand.

Cancer

- Increases HIF-1α and NF-κB protein levels through ROS activation

Cancer

- Upregulates TNC, MMP-9 and COX-2 through oxidative stress-induced NFkB.

Cancer

- Repeated episodes of hypoxia/reoxygenation increased stemness and tumorigenic potential of breast cancer cells.

Cancer

- Reduces radiotherapy-induced apoptosis.

Cancer

- Reduced radiotherapy-induced apoptosis.

Cancer

- Promotes permanent changes in the expression of p53, E-cadherin, and HIF1α in cancer cells conferring increased drug resistance and survival under hypoxia.

OSA

- Does not change tumor growth.

24 h 21% O2 10 cycles Chen et al. (2015)

U251, U87 and GBM8401 (human glioblastoma)

1 h 0.5–1% O2 30 min 21% O2 (3 cycles)

Gutsche et al. (2016)

SUM149PT (human breast cancer)

24 h 0.2% O2 24 h 20.9% O2 45 days

Louie et al. (2010)

MDA-MB 231/BCM2 (human breast cancer)

1 wk 1% O2 1–3 wks 21% O2 (x3 cycles)

Martinive et al. (2006)

FsaII (human fibrosarcoma)

1 h 7% O2 30 min 21% O2 (x3 cycles)

Martinive et al. (2006)

TLT (human liver cancer)

1 h 7% O2 30 min 21% O2 (x3 cycles)

Verduzco et al. (2015)

MCF10A (human breast cancer)

16 h 0.2% O2

8 h 21% O2 Vilaseca et al. (2017)

RENCA (mouse kidney cancer)

20 s 6% O2 followed by 40 s 21% O2 6 h/day–35 days

- Increases VEGF synthesis in macrophages. - Promotes tumor vascularization.

upregulated the expression of stem cell-associated genes in neuroblastoma cells. In another study, Bhaskara and colleagues used neuroblastoma SH-SY5Y cells, and showed that IH increased their migratory behavior and upregulated osteoclastogenic factors, VEGF and the receptor activator of nuclear factor kappa-B ligand (RANKL) (Bhaskara et al., 2014). In addition, conditioned medium from IH-exposed neuroblastoma cells enhanced osteoclastogenesis and triggered the activation of extracellular-signal-regulated kinase (ERK), c-Jun N-terminal kinase (JNK), and p38 in macrophages. Finally, neuroblastoma cells pre-treated with IH showed evidence of accelerated tumor growth when injected into mice (Bhaskara et al., 2014).

mice were then exposed to either IH (cycles of 20 s 6% O2 followed by 40 s 21% O2, 6 h/day for 21 days) or normoxia (RA). IH promoted an increased percentage of vascular progenitor cells and endothelial cells in the tumor tissue. As reported previously in the melanoma model, plasma VEGF levels were significantly higher in the IH-exposed group. At the end of the experiments, immunohistological analyses revealed that IH increased vascularization of kidney tumors, leading to the assumption that the IH-induced vascularization may be mediated in part by the increased circulating levels of VEGF. Using in vitro exposures to IH similar to those applied in vivo, Vilaseca et al. identified that only macrophages presented a marked increase in VEGF expression in response to IH, without evidence of VEGF expression changes in endothelial and tumor cells. In contrast with previous findings in melanoma and lung adenocarcinoma, IH did not accelerate tumor growth or alter the phenotype of TAMs.

2.6. Glioblastoma Chen et al. investigated whether short cyclic hypoxia exposures as occurs in tumors could confer chemo-resistance to glioblastoma cells (Chen et al., 2015) (Table 3). To carry out this study, they exposed three different glioblastoma cell types (U251, U87 and GBM8401) to IH consisting of 3 cycles of 0.5–1% O2 for 1 h interrupted by 5% CO2 and air for 30 min or control in vitro. IH induced HIF-1α and NF-κB protein levels in U87 and U251 glioblastoma cells. Accordingly, they explored whether IH-induced reactive oxygen species (ROS) could contribute to HIF-1α and NF-κB activation. To this end, the glioblastoma cells were treated with Tempol during IH exposures, and this anti-oxidant treatment abrogated the increases in HIF-1α and NF-κB activation under IH.

