Role of RKIP protein in the modulation of lung cancer cell metabolism

Universidade do Minho Escola de Medicina Joana Filipa Barbosa Pinheiro Role of RKIP protein in the modulation of lung cancer cell metabolism janeiro de 2021 UMinho | 2021 Joana Filipa Barbosa Pinheiro Role of RKIP protein in the modulation of lung cancer cell metabolism Role of RKIP protein in the modulation of lung cancer cell metabolism Doutora Olga Catarina Lopes Martinho Doutora Sara Costa Granja Universidade do Minho Escola de Medicina ii Direitos de autor e condições de utilização do trabalho por terceiros Este é um trabalho académico que pode ser utilizado por terceiros desde que respeitadas as regras e boas práticas internacionalmente aceites, no que concerne aos direitos de autor e direitos conexos. Assim, o presente trabalho pode ser utilizado nos termos previstos na licença abaixo indicada. Caso o utilizador necessite de permissão para poder fazer um uso do trabalho em condições não previstas no licenciamento indicado, deverá contactar o autor, através do RepositóriUM da Universidade do Minho. Atribuição-NãoComercial-SemDerivações CC BY-NC-ND https://creativecommons.org/licenses/by-nc-nd/4.0/ iii Agradecimentos O trabalho que aqui apresento é para mim uma conquista. E porque não cheguei aqui sozinha, gostaria de agradecer a todas as pessoas que ajudaram a tornar este momento possível. Desta forma começo por agradecer à pessoa que me acolheu e fez “crescer” neste mundo da ciência, à minha orientadora, Olga Martinho. Obrigada por tudo! Por todos os ensinamentos, por toda a paciência e compreensão, pelo teu humanismo e bondade, pela incansável dedicação, e sobretudo por acreditares em mim e me inspirares a dar o meu melhor todos os dias. Obrigado por me teres feito crescer quer a nível profissional, quer como pessoa, sem a tua ajuda nada disto seria possível! À Doutora Sara Granja pela coorientação e prontidão em embarcar neste projeto. Obrigada por toda ajuda, disponibilidade, apoio prestado, e todas as sugestões que ajudaram à realização deste projeto. À Diana e à Raquel, um grande obrigado por tudo o que me ensinaram, por estarem sempre lá quando precisei, pelos momentos, amizade e por alegrarem aquele laboratório como só vocês sabem! À minha Gabi e à minha Patrícia, obrigada por todo o companheirismo, pela amizade, e por toda a força que me dão. A todas as colegas de laboratório, por toda a ajuda, pela alegria, pela companhia e sobretudo pela paciência em ter de repetir as frases vezes e vezes sem conta até eu ouvir. A ti Diogo, um grande obrigado por todo o apoio incondicional, companheirismo, carinho e dedicação em me animares mesmo nos piores dias. Obrigada por me motivares todos os dias! Aos meus amigos de sempre e para sempre, e às melhores que bioquímica me deu, um grande obrigado por me ouvirem, por todos os momentos, e sobretudo pela vossa amizade e apoio. Aos meus pais e às minhas “pequenas”, Gabi e Matilde, obrigada por serem os pilares da minha vida, por todo o amor e carinho, e acima de tudo por acreditarem em mim, são o meu orgulho! À minha restante família, obrigada por todo o apoio e orgulho que demonstram em mim. The work presented in this thesis was performed in the Life and Health Sciences Research Institute (ICVS), Minho University. Financial support was provided by grants from the ICVS Scientific Microscopy Platform, member of the national infrastructure PPBI - Portuguese Platform of Bioimaging (PPBI-POCI01-0145-FEDER-022122; by National funds, through the Foundation for Science and Technology (FCT) - project UIDB/50026/2020; and by the projects NORTE-01-0145-FEDER-000013 and NORTE-01-0145FEDER-000023,supported by Norte Portugal Regional Operational Programme (NORTE 2020), under the PORTUGAL 2020 Partnership Agreement, through the European Regional Development Fund (ERDF); partially by the PTDC/MED-ONC/31423/2017 project (Ref: POCI-01-0145-FEDER-031423). iv Statement of integrity I hereby declare having conducted this academic work with integrity. I confirm that I have not used plagiarism or any form of undue use of information or falsification of results along the process leading to its elaboration. I further declare that I have fully acknowledged the Code of Ethical Conduct of the University of Minho. Assinado por : JOANA FILIPA BARBOSA PINHEIRO Num. de Identificação: BI15331707 Data: 2021.01.21 22:04:16+00'00' Papel da proteína RKIP na modelação do metabolismo celular em cancro do pulmão v Resumo Os tumores sólidos são caracterizados por uma dependência exacerbada de glucose, sendo a reprogramação metabólica das células tumorais considerada um hallmark do cancro. Hoje é sabido que esta reprogramação em tumores é frequentemente controlada e induzida pela ativação oncogénica de vias de sinalização celular. O cancro do pulmão, é um exemplo clássico dos tipos tumorais em que já foi identificada uma alta frequência de mutações oncogénicas, cujo papel na ativação aumentada de vias de sinalização celular, está mais que estabelecido. A complexa maquinaria de ativação concomitante de várias vias de sinalização celular tem sido associada à aquisição de um fenótipo metabólico altamente heterogéneo. Desta forma, o uso da reversão metabólica como abordagem terapêutica torna-se muito mais complexa, sendo necessário a contínua investigação na procura de novas moléculas que possam estar envolvidas nesta reprogramação metabólica heterogénica. A proteína supressora tumoral RKIP, alterada em vários tipos tumorais como o cancro de pulmão, é funcionalmente uma proteína sinalizadora intracelular impactante na agressividade tumoral, através da regulação de várias vias de sinalização celular, que entre outras, podem controlar o metabolismo tumoral. Assim, juntamente com o conhecimento de que a RKIP poderá estar diretamente associada a alterações metabólicas, surge-nos a hipótese de que a RKIP pode ser uma das proteínas subjacentes à heterogeneidade metabólica em cancro do pulmão. Para explorar e validar a nossa hipótese, começamos por fazer uma análise in silico , utilizando a base de dados TCGA, para determinar se existe uma assinatura molecular de vias metabólicas associadas à expressão de RKIP, em pacientes com adenocarcinoma pulmonar. Ademais, foi primeiramente realizada a caracterização metabólica dos modelos in vitro escolhidos para o estudo, a fim posteriormente manipular geneticamente a expressão da RKIP e determinar quais as alterações metabólicas especificamente associadas à RKIP. De forma geral, observamos in silico que a RKIP está associada à alteração de expressão de genes, proteínas e vias de sinalização mais relacionadas com fosforilação oxidativa do que glicólise. Funcionalmente, a manipulação genética da RKIP alterou principalmente os padrões de expressão de proteínas metabólicas, e discretamente os níveis de glucose e lactato in vitro , indicando de uma forma geral que a RKIP inibe o metabolismo glicolítico. Em conclusão, os nossos resultados forneceram as primeiras provas funcionais de que a RKIP poderá de facto modular o metabolismo celular em cancro do pulmão, abrindo uma enorme janela para trabalhos futuros. Palavras-chave: Cancro do pulmão, Reprogramação metabólica, RKIP Role of RKIP protein in the modulation of lung cancer cell metabolism vi Abstract Solid tumours are characterized by an exorbitant glucose metabolism, being tumour cell metabolic reprogramming now considered a hallmark of cancer. Oncogenic signalling has been linked to metabolic rewiring of cancer cells. Lung cancer in specific, is often driven by mutations that lead to oncogenic activation of tyrosine kinases, known to produce metabolic alterations in tumour cells. Growing evidence suggest that tumour cells acquire heterogeneous metabolic phenotypes to withstand the hindrances associated to tumour growth and progression, which makes the search for novel players in tumour metabolic reprogramming of most importance. As a stablish tumour suppressor, RKIP downregulation has been associated with tumoral aggressiveness and patient’s poor prognosis, in several tumour types, including lung cancer. Being a master regulator of several intracellular signalling pathways, together with recent findings suggesting that RKIP could potentially be associated with alterations in tumour cell metabolism, RKIP arises to us as one of the potential proteins behind cancer metabolic heterogeneity. which give support to our hypothesis. Therefore, it was herein aimed to explore RKIP role as a modulator of lung cancer metabolism. First, using the TCGA database, an in silico analysis was performed to determine whether there is a RKIP-associated metabolic signature in lung cancer patients. Further, it was performed a metabolic characterization of the in vitro models chosen for the work, in order to genetically manipulate their RKIP expression and determine the specific RKIP-associated metabolic changes. In general, we observed that RKIP is significantly associated with alterations in metabolism-related genes and signalling pathways involved in the regulation of metabolic processes, both at mRNA and protein level. Specifically, RKIP expression is positively correlated with oxidative phosphorylation-related genes, and inversely with glycolysis-related genes, in lung adenocarcinoma patients. Functionally, we found that, although not significant, RKIP genetic manipulation in vitro led to some metabolic alterations, mainly in the expression patterns of metabolic proteins, but also in glucose and lactate levels, which in general indicate that RKIP leads to inhibition of glycolytic metabolism. In conclusion, our results provided the first functional evidence that RKIP could in fact modulate lung cancer cell metabolism and opened a huge window for future work. Keywords: Lung cancer, Metabolic reprogramming, RKIP vii Table of contents Direitos de autor e condições de utilização do trabalho por terceiros ..................................................... ii Agradecimentos .................................................................................................................................. iii Statement of integrity .......................................................................................................................... iv Resumo............................................................................................................................................... v Abstract.............................................................................................................................................. vi Table of contents ............................................................................................................................... vii List of abbreviations ............................................................................................................................ ix Table of Figures ..................................................................................................................................xii List of Tables ..................................................................................................................................... xiii CHAPTER 1: General Introduction ........................................................................................................ 1 1.1 Cancer ............................................................................................................................................. 2 1.1.1 Lung cancer ................................................................................................................................. 3 1.2 Raf Kinase Inhibitor Protein (RKIP) .................................................................................................... 5 1.2.1 RKIP as a signalling modulator...................................................................................................... 6 1.2.2 RKIP role in cancer ....................................................................................................................... 