2.5. Neuroblastoma Bhaskara et al. exposed NB1691 neuroblastoma cells in vitro to 1, 5 or 10 cycles of hypoxia (IH) and RA (Bhaskara et al., 2012) (Table 3). Each cycle consisted in 24 h in 1% O2 followed by 24 h recovery under RA (21% O2). After these IH challenges, NB1691 presented increased HIF-1α expression with a marked accumulation of HIF-1α protein in the nucleus of these cells, which also was associated with an increased expression of VEGF and with increased cell survival. IH also 6

Respiratory Physiology & Neurobiology xxx (xxxx) xxx–xxx

I. Almendros, D. Gozal

Acknowledgement

Finally, tempol treatment also attenuated the chemo-resistance induced in tumor cells by IH (Chen et al., 2015).

This work was supported by SEPAR (086/2014; 139/2015). 2.7. Liver carcinoma

References

Similar to glioblastomas, liver carcinoma has been appraised in regards to assessment of the effects of IH on cancer treatment (Table 3). Martinive et al. exposed transplantable liver tumor (TLT) implanted in mice to three cycles of 1-h hypoxia (7% O2) interrupted by 30-min reoxygenation periods (21% O2) (Martinive et al., 2006). One hour later, the mice were locally irradiated (10 Gy) and tumors were harvested after 24 h to evaluate the extent of cellular apoptosis. IH influenced the survival of both vascular and tumor cells in vivo. Specifically, IH promoted a ∼2-fold reduction in the extent of apoptotic-positive vascular structures, as well as reduced the extent of radiotherapy-induced apoptosis of tumor cells.

Akbarpour, M., Khalyfa, A., Qiao, Z., Gileles-Hillel, A., Almendros, I., Farre, R., Gozal, D., 2017. Altered CD8+ T-Cell lymphocyte function and TC1 cell stemness contribute to enhanced malignant tumor properties in murine models of sleep apnea. Sleep 40. Almendros, I., Garcia-Rio, F., 2017. Sleep apnoea, insulin resistance and diabetes: the first step is in the fat. Eur. Respir. J. 49. Almendros, I., Montserrat, J.M., Ramirez, J., Torres, M., Duran-Cantolla, J., Navajas, D., Farre, R., 2012a. Intermittent hypoxia enhances cancer progression in a mouse model of sleep apnoea. Eur. Respir. J. 39, 215–217. Almendros, I., Montserrat, J.M., Torres, M., Bonsignore, M.R., Chimenti, L., Navajas, D., Farre, R., 2012b. Obesity and intermittent hypoxia increase tumor growth in a mouse model of sleep apnea. Sleep Med. 13, 1254–1260. Almendros, I., Montserrat, J.M., Torres, M., Dalmases, M., Cabanas, M.L., CamposRodriguez, F., Navajas, D., Farre, R., 2013a. Intermittent hypoxia increases melanoma metastasis to the lung in a mouse model of sleep apnea. Respir. Physiol. Neurobiol. 186, 303–307. Almendros, I., Wang, Y., Farre, R., Gozal, D., 2013b. Intermittent hypoxia enhances melanoma cell proliferation only in co-culture with macrophages. Am. J. Respir. Crit. Care Med. 187, A6061. Almendros, I., Wang, Y., Becker, L., Lennon, F.E., Zheng, J., Coats, B.R., Schoenfelt, K.S., Carreras, A., Hakim, F., Zhang, S.X., et al., 2014. Intermittent hypoxia-induced changes in tumor-associated macrophages and tumor malignancy in a mouse model of sleep apnea. Am. J. Respir. Crit. Care Med. 189, 593–601. Almendros, I., Gileles-Hillel, A., Khalyfa, A., Wang, Y., Zhang, S.X., Carreras, A., Farre, R., Gozal, D., 2015. Adipose tissue macrophage polarization by intermittent hypoxia in a mouse model of OSA: effect of tumor microenvironment. Cancer Lett. 361, 233–239. Almendros, I., Khalyfa, A., Trzepizur, W., Gileles-Hillel, A., Huang, L., Akbarpour, M., Andrade, J., Farre, R., Gozal, D., 2016. Tumor cell malignant properties are enhanced by circulating exosomes in sleep apnea. Chest 150, 1030–1041. Bhaskara, V.K., Mohanam, I., Rao, J.S., Mohanam, S., 2012. Intermittent hypoxia regulates stem-like characteristics and differentiation of neuroblastoma cells. PLoS One 7, e30905. Bhaskara, V.K., Mohanam, I., Gujrati, M., Mohanam, S., 2014. Intermittent hypoxia effect on osteoclastogenesis stimulated by neuroblastoma cells. PLoS One 9, e105555. Campillo, N., Jorba, I., Schaedel, L., Casals, B., Gozal, D., Farré, R., Almendros, I., Navajas, D., 2016. A novel chip for cyclic stretch and intermittent hypoxia cell exposures mimicking obstructive sleep apnea. Front. Physiol. 29 (7), 319. Campillo, N., Torres, M., Vilaseca, A., Nonaka, P.N., Gozal, D., Roca-Ferrer, J., Picado, C., Montserrat, J.M., Farre, R., Navajas, D., et al., 2017. Role of cyclooxygenase-2 on intermittent hypoxia-induced lung tumor malignancy in a mouse model of sleep apnea. Sci. Rep. 7, 44693. Carreras, A., Zhang, S.X., Almendros, I., Wang, Y., Peris, E., Qiao, Z., Gozal, D., 2015. Resveratrol attenuates intermittent hypoxia-induced macrophage migration to visceral white adipose tissue and insulin resistance in male mice. Endocrinology 156, 437–443. Chaudhuri, O., Koshy, S.T., Branco da, C.C., Shin, J.W., Verbeke, C.S., Allison, K.H., Mooney, D.J., 2014. Extracellular matrix stiffness and composition jointly regulate the induction of malignant phenotypes in mammary epithelium. Nat. Mater. 13, 970–978. Chen, W.L., Wang, C.C., Lin, Y.J., Wu, C.P., Hsieh, C.H., 2015. Cycling hypoxia induces chemoresistance through the activation of reactive oxygen species-mediated B-cell lymphoma extra-long pathway in glioblastoma multiforme. J. Transl. Med. 13, 389. Cortese, R., Almendros, I., Wang, Y., Gozal, D., 2015. Tumor circulating DNA profiling in xenografted mice exposed to intermittent hypoxia. Oncotarget 6, 556–569. Dal, M.M., Brant, A., Blackford, A.L., Griffin, J.F., Shindo, K., Barkley, T., Rezaee, N., Hruban, R.H., Wolfgang, C.L., Goggins, M., 2016. Obstructive sleep apnea and pathological characteristics of resected pancreatic ductal adenocarcinoma. PLoS One 11, e0164195. DeBerardinis, R.J., Chandel, N.S., 2016. Fundamentals of cancer metabolism. Sci. Adv. 2, e1600200. Dehne, N., Mora, J., Namgaladze, D., Weigert, A., Brune, B., 2017. Cancer cell and macrophage cross-talk in the tumor microenvironment. Curr. Opin. Pharmacol. 35, 12–19. Dewan, N.A., Nieto, F.J., Somers, V.K., 2015. Intermittent hypoxemia and OSA: implications for comorbidities. Chest 147, 266–274. Eubank, T., Sherwani, S., Peters, S., Gross, A., Evans, R., Magalang, U.J., 2013. Intermittent hypoxia augments melanoma tumor metastases in a mouse model of sleep apnea. Am. J. Respir. Crit. Care Med. 187, A2302. Forster, J.C., Harriss-Phillips, W.M., Douglass, M.J., Bezak, E., 2017. A review of the development of tumor vasculature and its effects on the tumor microenvironment. Hypoxia (Auckl.) 5, 21–32. Gao, J., Wu, Y., Su, Z., Amoah, B.P., Jiao, Z., Bie, Q., Lu, L., Wang, S., Xu, H., 2014. Infiltration of alternatively activated macrophages in cancer tissue is associated with MDSC and Th2 polarization in patients with esophageal cancer. PLoS One 9, e104453. Gileles-Hillel, A., Kheirandish-Gozal, L., Gozal, D., 2016. Biological plausibility linking sleep apnoea and metabolic dysfunction. Nat. Rev. Endocrinol. Gileles-Hillel, A., Almendros, I., Khalyfa, A., Nigdelioglu, R., Qiao, Z., Hamanaka, R.B., Mutlu, G.M., Akbarpour, M., Gozal, D., 2017. Prolonged exposures to intermittent