8 1.3 Cancer metabolism reprogramming ................................................................................................ 11 1.3.1 Molecular basis of cancer metabolism reprogramming ................................................................ 12 1.3.2 Targeting cancer cell metabolism ................................................................................................ 16 1.3.3 Metabolic reprogramming in lung cancer .................................................................................... 18 CHAPTER 2: Research Objectives ..................................................................................................... 20 CHAPTER 3: Materials and Methods ................................................................................................. 22 3.1 Cell lines and cell culture ................................................................................................................ 23 3.2 Drugs ............................................................................................................................................. 23 3.3 In silico analysis ............................................................................................................................. 23 3.4 In vitro RKIP genetic modulation ..................................................................................................... 24 3.5 Cell viability assay ........................................................................................................................... 24 3.6 Extracellular glucose and lactate measurements .............................................................................. 25 3.7 Western blot analysis ...................................................................................................................... 25 3.8 Immunofluorescence analysis ......................................................................................................... 26 3.9 Statistical analysis .......................................................................................................................... 27 CHAPTER 4: Results ......................................................................................................................... 28 4.1 RKIP and lung cancer: In silico analysis ........................................................................................... 29 4.2 Metabolic characterization of lung cancer cell lines .......................................................................... 32 CHAPTER 1: General Introduction 2 1.1 Cancer Cancer is a heterogeneous and complex disease that remains one of the main causes of death worldwide, with rapid growth both in incidence and mortality. About 18.1 million new cases of cancer and 9.6 million cancer deaths were estimated to have occurred in 20181. Thus, understanding the biology and complexity underlying the evolution of this disease is important for reducing its global burden. In short, cancer is a disease of the genome, caused by the acquisition of somatic mutations in several genes that collectively dictate malignant growth2,3. In the early 2000s, Hanahan and Weinberg proposed the hallmarks of cancer as an organizing principle for rationalizing the complexities of neoplastic disease4. The hallmarks of cancer comprise several biological capabilities acquired during the multistep development of human tumours. Among them are, sustaining proliferative signalling, evading growth suppressors, resisting cell death, inducing angiogenesis, enabling replicative immortality, activating invasion and metastasis, genome instability and mutation, tumour-promoting inflammation, avoidance immune destruction and deregulation of cellular energetics5 (Figure 1). Figure 1: Hallmarks of cancer. Schematic illustration of the set of features acquired by tumours cells that enables tumour growth and metastatic development. Adapted from5. 3 1.1.1 Lung cancer Lung cancer remains the most diagnosed cancer and the leading cause of cancer deaths worldwide in both sexes combined, with 2.1 million (11.6% of the total cases) new cases and 1.8 million (18.4% of the total cancer deaths) deaths predicted in 20181. The main risk factor of lung cancer is tobacco use, accounting for more than 80% of the cases in Western populations6. Although the numbers are high, major shifts in the global distribution of this disease have occurred, reflecting the temporal changes in the tobacco consumption patterns7. Declines in smoking, due to tobacco control measures applied in the last decades, as well as improvements in early detection and treatment, have resulted in a continuous decline in the lung cancer incidence and mortality rates. This incidence decrease has shown to be much faster in men than in women, which may be due to the historical and gender differences in tobacco uptake and cessation1,6. Besides that, was recently reported that women have a higher risk to develop lung cancer upon smoking than men, which can also explain this tendency8. Lung cancer can be categorized in two main categories based on their histological features: small cell lung cancer (SCLC) and non-small cell lung cancer (NSCLC). The latter, which accounts for approximately 85% of all lung cancer cases, is further divided into three subtypes, adenocarcinoma (AC), squamous cell carcinoma (SCC) and large cell carcinoma (LCC)9,10. Of all the types, lung adenocarcinomas are the most common and have the poor survival rates, with an average five-year relative survival rate of 5%, mainly because of late-stage detection and lack of treatments in advanced stages6,11. Although smoking is unquestionably the leading cause of lung cancer, approximately 25% of lung cancer cases worldwide are not attributable to tobacco consumption10. Striking differences in the epidemiological, clinical and molecular characteristics of lung cancer that arises in never smokers versus smokers have been identified in the last decade10. Globally, lung cancer in never smokers is more common in women and in East Asia, and has been associated with environmental exposures, pollution, other occupational carcinogens, and inherited genetic susceptibility1,12. Despite all the major histological types are associated with tobacco use, the association is stronger for SCLC and SCC than for adenocarcinoma10. Molecular epidemiology studies demonstrate that the genomic landscape of lung tumorigenesis between never smokers and smokers, providing more evidence that these cancers arise through different molecular mechanisms (Figure 2). As example, adenocarcinomas in never smoker patients frequently present mutations in the epidermal growth factor receptor ( EGFR ) gene. By contrast, smokers’ patients often harbour mutations in KRAS and TP53 genes10,12,13. This can be explained by the exposure of smokers to highly carcinogenic agents like tobacco smoke, that contain agents that are known to bind to DNA 4 creating DNA adducts which, eventually, can lead to mutations in KRAS or TP53 genes, which are key events in lung cancer pathogenesis10,14,15. Although EGFR and KRAS mutations in lung tumours are most entirely mutually exclusive, both lead to constitutive activation of receptor tyrosine kinase (RTK) signalling9,10 (Figure 2). RTK signalling pathways, namely RAS/MAPK, PI3K/AKT and STAT pathways, govern fundamental physiological processes, such as cell proliferation, differentiation, metabolism and cell death and survival16,17. Notably, dysregulation and constitutive activation of these signalling pathways have been implicated in the initiation, progression and metastatic spread of lung cancer, either by mutations in RTKs (e.g., EGFR ) or in its downstream effectors (e.g., KRAS ), which makes them an attractive target for lung cancer therapy16. Figure 2: Two pathways to adenocarcinoma. A) Ligand binding to EGFR induces homoand hetero-dimerization of the receptor, resulting in activation of downstream effectors including the RAS/MAPK (mitogen activated protein kinase), PI3K (phosphatidylinositol 3-kinase)/AKT, and STAT (signal transducer and activator of transcription) pathways that lead to cell proliferation, survival and many other effects associated with carcinogenesis. The EGFR pathway is frequently activated in never smokers by mutations in the EGFR gene. B) In smokers, mutations of the KRAS gene often occur, resulting in the release of growth factors that bind to EGFR, activating its pathway. In addition, Ras directly activates the PI3K/AKT pathway. The result of KRAS or EGFR mutations are virtually identical, and mutations of both genes in adenocarcinomas of the lung are rarely seen. Other methods of activation of these pathways include gene amplification and mutations in BRAF , PIK3CA (a subunit of PI3K), and mutations in other receptors tyrosine kinase (RTKs), such as ERBB2 (also known as HER2 ). Adapted from10. 5 Concerning lung cancer treatment, the first approach is usually surgical resection, followed by chemotherapy or radiotherapy, depending on the staging of the tumour18. Despite the high relevance and use of the traditional cytotoxic treatments, they present several limitations such as the lack of selectivity to tumour cells, the gain of resistance and systemic toxicity19. Fortunately, over the past two decades, important advancements in lung cancer treatment have been achieved. An expansion on the understanding of these malignancies, coupled with advances in the molecular characterization of tumours, enable the development of targeted therapies that introduce the era of personalized medicine2022. In fact, it is now the standard approach, in the treatment of lung cancers, to genotype tumours at diagnosis, that allow individualized therapy, leading to higher efficiency of the treatment, less secondary effects, and ultimately an improvement of patient life quality and survival13,23. For instance, the use of small molecule tyrosine kinase inhibitors (TKI), such as erlotinib and gefitinib, for lung adenocarcinoma patients harbouring EGFR mutations, show longer progression-free survival and more favourable tolerability24,25. Although promising, NSCLC tumours treated with these first-generation TKIs inevitably develop resistance. Several resistance mechanisms were already described and other target therapy drugs, such as second and third generation TKIs (afatinib and Osimertinib, respectively), were clinical developed and are now used in lung cancer patients treatment26. Besides that, the emergence of immunological checkpoint inhibitors (ICIs) has signalled a new direction for lung cancer treatment. Current approved ICIs for NSCLC treatment include anti-PD-1 antibodies, nivolumab and pembrolizumab, as well as anti-PD-L1 antibody atezolizumab18,27-29. Unfortunately, the clinical benefits of ICIs have been proven limited and unsatisfactory, with the overall response rate of monotherapy about 10-20%, which may be due to the complexity of the tumour immune microenvironment26. Despite the major progress in understanding the biology and management of lung cancer in the past decades, the overall cure and survival rates remain low, particularly in metastatic disease. Therefore, understanding the mechanisms behind treatment resistance and a continued search for key molecules involved in the regulation of lung tumorigenesis is necessary. 1.2 Raf Kinase Inhibitor Protein (RKIP) Raf kinase inhibitor protein (RKIP), first discovered as phosphatidylethanolamine-binding protein 1 (PEBP1), is a small (23kDa) cytosolic protein originally purified from bovine brain, that belongs to a highly conserved family of proteins30,31. This protein is widely expressed in normal human tissues and has been recognised for its important role in multiple physiological processes, such as membrane biosynthesis, spermatogenesis, cardiac output, neural development, inflammation and others31-33. Therefore, 6 dysregulated RKIP expression has been related with some human diseases, such as Alzheimer’s disease, metabolic and inflammatory disorders and cancer30-33. This multifunctional capacity of RKIP is associated with its involvement in the modulation of several signalling pathways (Figure 3) that control important cellular processes34-36. 1.2.1 RKIP as a signalling modulator Regarding RKIP role as a signal transduction modulator, this protein was first described as an endogenous inhibitor of the Raf-MEK-ERK (MAPK) pathway, accounting for its current name. Yeung et al . first reported that RKIP was able to block Raf-1/MEK interaction by acting as a competitive inhibitor of MEK phosphorylation and by reducing the affinity between Raf-1/MEK37 (Figure 3). Moreover, it was described that RKIP binds directly to the N-region of the Raf-1 kinase domain, preventing Ser338 and Tyr340/341 phosphorylation by PAK and Src kinases, needed for Raf-1 activation38,39. RKIP has been also reported to indirectly interfere with upstream activators of Raf-1, such as Gprotein coupled receptors (GPCR), being the nature of this interference dictated by the phosphorylation status of RKIP. Protein Kinase C (PKC)-mediated phosphorylation of RKIP at serine 153 decreases RKIP affinity for Raf-1 and increase its affinity for G protein-coupled receptor kinase 2 (GRK2), which is an endogenous inhibitor of GPCR activation40,41. Thus, the binding of phosphorylated RKIP to GRK2 leads to a dissociation of GRK2 from GPCR, allowing GPCR activation and phosphorylation of downstream targets, including Raf-1, which implies a role of RKIP as an endogenous modulator of cell response to growth factor stimuli42,43 (Figure 3). Furthermore, RKIP was identified as a negative modulator of nuclear factor kappa B (NF-κB) signalling, by antagonizing its upstream signal transducers44 (Figure 3). The inhibitory effect of RKIP is exerted by its association with the upstream kinases TAK, NIK, IKKα and IKKβ, abolishing their kinase activity, which results in the elimination of inhibitory kappa B (IκB) phosphorylation and degradation, blocking NF-κB translocation to the nucleus and consequently NF-κB-mediated transcription of several genes with anti-apoptotic features44,45. Likewise, it has recently described that RKIP acts as a negative modulator of the signal transducer and activator of transcription 3 (STAT3), Sonic hedgehog (Shh) signalling and Notch1 (Figure 3). Concerning STAT3, RKIP blocks its activation by preventing its phosphorylation by upstream kinases, controlling this way the transcription of genes related to cell growth, apoptosis, survival and differentiation46,47. Additionally, it was demonstrated that RKIP can bind to signal transducer Smoothened (SMO), keeping it inactive and consequently blocking the activation of the zinc-finger transcription factor 7 Gli1, the final target of Shh signalling, which is involved in the regulation of proliferation, differentiation and cancer stem cells (CSCs) activation48. Moreover, RKIP directly interacts with Notch1 preventing its proteolytic cleavage and the release of Notch intracellular domain (NCID), which stimulates the epithelial to mesenchymal transition (EMT) and metastasis49. In contrast, RKIP can also act as a positive modulator intracellular signalling, as it able to activate glycogen synthase kinase-3β (GSK3β) (Figure 3), that is involved in the regulation of many cellular functions besides its initially described role in glycogen metabolism50,51. Al Mulla and colleagues described that RKIP stabilizes GSK3β expression by preventing its phosphorylation at the inhibitory T390 residue by p38 MAPK, which is activated under oxidative stress augmented upon RKIP depletion or downregulation. The loss of RKIP de-represses GSK3β inhibition of oncogenic substrates causing stabilization of cyclin D, which induces cell-cycle progression, Snail, and Slug, which promote EMT50. Figure 3: RKIP as a modulator of intracellular signalling pathways. On the left, RKIP binds to IKK complex preventing Ikβ phosphorylation and degradation which ultimately blocks the translocation of NF-kβ to the nucleus. Next, RKIP binds to the Notch Intracellular Domain (NICD) preventing its proteolytic cleavage, inhibiting the translocation of NICD to the nucleus. In the middle, RKIP act as an inhibitor of the Shh signalling pathway by binding to the SMO receptor, keeping it inactive and preventing Gli1 transcription. Also in the middle, RKIP depletion enhances oxidative stress–mediated activation of the p38 8 MAPK, which, in turn, inactivates GSK3β by phosphorylating it at the inhibitory T390 residue. On the right, RKIP is bound to Raf preventing the phosphorylation of MEK by Raf leading to the inhibition of the Raf/MEK/ERK/STAT3 signalling. Further, RKIP blocks Snail through MAPK inhibition and NF2 stabilization. In the nucleus, Snail acts as a p53 suppressor GPCRs are desensitized and internalized in response to phosphorylation by GRK2. After cells stimulation (e.g. growth factors), PKCmediated phosphorylation of RKIP, at S153, inactivates RKIP as an inhibitor of Raf-1, and converts it to a GRK2 inhibitor. GPCR signalling through ERK/MAPK can therefore persist. Adapted from52. 1.2.2 RKIP role in cancer Given the major role of RKIP as a modulator of important intracellular signalling pathways that are often deregulated in cancer, RKIP has proved to play a crucial role in controlling both tumour aggressiveness and therapeutic response, being nowadays considered a well-establish tumour suppressor43. The first association between RKIP and cancer was established in prostate metastatic cell lines, which displayed lower RKIP expression levels when compared to primary tumour cell lines53. Besides that, Fu et al . also demonstrated that reestablishment of RKIP expression in the metastatic cell lines lead to an inhibition of their invasion capability, a decreased development of lung metastases, but did not affect the growth of the primary tumour53. This suggested that RKIP may not have an essential role in the primary tumour, but instead, has great importance as a metastasis suppressor, as evidenced in further studies on different tumour types43,54,55. In accordance, loss or reduction of RKIP expression has been associated with malignancy and poor prognosis in several solid tumours, as described by our group56-62 and others36,43,63. In fact, due to its negative association with metastasis, RKIP has been implicated as a strong predictive biomarker for metastatic risk in patients as well as an independent prognostic marker for overall survival and disease-free survival in many solid tumours (reviewed by us in52). Despite the growing importance of RKIP as a metastasis and prognostic marker in human tumours, the mechanisms behind RKIP downregulation remains elusive, however, it was already demonstrated that RKIP expression can be regulated at multiple levels36. For instance, at the epigenetic level, promotor methylation was pointed as a mechanism behind RKIP loss of expression in some tumour types, yet this mechanism seems to be tumour-type specific43,64-66. Regarding the regulation at the transcriptional level, some transcription factors, such as Snail and BACH1, can bind directly to RKIP promoter suppressing its transcription and expression36,67,68. Moreover, it was found that RKIP mRNA can be targeted and consequently suppressed by microRNAs, such as miR-27a69, miR-53470, miR-23a71 and miR-22472. Ultimately, as mention before, RKIP can be modulated at the post-transcriptional level by PKC-mediated 9 phosphorylation at Serine 153, which also accounts for the loss of RKIP activity in several cancer types40,41,73. As a master metastasis suppressor, RKIP exerts its inhibitory functions at different steps of the metastatic process, including those involved in its initiation, such as angiogenesis, EMT, cell migration and invasion43. It is now known that RKIP targets signalling networks that directly or indirectly regulate metastatic functions and tumour progression, which are mostly under the control of the central signalling pathways described above to be regulated by RKIP (Figure 3). For instance, Rosner and colleagues identified a pro-metastatic signalling cascade involving MYC/LIN28/let-7 as an RKIP target in breast cancer models. They demonstrated that, through its inhibitory action in the Raf/MEK/MAPK pathway, RKIP leads to inhibition of LIN28 transcription by MYC, which in turn, enhances let-7 expression74. The potentiation of let-7 expression by RKIP leads to a negative regulation of its targets, such as BACH1, a transcription factor that induces matrix metalloproteinase1 (MMP1) expression and promotes metastasis75,76. Furthermore, it has also been describe that RKIP can negatively regulate the early metastatic events, such as EMT, through the NF-κB/Snail/YY1/RKIP circuitry77,78. Snail, a well-described EMT inducer, is transcriptionally regulated by NF-κB signalling, that is known to repress both RKIP and E-cadherin transcription, two key proteins in cancer metastasis67. In turn, RKIP was demonstrated to inhibit NF-κB pathway, leading to negative regulation of EMT inducers, such as Snail and YY1 and to an upregulation of E-cadherin, supressing this way the EMT process79,80. Besides that, recent findings strongly suggest that RKIP’s anti-metastatic properties can also be mediated through modulation of the tumour microenvironment (TME), as recently reviewed by us33. It was demonstrated that RKIP modulates the breast cancer TME by controlling the infiltration of specific immune cells and secretion of pro-metastatic factors, for example by blocking the recruitment of prometastatic macrophages, through vital regulation of chemokines expression81,82. Moreover, Bainer et al . also demonstrated that gene expression in metastatic breast tumours was significantly correlated with gene expression in local stroma, and that alterations in stromal gene expression elicited by tumours positive or negative for RKIP expression is a better predictor of breast cancer subtype and patient survival that tumour gene expression alone83. Together, this emphasizes the importance of exploiting RKIP’s microenvironmental functions for a better understanding of its role in carcinogenesis. It is acknowledged that tumour cells often develop therapeutic resistance along with increased metastatic potential, resulting in a poorer overall survival. Notably, downregulation of RKIP expression has been associated to the development of resistance to conventional cytotoxic drugs in tumour cells61,84,85. In that sense, RKIP expression has also been implicated in the regulation of tumour cells resistance to 10 conventional chemotherapy, radiotherapy, and more recently to immune-mediated cytotoxicity43,86,87. As an apoptotic inducer, RKIP has shown to cause the re-sensitization of resistant tumours to therapy and also sensitivity to host immune-surveillance, via multiple interactions with signalling modules33,43,88. Examples of these RKIP-modulated cascades are the NF-κB, STAT3, Shh, MAPK and PI3K/AKT signalling pathways, which are involved in the regulation of both metastasis and tumour cell sensitivity to apoptotic stimuli43,48. Since RKIP is commonly downregulated in many human cancers, its induction could be an attractive approach to sensitize tumour cells to therapy response. In fact, some agents have already been reported as able to induct RKIP expression contributing this way to cells sensitization to apoptosis, such as rituximab89, nitric oxide (NO) donors79, didymin90 and the proteosome inhibitor NPI-005280. Concerning lung cancer in specific, the role of RKIP as a prognostic marker in these malignancies is not yet clear, as the studies regarding its expression and clinical significance are still scarce and not concordant among them, however a clinical association between low expression of RKIP and higher TNM stage or presence of lymph node metastasis is found across the studies (as reviewed in62). Nonetheless, the implications of RKIP in these tumours related signalling, progression and therapy resistance is quite evident62. For instance, it was demonstrated, using NSCLC cell lines, that RKIP is able to inhibit ERK and STAT3 phosphorylation and, consequently, supress cell migration and tumour metastasis46. Moreover, RKIP has shown to modulate EMT, invasion and metastasis in NSCLC through the modulation of EMTcontrolling pathways such as Nocth1 and the NF-κB/Snail/YY1/RKIP loop49,91. Further, RKIP was also identified as a p53 modulator in lung cancer by blocking Snail, that acts as a p53 suppressor, through MAPK signalling inhibition and NF2 stabilization92. Beyond this, it was demonstrated that RKIP reduction enhances radioresistance by activating the Shh signalling pathway in NSCLC48. Importantly, RKIP show to be a key molecular player in the modulation of NSCLC cells response to conventional therapy, such as the chemotherapeutic agents adriamycin and cisplatin, as well as radiotherapy48,69,93. Finally, Giovannetti et al. demonstrated that RKIP is also implicated in lung cancer cells response to targeted therapies. By studying the synergistic interaction between sorafenib, a multikinase inhibitor, and erlotinib, an EGFR inhibitor, they found that sorafenib-related reduction of AKT/ERK phosphorylation in erlotinib-resistant cells was associated with significant RKIP upregulation94. Taking these together, RKIP has already shown to be an important modulator of relevant signalling pathways, including MAPK, as well as to be both a driver and predictor of therapy response in lung cancer, which emphasizes the importance of a continued study of RKIP biological role in these malignancies. 