2.8. Limitations and future challenges There is rapidly increasing evidence linking tumor aggressiveness and resistance to therapy with the presence of a hypoxic tumor microenvironment. However, the studies investigating IH are still relatively scarce. A first limitation is the lack of consensus and homogeneity in the magnitude and frequency of IH exposures in both cancer and OSA-related studies. This is possibly less of a critical issue in those studies focused in tumor hypoxia since the IH pattern in the tumor is aberrant and not repetitive. However, in those studies carried out in vitro and in vivo aiming to study the effects of OSA on tumor behavior, it is essential that the IH profiles used mimic as much as possible the IH experienced in moderate/severe OSA patients. In this regard, it is also important to mention that the mechanisms involved in the activation of HIF could presumably depend on the magnitude and frequency of IH. It is well known that hypoxia reduces the activity of prolyl hydroxylases promoting the accumulation of HIF facilitating the hypoxic tolerance through the activation of several transcription factors. However, at high frequency cycling hypoxia, it is probable that HIF stabilization occurs via distinct mechanisms mediated by the presence of pro-inflammatory mediators or oxidative reactive species. To overcome some of these limitations, a novel in vitro model has been developed very recently (Campillo et al., 2016). In the context of OSA, only a restricted number of cancer types has been investigated to date. As depicted in this review, breast cancer, neuroblastoma, glioblastoma and liver carcinoma could also respond and be affected to high frequency IH. As future challenges in cancer-OSA research, we foresee the following items for specific attention: i) to increase the knowledge of OSAcancer based on other cancer types ii) to better understand the mechanisms governed through IH and modulating cancer malignancy iii) to study the potential interactions of IH-induced cancer malignancy with other factors that co-exist in OSA, such as sleep fragmentation, obesity, aging, menopause, etc… and iv) the potential effects of coexistent IH in the context of other respiratory diseases such as lung fibrosis and chronic pulmonary obstructive disease (overlap of IH with chronic hypoxia, acidosis, etc…)

3. Summary In this review, we have delineated the available evidence linking IH and tumor biology. The epidemiological evidence appears to support that at least some types of solid tumors will manifest increased incidence and adverse prognosis in the context of OSA, and this is an area of rapidly evolving and intensified research (Dal et al., 2016; Gozal et al., 2015; Lee et al., 2017; Li et al., 2017; Nieto et al., 2012; Owens et al., 2016). Therefore the translational and clinical implications of the studies reviewed herein are obvious and should stimulate the field. 7

Respiratory Physiology & Neurobiology xxx (xxxx) xxx–xxx

I. Almendros, D. Gozal

Nieto, F.J., Peppard, P.E., Young, T., Finn, L., Hla, K.M., Farre, R., 2012. Sleep-disordered breathing and cancer mortality: results from the Wisconsin Sleep Cohort Study. Am. J. Respir. Crit. Care Med. 186, 190–194. Owens, R.L., Gold, K.A., Gozal, D., Peppard, P.E., Jun, J.C., Lippman, S.M., Malhotra, A., 2016. Sleep and Breathing … and cancer? Cancer Prev. Res. (Phila) 9, 821–827. Perini, S., Martinez, D., Montanari, C.C., Fiori, C.Z., 2016. Enhanced expression of melanoma progression markers in mouse model of sleep apnea. Rev. Port. Pneumol. 2006. Qiu, G.Z., Jin, M.Z., Dai, J.X., Sun, W., Feng, J.H., Jin, W.L., 2017. Reprogramming of the tumor in the hypoxic niche: the emerging concept and associated therapeutic strategies. Trends Pharmacol. Sci. Rius, J., Guma, M., Schachtrup, C., Akassoglou, K., Zinkernagel, A.S., Nizet, V., Johnson, R.S., Haddad, G.G., Karin, M., 2008. NF-kappaB links innate immunity to the hypoxic response through transcriptional regulation of HIF-1alpha. Nature 453, 807–811. Rofstad, E.K., Gaustad, J.V., Egeland, T.A., Mathiesen, B., Galappathi, K., 2010. Tumors exposed to acute cyclic hypoxic stress show enhanced angiogenesis, perfusion and metastatic dissemination. Int. J. Cancer 127, 1535–1546. Semenza, G.L., 2017. Hypoxia-inducible factors: coupling glucose metabolism and redox regulation with induction of the breast cancer stem cell phenotype. EMBO J. 36, 252–259. Tarrado-Castellarnau, M., de, A.P., Cascante, M., 2016. Oncogenic regulation of tumor metabolic reprogramming. Oncotarget 7, 62726–62753. Tellier, C., Desmet, D., Petit, L., Finet, L., Graux, C., Raes, M., Feron, O., Michiels, C., 2015. Cycling hypoxia induces a specific amplified inflammatory phenotype in endothelial cells and enhances tumor-promoting inflammation in vivo. Neoplasia 17, 66–78. Tiemessen, M.M., Jagger, A.L., Evans, H.G., van Herwijnen, M.J., John, S., Taams, L.S., 2007. CD4 + CD25 + Foxp3+ regulatory T cells induce alternative activation of human monocytes/macrophages. Proc. Natl. Acad. Sci. U. S. A. 104, 19446–19451. Toth, R.K., Warfel, N.A., 2017. Strange bedfellows: nuclear factor, erythroid 2-Like 2 (Nrf2) and hypoxia-inducible factor 1 (HIF-1) in tumor hypoxia. Antioxidants (Basel) 6. Verduzco, D., Lloyd, M., Xu, L., Ibrahim-Hashim, A., Balagurunathan, Y., Gatenby, R.A., Gillies, R.J., 2015. Intermittent hypoxia selects for genotypes and phenotypes that increase survival, invasion, and therapy resistance. PLoS One 10, e0120958. Vilaseca, A., Campillo, N., Torres, M., Musquera, M., Gozal, D., Montserrat, J.M., Alcaraz, A., Touijer, K.A., Farre, R., Almendros, I., 2017. Intermittent hypoxia increases kidney tumor vascularization in a murine model of sleep apnea. PLoS One 12, e0179444. Wong, P.P., Bodrug, N., Hodivala-Dilke, K.M., 2016. Exploring novel methods for modulating tumor blood vessels in cancer treatment. Curr. Biol. 26, R1161–R1166. Xie, J., Xiao, Y., Zhu, X.Y., Ning, Z.Y., Xu, H.F., Wu, H.M., 2016. Hypoxia regulates stemness of breast cancer MDA-MB-231 cells. Med. Oncol. 33, 42.