17 and one of the most prescribed drugs in the world for the treatment of type 2 diabetes mellitus. Notably, retrospective clinical studies suggested that the use of metformin improve the prognosis of diabetic patients with multiple cancers and prevent tumour initiation and reoccurrence163,164. Moreover, metformin has shown to inhibit proliferation, cell survival and induce apoptosis in several cancer models165,166. It is now recognized that metformin acts as an anticancer agent by inhibiting mitochondrial ETC complex I (NADH: ubiquinone oxidoreductase), blocking this way mitochondrial ATP production also its biosynthetic capacity, which can lead to an increase of cellular energy stress, and ultimately cell death107,165. Furthermore, several studies demonstrated synergistic effects of metformin in cancer cells when combined with other treatments, such as glycolysis inhibitors, chemotherapeutic drugs, and targeted therapies155,167-169. Figure 6: Potential molecular targets in cancer cell metabolism. The pathways of central carbon metabolism are presented. Some of the metabolic enzymes deregulated in cancer that are currently being considered as molecular targets for therapy are marked with a target (shown as a pink circle in the figure). Five drugs that influence metabolism and have been tested in humans are shown in pink boxes. Adapted from156. 18 1.3.3 Metabolic reprogramming in lung cancer Lung cancer is a molecularly heterogeneous disease, therefore, understanding its biology is crucial for the development of effective therapies13. Like many other solid tumours, lung cancer requires a high intake of glucose, and FDG-PET scan already became a standard diagnostic tool to assess tumour extension9. Moreover, data from patients indicate that lung tumours are dependent on glucose metabolism, being the increased expression of glycolytic markers correlated with poor prognosis in lung cancer patients170-172. Despite the clear association between high glucose metabolism and poor prognosis, the pathways through which glucose is metabolized in this type of cancer remain unclear. Studies using surgical resections from NSCLC patients after infusion with radioactive glucose demonstrated that these tumours displayed enhancement of both glycolysis and OXPHOS, which was translated in the high levels of lactate and TCA cycle intermediates in tumour samples compared with normal tissue137,173. These observations suggested that glycolysis and OXPHOS can function in simultaneous if not in the same cancer cell at least in the same tumour, in which metabolic symbiosis can be established137. Hensley et al. also described that glucose oxidation via PDH, and TCA cycle were higher in NSCLC tumours compared to adjacent normal lung137. In addition, they demonstrated that NSCLC tumours oxidize multiple types of nutrients in vivo besides glucose, such as lactate, fatty acids and amino acids, and this was, in part, dependent on the perfusion status137. Notably, further studies reported the importance of glucose-derived lactate as a carbon source for the TCA cycle in tumours from NSCLC patients174,175. Despite commonalities in metabolic reprogramming, growing evidence suggest that metabolic changes in NSCLC tumours are highly heterogeneous137,176-178. Given the high mutation burden of lung cancer21,22, its genetic heterogeneity can also influence lung cancer metabolic diversity, since oncogenic signalling is tightly linked to metabolic reprogramming109. Indeed, some lung cancer driver mutations have been already described to produce metabolic alterations in these tumours179,180. In specific, NSCLC tumours are often driven by activation of tyrosine kinase signalling, by EGFR and KRAS mutations for instance9,13. Remarkably, activating mutations in EGFR gene promote metabolic rewiring in NSCLC by increasing aerobic glycolysis, PPP pathway, altered pyrimidine biosynthesis and redox metabolism181,182. Moreover, it was demonstrated that PI3K/AKT signalling plays an important role in regulation of metabolic activities in EGFR -mutant lung adenocarcinoma cells, by facilitating GLUT1 proper cellular localization181. Furthermore, KRAS mutations, that occur in approximately 30% of adenocarcinoma patients, have also been implicated in the metabolic reprogramming in NSCLC, including in the upregulation of glucose uptake, glutamine utilization and aerobic glycolysis151,154,183. For instance, a study has reported that KRAS - 19 mutated NSCLC cells express higher levels of enzymes involved in glycolysis and PPP when compared to non-malignant cells, suggesting alterations in glucose metabolism is NSCLC cells carrying mutant KRAS 184. Further, Kerr et al. described that glycolysis is acquired as a function dependent on the number of copies of mutant KRAS 185. Moreover, expression and activity of HK2 was induced by KRAS -mutation in NSCLC models. In particular, deletion of HK2 supressed KRAS -mutant lung cancer development in murine models and reduced cell growth both in vitro and in vivo by inhibiting nucleotides synthesis and glutamine-derived carbon utilization in TCA cycle122. In addition, KRAS mutations already show to enhance glutamine metabolism in vitro through PI3K/AKT dependent signalling in NSCLC cells186. Overall, the literature herein reviewed provides a strong evidence of metabolic rewiring in lung cancer. Concomitantly, the acquired heterogeneous metabolic phenotype by tumours cells to withstand the complex challenges linked to cancer progression137,176, is still puzzling and challenging to search for more and new oncogenic alterations behind this metabolic heterogeneity. Given the key role of RKIP in the regulation of several intracellular signalling pathways, more than just MAPK, and its recognized importance in tumours malignancy52 in a frequency which seems to be much higher than KRAS and with a higher penetrance among solid tumours, including lung cancer62, RKIP arises to us as one of the potential proteins behind cancer metabolic heterogeneity. To support our believe, there is on one hand, an in silico study reporting that RKIP is inversely correlated with genes that directly regulate cell metabolism, in clear cell renal cell carcinoma (ccRCC)187. By the other hand, it was demonstrated that the pro-metastatic transcription factor BACH1, which is negatively regulated by RKIP, promotes aerobic glycolysis in breast cancer188,189. Curiously, until now there are no studies exploiting the direct functional role of RKIP in the modulation of cancer cells metabolism. CHAPTER 2: Research Objectives 21 2. Research Objectives The reprogramming of cellular metabolism is now widely recognised as a hallmark of cancer. Remarkably, oncogenic signalling has been linked to tumour metabolic reprogramming, being cancer cells able to maintain growth factor–independent glycolysis and survival, through the expression of oncogenic kinases. Overall, the literature provides strong evidence of an oncogenic signalling dependent metabolism in lung cancer. In fact, lung cancers exhibit a high mutation burden, but specifically NSCLC tumours, are known to be driven by oncogenic activation of tyrosine kinases, such as mutations in KRAS and EGFR genes, and many other pathways, which are well described as modulators of metabolic rewiring in lung cancer cells. Nowadays, one of the emerging themes in cancer research relies on the ability of tumours cells acquiring heterogeneous metabolic phenotypes to withstand the complex challenges linked to cancer progression, consequently, there is a demand for the search of new oncogenic alterations that can be behind this metabolic heterogeneity. RKIP is a well-establish tumour suppressor, whose loss of expression has been observed in a plethora of solid tumours, including lung cancer. Herein, due to its important role in carcinogenesis and its master role in the regulation of several intracellular signalling pathways, more than the first described MAPK, which are known to be crucial in cancer cell metabolic rewiring, RKIP arises to us as one of the potential proteins behind lung cancer metabolic heterogeneity. Notably, recent evidence indirectly raised the suspicion that RKIP loss may be involved in tumoral metabolic processes, but without a direct functional validation. Thus, passing the redundancy, in the present study we hypothesized that RKIP could be a key player in the metabolic reprogramming of NSCLC tumours cells. To achieve our hypothesis, the present project was divided in four main aims: The first was to determine whether a RKIP-associated molecular signature is correlated with metabolic alterations in lung adenocarcinoma patients. For that, an extensive in silico analysis was performed using TCGA database. Secondly, it was intended to metabolic characterize the chosen NSCLC cell lines for cell proliferation, metabolic profile and RKIP expression. In third, to unravel the effect of RKIP on the modulation of NSCLC cells metabolism, we pursued to positively and negatively modulate its expression and determine the cells metabolic behaviour as well as its expression patterns of metabolic proteins, upon transfection. Finally, in fourth, it was planned to carry out a focused in silico analysis to validate and understand the pertinent obtained results. CHAPTER 3: Materials and Methods 23 3.1 Cell lines and cell culture In this study, a panel of different NSCLC cell lines were used (Table 1). All cells were grown and maintained in Dulbecco’s Modified Eagle’s Medium (DMEM, GIBCO, Invitrogen) supplemented with 10% of Fetal Bovine Serum (FBS, GIBCO, Invitrogen) and 1% penicillin/streptomycin (GIBCO, Invitrogen). For normoxia conditions, cells were incubated in a humidified atmosphere of 21% O2, 5% CO2 and 74% N2 at 37ºC. For hypoxia conditions, cells were placed in airtight chamber with hypoxic gas mixture 1% O2, 5% CO2 and 95% N2 and incubated at 37ºC for a specific period time, according to the assay carried out. Table 1: Non-small cell lung cancer cell lines used in this study. Cell line Histological type Relevant alterations A549 Adenocarcinoma KRAS mutant HCC827 Adenocarcinoma EGFR mutant H292 Mucoepidermoid Carcinoma EGFR and KRAS WT WT: Wild-type 3.2 Drugs 2-Deoxyglucose and metformin were purchased from Sigma-Aldrich and MedChemExpress (MCE), respectively. Both drugs were prepared in stock solutions of 200mM by dilution in sterile water and stored at -20ºC. In all experimental conditions the drugs were diluted in culture medium to a final concentration of 1.5mM. Cell culture medium was also used as the vehicle control in all experiments. 3.3 In silico analysis The cBioPortal for Cancer Genomics (http://www.cbioportal.org) is a repository of multidimensional cancer genomics datasets190,191, that was used to determine RKIP-associated signature in lung adenocarcinoma patients. In this study, mRNA expression and protein expression data were analysed from a total of 586 samples from 584 adenocarcinoma patients from the provisional TCGA dataset: Lung Adenocarcinoma (TCGA, Firehose Legacy). According to the TCGA guidelines (http://cancergenome.nih.gov/publications/publicationguidelines), this dataset has no limitations or restrictions. Enrichment analysis were performed to determine the expression profiles of genes that were positively or negatively associated to PEBP1 gene (encoding RKIP) and specific correlations between PEBP1 gene and other genes of interest were determined by spearmen correlation. Significant alterations in mRNA expression (RNA Seq V2 RSEM) and protein expression (RPPA) were determined by Z-score threshold of ±2. 24 Data from the enrichment analysis (mRNA and/or protein data) that were positively or inversely correlated with RKIP expression were retrieved for functional protein association network analysis that was done using STRING (https://string-db.org/), a database that integrates publicly available evidence, computational prediction data and textmining information into comprehensive protein-protein interaction networks192. The end date of the analysis was 30th of September 2020. 3.4 In vitro RKIP genetic modulation For generation of stable RKIP knockout (KO), H292 and HCC827 were previously knocked out using the CRISPR/Cas9 technology, a well-established genome editing tool193. For that, the CRISPR/Cas9 knockout kit from Santa Cruz Biotechnology (CRISPR/Cas9 KO Plasmid - sc-401270-KO-2 and HDR Plasmid – sc-401270-HDR-2) was used193. Cells transfected with both plasmids for RKIP knockout were so called as RKIP KO cells, while control cells transfected only with HDR plasmid were so called as CTR. For generation of stable RKIP overexpressing (OE) cell lines, we used a pcDNA3 vector containing the full cDNA of RKIP in H292 and HCC827. This overexpression is based on a transfection with a pcDNA vector that contains a multiple cloning site in which the full cDNA of RKIP is inserted, and also a Geneticin (G418) resistance gene that will be useful for selection of the successfully transfected cells. The cells were transfected with the empty vector for control (so called as CTR cells), and with RKIP full cDNA containing vector to overexpressed it (so called as RKIP OE cells). The cells were plated into 6-well plates at a density of 5x105 cells per well, allowed to adhere overnight and transfected in serum free Opti-MEM media for 24 hours. The transfection was done using the FUGENE HD reagent (Roche) according to the manufacturer’s protocol, with 2µg of plasmid at a ratio of 6:2 (reagent:plasmid). Then the stable transfectants were selected with varying concentrations of G418 (H292=800 µg/ml; HCC827=300 µg/ml). 3.5 Cell viability assay The sulforhodamine B (SRB) assay was used to assess total protein (expressed as total biomass) overtime or in a specific time point. The cells were seeded into 48-well plates in triplicate at a density of 3x104 (A459 and H292) and 6x104 (HCC827) cells per well and allowed to adhere overnight. In the following day the cells were submitted to different conditions of oxygen availability (normoxia and hypoxia), glucose concentrations (1g/L, 2.5g/L and 4.5g/L glucose) or drug treatment (1.5mM of 2-deoxyglucose or metformin) depending on the assay to be carried out. The day of conditions imposition is considered the 0h time point in the overtime assays. At the 0,24,48 or 72 hours, the cells were fixed with cold 10% 25 trichloroacetic acid for at least 1 hour at 4ºC and stained with Sulforhodamine B (Sigma-Aldrich) for 30 minutes. Next, cells were washed with 1% acetic acid to remove the excess of dye and the protein-bound dye was dissolved in 10mM of Tris-Base solution (pH=10.5) for absorbance measurement at 490nm using the Thermo-Scientific Varioskan Flash SkanIt software (Thermo-Scientific). The results were then calibrated to the starting value (time 0 h, considered as 100% of viability) and expressed as the mean ± SD. The assays were done in triplicate at least three times. 3.6 Extracellular glucose and lactate measurements Glucose consumption and lactate production were determined through extracellular glucose and lactate measurements, respectively. The cells were seeded into 48-well plates at a density of 3x104 (A459 and H292) and 6x104 (HCC827) cells per well and allowed to adhere overnight. In the following day the cells were submitted to different conditions of oxygen availability (normoxia and hypoxia), glucose concentrations (1g/L, 2.5g/L and 4.5g/L glucose) or drug treatment (1.5mM of 2-deoxyglucose or metformin) for 24, 48, or 72 hours depending on the assay to be carried out. Extracellular glucose and lactate content were analysed in cell culture medium at the respective time points using commercial kits (Spinreact), in which 100µL of the commercial reagent is added to 2µL of the supernatant sample. After incubation time, the absorbance was measured at 490nm in the Varioskan Flash reader using the SkanIt software (Thermo-Scientific). Glucose and lactate quantity were calculated through a standard curve and normalized to total biomass determined by SRB assay, as previously described129,159. Results are expressed as mean mg/total biomass. The assays were done in triplicate at least three times. 3.7 Western blot analysis The cells were seeded in 6-well plates at a density of 1x106 cells per well and allowed to adhere overnight. The cells were then incubated for 24 hours in different conditions of oxygen availability (normoxia and hypoxia) and glucose concentrations (1g/L and 4.5g/L glucose) depending on the assay to be carried out. Afterwards, cells were washed and scrapped in cold PBS and lysed in lysis buffer containing phosphatases and proteases inhibitors (Roche). After a centrifugation at 13000 rpm for 15 minutes at 4°C, total protein was quantified using the Bradford reagent (Sigma-Aldrich). Aliquots of 40µg of total protein from each sample were prepared and separated on a standard 12% polyacrylamide gel by sodium dodecyl sulfate polyacrylamide gel electrophoresis (100V) and transferred onto a nitrocellulose membrane (Amersham Biosciences) using the Trans-Blot Turbo Transfer System (Bio-Rad) (25V, 1A for 30 min). Next, the membranes were blocked with 5% milk in Tris-Buffered Saline/0.1% Tween (TBS- 26 Tween) for 1 hour at RT and incubated overnight with the primary antibodies at 4°C (Table 2). After washing in TBS-Tween, the membranes were incubated with the respective secondary antibodies coupled with horseradish peroxidase (1:2500, Cell Signalling) for 1 hour at RT. Tubulin was used as loading control. Blots detection was done by chemiluminescence (Supersignal West Femto kit, Pierce, Thermo Scientific) using the Sapphire Biomolecular Imager (Azure Biosystems). Table 2: Details of the primary antibodies used for western blot. Protein Reference Dilution (Secondary Antibodies) HK2 C64G5 (CS) 1:1000 (Rabbit) PFKP D4B2 (CS) 1:1000 (Rabbit) PKM2 D78A4 (CS) 1:1000 (Rabbit) LDHA C4B5 (CS) 1:1000 (Rabbit) p-GSK3β (Ser9) D85E12 (CS) 1:1000 (Rabbit) PDH C54G1 (CS) 1:1000 (Rabbit) ACO2 D6D9 (CS) 1:1000 (Rabbit) SDHA D6J9M (CS) 1:1000 (Rabbit) DLST D22B1 (CS) 1:1000 (Rabbit) RKIP D42F3 (CS) 1:1000 (Rabbit) α-Tubulin SC-73242 1:5000 (Mouse) CS: Cell Signalling Technology; SC: Santa Cruz Biotechnology; 3.8 Immunofluorescence analysis The cells were seeded on glass cover slips placed into 12-well plates until 60% of confluence and allowed to adhere overnight. Cells were incubated in normoxic or hypoxic conditions for 24 hours. Afterwards, the cells were fixed and permeabilized in cooled methanol for 10 minutes and blocked with 5% bovine serum albumin for 30 minutes. Slides were incubated overnight at RT with the primary antibodies (Table 3). After washing in Phosphate Buffered Saline (PBS), the TRITC (anti-rabbit) and FITC (anti-mouse) Alexa Fluor-conjugated secondary antibodies (Molecular Probes, Invitrogen) were applied at a dilution of 1:500 for one hour at RT protected from the light. Finally, after washing in PBS, the nucleus was stained with 4’,6-diamino-2-phenylindone (DAPI) (Sigma) and slides were covered using Vectashield Mounting Media. Images were obtained with a fluorescence microscope (Olympus BX61) at 200X magnification, using Cell P software. 33 case of low glucose medium the cells consume all the glucose available in the first 24/48 hours, leading us to conclude that those cells are completely dependent on the presence of glucose to maintain an exponential proliferative rate, since in medium with high glucose concentrations the optimal cell growth is not affected (Figure 8). Overall, from this first analysis, A549 cell line present the higher proliferative rates, followed by H292 and finally HCC827 cells, in all culture conditions (Figure 8A). In the opposite, the HCC827 cell line was the faster consuming the culture medium glucose, while H292 did it more slowly, meaning that the culture conditions affect cell proliferation, but their proliferative rates do not determine how they behave metabolically. Figure 8: Effect of the culture conditions in the cell proliferation of NSCLC cell lines. Total biomass (SRB assay) (A) and total extracellular glucose (ug) (B) were measured at 24, 48 and 72 hours in A549, H292 and HCC827 cell lines maintained in normoxia (N) or hypoxia (H), and different glucose medium conditions: 1g/L, 2.5g/L and 4.5g/L glucose. Results are represented as mean ±standard deviation (***p<0.001) (N=3). One of the metabolic hallmarks of cancer cells is their dependency on aerobic glycolysis, which is characterized by a high uptake of glucose and a consequent high lactate production98. Thus, to fully metabolic characterize our cell lines, we next evaluated their glucose consumption and lactate production rates overtime (Figure 9) but excluding the low glucose (1g/L) condition since it affects the proliferation rates. By analysing Figure 9, we can observe that hypoxic conditions significantly promote the glycolytic phenotype, as the cells consume more glucose and produce more lactate than in normoxic conditions, as expected. At normoxic conditions, the cells tend to stabilize the glucose consumption rates upon 24 hours (Figure 9A), while the extracellular lactate tends to decrease (Figure 9B), probably because is being 34 converted into pyruvate to enter in the OXPHOS at the mitochondria. It is also important to note that cells are more glycolytic under intermediate glucose concentrations (2.5g/L glucose), than when are cultured in high glucose conditions (4.5g/L glucose). Figure 9: Effect of the culture conditions in the glucose and lactate metabolism of NSCLC cell lines. Glucose consumption (A) and lactate production (mg/total biomass) (B) was measured at 24, 48 and 72 hours under normoxic (N) or hypoxic (H) conditions, and at different glucose medium conditions (2.5g/L and 4.5g/L glucose). Results were calibrated to the total biomass, assessed by SRB, and are represented as mean ±standard deviation (***p<0.001) (N=3). Comparing the results from the different cell lines, there are no statistically significant differences between them (Figure 9). However, even after normalizing the results to the total biomass, the tendency observed before (Figure 8B) is somehow maintained: HCC827 present higher rates of glucose consumption and lactate production overtime, mainly at normoxia and under 2.5g/L of glucose conditions, while for H292 cell line is the opposite (Figure 9). Moreover, focusing on the 24 hours we can observe that the glucose consumption levels for HCC827 cells are similar between normoxia and hypoxia conditions (Figure 9A), which indicate