hypoxia promote visceral white adipose tissue inflammation in a murine model of severe sleep apnea: effect of normoxic recovery. Sleep 40. Gilkes, D.M., Semenza, G.L., Wirtz, D., 2014. Hypoxia and the extracellular matrix: drivers of tumour metastasis. Nat. Rev. Cancer 14, 430–439. Gozal, D., Farre, R., Nieto, F.J., 2015. Putative links between sleep apnea and cancer: from hypotheses to evolving evidence. Chest 148, 1140–1147. Gozal, D., Gileles-Hillel, A., Cortese, R., Li, Y., Almendros, I., Qiao, Z., Khalyfa, A.A., Andrade, J., Khalyfa, A., 2017. Visceral white adipose tissue after chronic intermittent and sustained hypoxia in mice. Am. J. Respir. Cell Mol. Biol. 56, 477–487. Gutsche, K., Randi, E.B., Blank, V., Fink, D., Wenger, R.H., Leo, C., Scholz, C.C., 2016. Intermittent hypoxia confers pro-metastatic gene expression selectively through NFkappaB in inflammatory breast cancer cells. Free Radic. Biol. Med 101, 129–142. Hakim, F., Wang, Y., Zhang, S.X., Zheng, J., Yolcu, E.S., Carreras, A., Khalyfa, A., Shirwan, H., Almendros, I., Gozal, D., 2014. Fragmented sleep accelerates tumor growth and progression through recruitment of tumor-associated macrophages and TLR4 signaling. Cancer Res. 74, 1329–1337. Lee, S., Kim, B.G., Kim, J.W., Lee, K.L., Koo, D.L., Nam, H., Im, J.P., Kim, J.S., Koh, S.J., 2017. Obstructive sleep apnea is associated with an increased risk of colorectal neoplasia. Gastrointest. Endosc. 85, 568–573. Li, A., Liang, M., Fang, Y., Cao, J., Chen, B., 2016. Antioxidant tempol ameliorates intermittent hypoxia-induced melanoma lung metastasis in a mouse model of OSA. Am. J. Respir. Crit. Care Med. 193, A7741. Li, L., Lu, J., Xue, W., Wang, L., Zhai, Y., Fan, Z., Wu, G., Fan, F., Li, J., Zhang, C., et al., 2017. Target of obstructive sleep apnea syndrome merge lung cancer: based on big data platform. Oncotarget 8, 21567–21578. Liu, Y., Song, X., Wang, X., Wei, L., Liu, X., Yuan, S., Lv, L., 2010. Effect of chronic intermittent hypoxia on biological behavior and hypoxia-associated gene expression in lung cancer cells. J. Cell. Biochem. 111, 554–563. Louie, E., Nik, S., Chen, J.S., Schmidt, M., Song, B., Pacson, C., Chen, X.F., Park, S., Ju, J., Chen, E.I., 2010. Identification of a stem-like cell population by exposing metastatic breast cancer cell lines to repetitive cycles of hypoxia and reoxygenation. Breast Cancer Res. 12, R94. Manoochehri, K.H., Afshar, S., Najafi, R., 2016. Hypoxia: a double-edged sword in cancer therapy. Cancer Invest. 34, 536–545. Martinive, P., Defresne, F., Bouzin, C., Saliez, J., Lair, F., Gregoire, V., Michiels, C., Dessy, C., Feron, O., 2006. Preconditioning of the tumor vasculature and tumor cells by intermittent hypoxia: implications for anticancer therapies. Cancer Res. 66, 11736–11744. Matsumoto, S., Yasui, H., Mitchell, J.B., Krishna, M.C., 2010. Imaging cycling tumor hypoxia. Cancer Res. 70, 10019–10023. Muz, B., de la Puente, P., Azab, F., Azab, A.K., 2015. The role of hypoxia in cancer progression, angiogenesis, metastasis, and resistance to therapy. Hypoxia (Auckl.) 3, 83–92.

8