that these cells are already dependent of glycolysis in normoxic conditions. Whereas for H292 cells we observe an increase in both glucose consumption and lactate production levels under hypoxic conditions (Figure 9), indicating a possible switch in these cells’ metabolism. Overall, it seems to us that the HCC827 cell line is more glycolytic, while H292 cell line 35 seems to be more dependent of OXPHOS. With that in mind and knowing that the metabolic reprogramming of cancer cells is regulated by the expression of several metabolic markers103,113, we complete the characterization of the cell lines by assessing the expression levels of several metabolic proteins and RKIP, our molecule of interest, in the above mention culture conditions (Figure 10 and Figure 11). Upon western blot analysis, it was possible to see that all cell lines express RKIP, although at distinct levels (Figure 10). At normoxic conditions, A549 cells present the highest expression levels, followed by H292 cells, and finally, HCC827 cells that have the lowest levels of RKIP, as our group determined before193. Interestingly, RKIP expression varies with the culture conditions: in A549 cell line RKIP expression decrease under hypoxic conditions, while for HCC827 increases, but in both RKIP tend to decrease under high glucose conditions. In contrast, for H292 cell line the RKIP expression slightly increase both at hypoxic conditions and with high concentrations of glucose (Figure 10). Figure 10: Metabolic proteins and RKIP expression in NSCLC cell lines. (A) Representative western blot for an assay in which cells were maintained under normoxic (N) or hypoxic (H), and low glucose (LG -1g/L) or high glucose (HG – 4.5g/L) conditions for 24 hours. The expression of several metabolic proteins was assessed: HK2, PKM2, LDHA - glycolytic metabolism; PDH and TCA cycleSDHA, DLST and ACO2. The expression of RKIP was assessed and quantification of WB for RKIP (B) was performed using band densitometry analysis with Image J software. Relative protein expression results are shown as the ratio between the proteins and α-Tubulin, and represented as the mean of two independent assays. Concerning the remain proteins assessed, it is clear that the three cell lines have very distinct metabolic expression profiles (Figure 10A and Figure 11). For A549 cell line, HK2 expression was not detectable however higher expression levels of others glycolytic markers (PKM2 and LDHA) were observed when compared to some TCA cycle proteins (SDHA, 36 DLST, and ACO2) (Figure 10A and Figure 11), suggesting that A549 can still rely on glycolysis. Concerning the H292 cell line it was observed a markedly higher expression of PDH, one entry point into the TCA cycle, and other TCA cycle proteins when compared with the remaining cell lines (Figure 10A and Figure 11), indicating a preference for the oxidative metabolism. In contrast, HCC827 cell line present highest levels of HK2 expression, mainly in normoxic conditions, when compared to A549 and H292 cell lines, and also presents the lowest PDH expression levels (Figure 10A and Figure 11), reenforcing our previous observation that of all the cell lines, HCC827 is the less dependent on the oxidative metabolism. Figure 11: Quantification of the western blot presented in Figure 10, relative to metabolic associated proteins. Cells were maintained under normoxic (N) or hypoxic (H), and low glucose (LG -1g/L) or high glucose (HG – 4.5g/L) conditions for 24 hours. The expression of several metabolic proteins was assessed, and quantification of WB was performed using band densitometry analysis with Image J software. Relative protein expression results are shown as the ratio between the proteins and α-Tubulin, and represented as the mean of two independent assays. 37 Furthermore, it can be observed that, as expected, the different culture conditions affect the expression of the metabolic markers, being this effect more pronounced in cells that were under hypoxia. An increased expression HK2, PKM2 and LDHA was observed in cells subject to hypoxic conditions (Figure 10A and Figure 11), which is in accordance with the well-described induction of glycolytic enzymes through stabilization of HIF-1α111,112. For instance, H292 cell line present increased expression of HK2 and LDHA, and a decreased expression of PDH at hypoxic conditions (Figure 10A and Figure 11), suggesting a metabolic adaptation towards a more glycolytic phenotype in the absence of oxygen. Moreover, the variations on RKIP expression under different metabolic conditions, and even the difference on RKIP expression level between the above mention cell lines, emphasizes our hypothesis that RKIP expression may play a role on cancer cells metabolic reprogramming. 4.3 Effect of RKIP expression in the modulation of cancer cells metabolism 4.3.1 RKIP role on the metabolic behaviour of lung cancer cells To pursue our hypothesis that RKIP could be a master regulator of lung cancer cells metabolism, we moved for some functional in vitro assays. For that, we selected two cell lines with apparently distinct metabolic phenotypes, H292 (oxidative) and HCC827 (glycolytic), and then genetically modulate RKIP expression using two different approaches: gene knockout (KO) and overexpression (OE). Taking into account our first characterizations of the WT cell lines, all the following assays were done using culture medium with 2.5 g/L of glucose, both in normoxia and hypoxia conditions. By western blot, as it can be observed in Figure 12A, we confirmed that the transfections were successful in both cell lines. Additionally, following the same tendency of the WT experiments (Figure 10A and Figure 11), in HCC827 cell line it was observed an increase on RKIP expression in hypoxic conditions, more evident in overexpressing (OE) cell lines. In contrast, some discrepant results were found for H292 cell line between the two transfection controls, however the differences are not significant (Figure 12A). Before we proceed to the evaluation of RKIP role on the metabolic behaviour of lung cancer cells, we first aimed to exclude the possibility of RKIP modulation be affecting cellular proliferation and compromising the following results. For that, we determined the proliferation rates of the transfected cells in the above cited culture conditions, but no significant differences were found between the clones in none of the cell lines (Figure 12B). It is important to note that the cells were not affected by the transfection process, and the distinct proliferation rates between the cell lines were even maintained (Figure 8 and 12B). 38 Figure 12: Effect of RKIP expression manipulation on cell proliferation. (A) Representative western blot analysis of RKIP expression for assessment of the transfection’s efficiency for H292 (KO and OE) and HCC827 (KO and OE) cell lines. Quantification of WB for RKIP expression was performed using band densitometry analysis with Image J software. Relative protein expression results are shown as the ratio between the proteins and α-Tubulin, and represented as the mean of the 2 independent assays. (B) Total biomass was measured at 24, 48 and 72 hours by SRB assay for H292 (KO and OE) and HCC827 (KO and OE) cells maintained in normoxia (N) or hypoxia (H) in 2.5g/L glucose concentration medium. Results are represented as mean ±standard deviation (N=3). Next, to evaluate the RKIP effect on the glycolytic metabolism of these cells, glucose consumption and lactate production rates were measured over time, under normoxia and hypoxia conditions, for the above mention transfected cell lines (Figure 13). By analysing Figure 13, we observed once again that in hypoxic conditions there is an induction of the glycolysis, which is translated in increased glucose consumption and lactate production, as expected. Additionally, it is possible to notice that in normoxic conditions there is a tendency to stabilization of both glucose consumption and lactate production over 39 the time, what may indicate that cells can be using alternative pathways to support their growth and proliferation, as observed for the WT cell lines (Figure 9). Figure 13: Effect of RKIP on cells glycolytic metabolism. Glucose consumption and lactate production was measured overtime in H292 (KO and OE) and HCC827 (KO and OE) cells maintained in normoxia (N) or hypoxia (H) in 2.5g/L glucose concentration medium. Results were calibrated for the total biomass, measured by SRB assay, and represented as mean ± standard deviation (N=3). 40 Unfortunately, and in contrast to what we were expecting, comparing the different manipulated cell lines for RKIP expression (KO and OE) with the respective controls, we did not observe significant differences between them overtime (Figure 13). In fact, some differences were observed for shorter time points (24 hours), but along time they tend to recover and equalize the levels of glucose consumption and lactate production. Since the effect of RKIP expression in the modulation of the glycolytic metabolism is almost null at basal conditions and overtime, we next decided to do the assays only for 24 hours but in the presence of some glycolytic modulators to see if in the presence of a “metabolic stressor” the RKIP manipulated cells respond differentially. Thus, we assessed glucose consumption and lactate production rates upon treatment with 2-deoxyglucose (glycolysis inhibitor) and metformin (OXPHOS inhibitor). By analysing Figure 14 we can perceive that overall, the treatment with 2-DG, an HK2 inhibitor, leads to an inhibition of the glycolytic metabolism, whereas treatment with metformin, an inhibitor of mitochondrial ETC complex I, promotes it, as expected128,165. It is interesting to note that the influence of metformin in HCC827 cells has the same effect as hypoxia, while in H292 cells the absence of oxygen has as a stronger effect in glycolysis induction than metformin. Also, 2-DG has a stronger glycolytic inhibitory effect in HCC827 cells than in H292 cells, which can indicate that HCC827 cells rely more on glycolysis, as previously observed. Concerning RKIP effect, again there are no statistically significant differences between the CTR and the RKIP manipulated cells (Figure 14). However, in general, there is a tendency for RKIP KO cells export more lactate in all conditions for both cell lines, even not being this tendency observed in the glucose consumption rates. Also, it is interesting to note that at least for HCC827 cell line, the opposite effect is observed, with OE cells exporting less lactate than the respective control, for all conditions (Figure 14). In the future will be important to deepen the study with different approaches, such mitochondrial activity assays, and shorter time points since the cells seems to have a high metabolic plasticity in vitro over time. 41 Figure 14: RKIP effect on cells glycolytic metabolism upon treatment with metabolic inhibitors. Glucose consumption and lactate production was measured overtime in H292 (KO and OE) and HCC827 (KO and OE) after treatment with 1.5mM 2-Deoxyglucose (2-DG) or metformin (Met) for 24 hours in 2.5g/L glucose concentration medium. Results are represented as mean ±standard deviation (N=3). 42 4.3.2 RKIP role in the modulation of metabolic proteins expression We previously perceived from the WT cell lines characterization that metabolic differences at protein levels do not necessarily translate into differences in metabolic function, at least for the assays used in this study. Also, the in silico analysis showed us that patients with RKIP altered expression have alterations in the expression of metabolic-associated genes, at mRNA and protein level. Thus, in the next step we intend to assess the protein expression levels of some metabolic regulators by western blot in this genetically manipulated cell lines for RKIP expression. In general, as it can be observed in the western blot, concordantly with what was observed for the WT cell lines, there is an upregulation of HK2 and LDHA, and downregulation of PDH upon hypoxic conditions (Figure 15 and Figure 16). Figure 15: Effect of RKIP on the modulation of metabolic proteins expression. Representative western blot analysis for H292 (A) and HCC827 (B) cell lines under normoxia (N) and hypoxia (H), upon RKIP genetic manipulation. The expression of several metabolic proteins was assessed: HK2, PFKP, PKM2, LDHA, and p-GSK3β - glycolytic metabolism; PDH and TCA cycleACO2 and SDHA. Tubulin was used as loading control (N=2). Regarding the effect of RKIP expression in metabolic-related proteins, we observed differences in some of these proteins’ expression upon RKIP gain or loss of expression, for both cell lines (Figure 15 and Figure 16). 49 Figure 20: In silico correlation between RKIP and HIF-1α in adenocarcinoma patients. (A) RNA Seq V2 data for AC patients (584 patients), showing the RKIP ( PEBP1 ) and HIF-1α ( HIF1A ) mRNA expression is inversely correlated. (B) Data from enrichment analysis in the same AC patients, showing HIF-1α is upregulated in the set of samples with RKIP mRNA downregulation. All data belongs to the Lung Adenocarcinoma (TCGA, Firehose Legacy) provisional database and is available at www.cbioportal.org. Moreover, RKIP showed to modulate the expression of several metabolic proteins (Figure 15, Figure 16 and Figure 17). Therefore, to validate the in vitro results, we explored the specific correlation between the metabolic proteins addressed in this study with RKIP mRNA expression in lung adenocarcinoma patients. Analysing Table 5, we can perceive that RKIP is inversely correlated with glycolysis associated genes, which is in accordance with the downregulation of some glycolytic proteins and consequent inhibition of glycolysis observed in RKIP OE in H292 cell line (Figure 19). Regarding TCA cycle associated genes, we observed a positive correlation of these genes with RKIP mRNA expression. However, at the protein level in vitro , RKIP showed to downregulate some of these proteins such as PDH and SDHA (Figure 15 and Figure 16). Additionally, to better understand the role of RKIP in the mitochondrial metabolism, namely in OXPHOS, we also determine the RKIP correlation with a set of electron transport chain (ETC) associated genes and found a significantly positive correlation with all these genes (Table 5). 50 Table 5: In silico correlation between RKIP and metabolism associated proteins in lung adenocarcinoma patients. Spearman’s correlations of mRNA data (RNA Seq V2 RSEM) for AC patients (584 patients), showing that RKIP is inversely correlated with glycolysis associated genes, and positive correlated with TCA cycle and ETC associated genes. All data belongs to the Lung Adenocarcinoma (TCGA, Firehose Legacy) provisional database and is available at www.cbioportal.org. Protein Cytoband Spearman's Correlation p -value Associated metabolic process GLUT1 1p34.2 -0.424 5.57e-24 Glycolysis HK2 2p12 -0.329 1.58e-14 Glycolysis PFKP 10p15.2 -0.265 9.89e-10 Glycolysis PKM 15q23 -0.173 7.729e-5 Glycolysis LDHA 11p15.1 -0.203 3.413e-6 Glycolysis MCT1 1p13.2 -0.277 1.46e-10 Glycolysis MCT4 17q25.3 -0.452 2.17e-27 Glycolysis PDHA1 Xp22.12 0.332 9.13e-15 TCA cycle PDHB 3p14.3 0.355 7.91e-17 TCA cycle IDH2 15q26.1 0.264 1.01e-9 TCA cycle ACO2 22q13.2 0.0945 0.0317 TCA cycle SDHA 5p15.33 0.0692 0.116 TCA cycle ATP5MC2 12q13.13 0.397 5.73e-21 ETC UQCRC1 3p21.31 0.232 9.23e-8 ETC UQCR10 22q12.2 0.248 1.09e-8 ETC ATP5MG 11q23.3 0.294 9.43e-12 ETC NDUFS3 11p11.2 0.315 2.41e-13 ETC COX17 3q13.33 0.256 3.50e-9 ETC ATP5PF 21q21.3 0.238 4.05e-8 ETC NDUFB4 3q13.33 0.170 1.003e-4 ETC NDUFB6 9p21.1 0.122 5.290e-3 ETC COX7C 5q14.3 0.379 4.47e-19 ETC ATP5MC1 17q21.32 0.273 2.82e-10 ETC ATP6V0B 1p34.1 0.223 3.07e-7 ETC COX7A2 6q14.1 0.258 2.78e-9 ETC NDUFA8 9q33.2 0.224 2.56e-7 ETC Altogether, this in silico analysis indicate us that RKIP is inversely correlated with the glycolytic metabolism in lung adenocarcinoma patients, validating some of our results. Along with that, we observed a positive correlation between RKIP and TCA cycle or ETC associated genes, which once again suggests that RKIP can play a role in distinct metabolic processes. To conclude, these results provide us a strong evidence for a role of RKIP in lung cancer metabolism and will be very useful in the future to deepen our work CHAPTER 5: General Discussion 52 5. General discussion Altered glucose metabolism is a widespread characteristic of several solid tumours and is often associated with tumour aggressiveness, metastasis, poor prognosis as well as therapy resistance103,131,157,194,195. While most cancers undergo metabolic rewiring, mounting evidence suggests that tumour metabolism signatures are context dependent, being influenced by a wide range of factors including, tissue of origin, tumour grade, microenvironment and oncogenic signalling137,153,154,196. As example, given that lung cancers exhibit a high mutation burden13,21,22, it is possible that their genetic heterogeneity is also reflected at the metabolic level, particularly since oncogene activation and loss of tumour suppressors can alter cellular metabolism109,154,197. Lung cancer cells often harbour mutations in genes and pathways, such as the PI3K/AKT pathway, the oncogene MYC , and the tumour suppressor genes TP53 and STK11 13,176,179. Specifically, in NSCLC tumours, known to be driven by oncogenic activation of tyrosine kinases, such as mutations in KRAS and EGFR genes, that were already described as modulators of metabolic rewiring in tumour cells154,179,180,197,198. These cell signalling pathways are implicated in cell metabolism by securely regulating the capacity of cells to obtain access to nutrients and subsequently process these compounds179. Moreover, oncogenic signalling supports the cells ability to maintain a growth factor–independent glycolysis and survival through the expression of oncogenic kinases95,107,108. Overall, the literature provides strong evidence of metabolic rewiring in lung cancer. Thus, nowadays, one of the emerging themes of cancer research rely on the acquired heterogeneous metabolic phenotype by tumours cells to withstand the complex challenges linked to cancer progression137,176, being puzzling to search for new oncogenic alterations behind this metabolic heterogeneity. RKIP protein arises to us as a potential modulator of cancer cells metabolism. Given the key role of RKIP in the regulation of several intracellular signalling pathways that have been described to be crucial to cancer cells metabolic rewiring, more than MAPK, and its involvement in many malignancies52, including NSCLC62, in this study we hypothesized that RKIP could be a key player in the metabolic reprogramming of tumours cells. Supporting our hypothesis, Rosner and Lee recently demonstrated the importance of the pro-metastatic transcription factor BACH1, that is negatively regulated by RKIP, in promoting aerobic glycolysis in breast cancer188,189. In addition, Kapoor et al . recently reported for the first time a link between RKIP downregulation and genes that directly regulate cell metabolism, namely fatty acid degradation and pyruvate metabolism, by analysing TCGA data for ccRCC patients187. 53 Keeping this in mind, our first approach was to perform an extensive in silico analysis to determine the RKIP-associated signature in lung adenocarcinoma patients, using the TCGA database. It was very interesting to verify that at the mRNA level, RKIP expression showed to be positively correlated with mitochondrial genes implicated in pyruvate and energy metabolism, electron transport chain and oxidative phosphorylation, and inversely correlated with genes related to glucose metabolism, indicating a potential negative association between RKIP and Warburg effect. At the protein level we found that RKIP expression is significantly correlated with genes involved in cancer signalling transduction and related with different biological processes, which is in accordance with the well-described role of RKIP as an important modulator of signalling molecules in cancer34,35,62. Importantly, RKIP showed to be correlated with players of MAPK signalling, which was expected considering RKIP role as an endogenous inhibitor of MAPK, as well as PI3K/AKT, AMPK and HIF1-α signalling, all described to play a crucial role in tumours metabolic reprogramming110,112,123,147,199,200. It is important to notice that, despite both enrichment analyses showed an association between RKIP and metabolic processes, no common genes were pointed. Probably because the protein data in TCGA datasets is still limited when compared to mRNA data, being protein expression data for RKIP and several metabolic proteins not available, or just because they are not included in the Top25 most related proteins. Interestingly, as cited above188,189, Lee et al. have reported also using TCGA database, that BACH1 gene expression is inversely correlated with ETC gene expression and OXPHOS mainly in patients with breast cancer, but also in other tumour types, including lung cancer188. In that study, the results were biologically validated at protein level, suggesting that mitochondrial metabolism can be exploited by targeting BACH1 to sensitize breast cancer cells to mitochondrial inhibitors188. In the same year, and following Lee et al. work188, other group reported the importance of BACH1 in driving lung tumours metastasis, suggesting that Heme oxygenase 1 (HO-1) inhibitors represent an effective therapeutic strategy to prevent lung cancer metastasis201. Altogether, given that BACH1 is a downstream target of RKIP, as well as a negative regulator of RKIP expression68,76,189, we are speculating a positive association between RKIP and ETC genes expression or OXPHOS induction. However, an association between BACH1 function and RKIP was never explored in lung malignancies. Moreover, besides all the above speculation, the direct and functional role of RKIP protein on cancer cells metabolism is completely unknown, being ours, the first study exploiting its role in the field. Thus, we moved to a more practical approach, starting by doing a metabolic characterization of the NSCLC cell line models to be used in the study. Given the fact that both nutrient and oxygen conditions, as well as the genomic context, play a crucial role in determining cell metabolic 54 behaviour103,176,196, three cell lines with different genetic backgrounds were selected and characterized in different culture conditions. First, we determined the effects of different glucose concentrations and oxygen conditions in the proliferation rates of the cells and observed that cells under stress conditions (hypoxia and low glucose concentration: 1g/L glucose) significantly proliferate less. This can be explained by a possible increased dependency of these cells on glucose, since it is described that cancer cells under hypoxia enhanced glycolysis to sustain the bioenergetic and biosynthetic requirements associated to cancer cells high-proliferative rates112. Also, the cells proliferative rates do not determine their metabolic behaviour, which is consistent with a previous study in NSCLC cell lines176. Finally, we observed that hypoxic conditions significantly promote glycolysis, which was translated in increased rates of glucose consumption and lactate production/export, as expected based on the well-documented induction of glycolysis in hypoxia111,112. Whereas, for normoxic conditions, a stabilization in glucose consumption and a tendency to a decrease in extracellular lactate content overtime was observed in some cases. We believe that this might be due to the metabolic flexibility of cancer cells that allows them to adapt to changes in the nutrient availability, therefore as the levels of extracellular glucose drop and extracellular lactate increases, cells can use lactate as fuel for the TCA cycle and the oxidative phosphorylation. Indeed, it has been described for different in vitro models, including A549 cell line, that exposure to lactic acid can reverse cancer cells phenotype, causing a shift from Warburg effect to OXPHOS202,203. Supporting this, some in vivo studies reported that lactate is metabolized in human lung tumours, and more importantly that glucose contribution to the TCA cycle occurs mostly indirectly, through circulating lactate174,175. A recent study demonstrated that NSCLC cells are remarkably diverse in the rates of glucose uptake and utilization, as well as in the pathways by which the glucose is metabolized176. Herein, we found that HCC827 cells presented the most glycolytic phenotype, followed by A549, and lastly by H292, a result that was further confirmed by the differences on the metabolic proteins expression profile by western blot. We noticed that H292 was the one that presents higher levels of TCA cycle enzymes, in particularly PDH, the enzyme responsible for the flux of glucose to TCA cycle140,141, whereas HCC827 present the lowest PDH expression of all the cell lines, suggesting that HCC827 cell line is more glycolytic and H292 more oxidative. In addition, it was possible to observe that the different conditions affect the expression of several metabolic markers as expected, being the effects more pronounced in cells that were subject to hypoxia, which was accordance with the well-described effect of HIF-1α stabilization in the induction of several metabolic genes111,112,142,147. Looking to what is described in the literature, some studies have reported that A549 cells rely mainly on OXPHOS rather than in glycolysis in normoxia128,204. Moreover, Chen et al. recently demonstrated that HCC827 cells present a higher ratio of lactate secretion/glucose consumption 55 than A549, meaning that HCC827 cell line has a preference for aerobic glycolysis176, supporting our believe that HCC827 cells are the most glycolytic. There are no studies on H292 cell line, hampering comparisons. Furthermore, it was very interesting to observe that the different glucose and oxygen culture conditions affected RKIP expression levels. Several studies have focused on the mechanisms by which hypoxia and activation of HIF-1-dependent signalling promotes metastasis through regulation of metabolic reprogramming, invasion, angiogenesis, immune suppression, epithelial-mesenchymal plasticity and resistance to apoptosis111,112,147,150. Given the role of RKIP as a metastasis suppressor, the possibility of RKIP also playing a role in the modulation of the cellular plasticity response to tumour hypoxic stress is not surprising. It was already demonstrated, in H1299 NSCLC cell line, that hypoxia-induced Notch1 activation was significantly inhibited by RKIP overexpression and stimulated by RKIP knockdown, supporting the notion that hypoxia-induced EMT can be functionally linked to RKIP expression, which subsequently modulates Notch signalling during metastasis49. Additionally, a recent computational study of protein-protein docking between HIF-1α and RKIP, suggested that RKIP can act as a negative regulator of HIF-1α, and in fact, cobalt chloride treatment (mimics hypoxic conditions) resulted in RKIP dissociation from HIF-1α in prostate cancer cell lines205. Thus, despite the inconsistent results we found between our in vitro models, which deserve further exploitation, the variations on RKIP expression under different metabolic conditions suggest that RKIP expression alterations could be a possible mechanism behind cells metabolic reprogramming. Following, to pursue the main question of this work, which is to functionally determine the role of RKIP on the modulation of cancer cells metabolism, we selected the two lung cancer cell lines with distinct metabolic and molecular phenotypes, HCC827 (glycolytic – EGFR mutant) and H292 (oxidative – WT for EGFR ), and then genetically modulate RKIP expression to stable knockout (KO) and overexpress (OE) its gene. Firstly, and as a transfection control, we determined whether RKIP expression manipulation affects the proliferation of lung cancer cell lines. For both the cell lines, no significant differences in the proliferation rates were observed, which lead us to conclude that neither RKIP expression nor the transfection process affected these cell lines proliferation capacity, and thus will not be a confounding indicator in the further studies. However, as a well-described tumour suppressor, RKIP acts as a multifunctional protein in carcinogenesis regulating several processes such as cellular growth, proliferation, migration, invasion and metastization36,43,52,63, thus we were expecting to see proliferation differences after RKIP KO or OE. Concerning NSCLC, no clear exploration of RKIP’s role in such processes 56 was done yet, being only demonstrated that NSCLC metastasis are inversely correlated with RKIP expression levels46. However, previous unpublished data from our group, also demonstrated no significant differences in cellular viability and migration rates between cells with and without RKIP in H292 cell line, although changes were seen in the expression of EMT proteins, which confirmed the presence of a RKIP phenotype193. The biological role of RKIP in lung cancer is yet being exploited in our group, but there were other tumour types for which an absence of RKIP impact in cells proliferation was found in vitro 58. Moving on to the dissection of RKIP role in metabolism modulation, we started by comparing the overtime glucose consumption and lactate export rates in the RKIP manipulated cell lines (H292 and HCC827), both in normoxia and hypoxia conditions. Overall, the metabolic behaviour of the two cell lines followed the same tendency observed with parental cells, meaning that they were not altered with transfection. When comparing the RKIP KO/OE cells with the respective controls, in both cell lines, no statistically significant differences were observed overtime. However, for shorter time points (24 hours) some differences were observed, but along time they tend to recover and equalize the levels of glucose consumption and lactate production. To circumvent this, we next did the same analysis but only for the shorter time point (24 hours), and upon treatment with metabolic inhibitors, to see whether RKIP manipulated cells would respond differently under stress conditions. From this analysis we observed that 2-deoxyglucose, a glycolysis inhibitor158, lead to a significant decrease of the glycolytic rates, while treatment with metformin, a OXPHOS inhibitor, result in glycolysis induction, as expected206,207. Moreover, it was also evident that the two cell lines respond differently to these inhibitors. Treatment with metformin, lead to a significant induction of glycolysis on H292 cells, suggesting that was able to inhibit the oxidative metabolism and cause a shift to glycolytic metabolism on these cells. For HCC827 cells, the metformin effect was not so evident, which can be explained by a supposed lack of oxidative metabolism dependence on these cells. Consistently, the opposite was observed after 2-deoxyglucose treatment, where glycolysis inhibition was more evident for HCC827 cells, indicating a higher dependency of these cells on glycolysis. Together these results corroborate the tendency observed in the metabolic characterization of the cells before transfection, where H292 cells seemed to rely more in the oxidative metabolism, whereas HCC827 is more dependent on glycolytic metabolism. Concerning the hypothetic role of RKIP expression on metabolism modulation, once again no statistically significant differences between the CTR and the RKIP manipulated cells were found, even when subjected to metabolic modulators. However, it was observed a tendency for RKIP KO cells export more lactate in both conditions and cell lines, which is in accordance with our hypothesis of glycolytic 57 phenotype induction upon RKIP loss. Interestingly, the opposite tendency was observed for RKIP overexpression in HCC827 cells. This tendency was not followed by the glucose consumption rates, to whom no differences were observed. Upon these results a few limitations warrant mention, such as the conditions where these assays were performed, and the limited sensitivity of the assays we used. Here we have to take in consideration that we are measuring extracellular levels of metabolites that can be constantly interconverted and used to fuel other metabolic pathways which can be a possible explanation for lactate levels not following the same tendency of glucose levels. For instance, besides the use of lactate to fuel TCA cycle, lactate has also been shown to serve as a gluconeogenic source for glucose generation in lung cancer cells grown with low glucose in vitro 208. Moreover, the presence of other nutrients in the culture medium can also influence the results. For instance, many reports point towards the essentiality of glutamine for NSCLC lines in vitro 151,186,209, thus we cannot discard the possibility of glutamine being interfering with our results. Intriguingly, that lack of significant metabolic differences, in the RKIP manipulated cell lines, was contradictory with the differences found on metabolic proteins expression, evaluated by western blot. For H292 cell line, some interesting things were spotted. Upon RKIP overexpression, we observed a decreased expression of some glycolytic markers, such as HK2 and LDHA, and also by immunofluorescence analysis a slight decreased expression of MCT1, MCT4, CD147, and GLUT1 was observed. This was consistent with the negative correlation that was found between RKIP and the genes that encode MCT4 and GLUT1 in the in silico analysis. In accordance, an increase in some glycolytic markers’ expression, HK2, PKM2 and LDHA was observed upon RKIP loss. Although at the first sight these results seem to be in accordance with our initial hypothesis of RKIP negative regulation of Warburg effect, when we look to TCA cycle enzymes expression, we also observed a decrease in some of these proteins expression in RKIP overexpressing cells. Indeed, the reduction on PDH expression, the enzyme that control one of the TCA entries points, was very evident in both western blot and immunofluorescence analysis. These results suggest that RKIP overexpression leads to inhibition of both glycolysis and OXPHOS in this cell line. Although being the Warburg effect the most well-described phenotype of cancer cells, most tumours acquire the ability to undergo enhanced aerobic glycolysis while maintaining the mitochondrial metabolism, as has been reported for lung tumours137,151. Concerning HCC827 cell line, the results were not so clear, yet some differences were spotted. We observed an upregulation of the glycolytic enzyme PFKP and a downregulation of the TCA cycle enzymes, PDH and SDHA, which can indicate on one hand, an enhancement of glycolysis and on the other hand a decrease in OXPHOS, therefore a Warburg-like phenotype upon RKIP loss, as we initially hypothesized. 58 Looking at RKIP OE results, the differences were less pronounced and seem to be dependent on oxygen conditions, which can be due to the fact that in this cell line RKIP expression showed to be affect by hypoxia. We were still able to see a downregulation of PKM2 enzyme for both conditions, indicating a reduction in glycolysis, which is consistent with what was observed in H292 RKIP OE. Curiously, a slight downregulation of PDH in hypoxic conditions was also observed in RKIP OE cells, thus presenting the same tendency that RKIP KO cells. Moreover, in contrast to H292 cell line, we did not find significant differences in glycolytic markers by immunofluorescence analysis upon RKIP overexpression. Regarding the existent literature for some of these metabolic proteins expression in lung cancer, it has been described that NSCLC tumours exhibited enhanced PDH activity when compared to benign lung samples137,151 and that PDH activity is associated with EMT and drug resistance in A549 and HCC827 cells210. In addition, analysis of tumour samples revealed that low expression of PDK4, one enzyme that inhibits PDH activity, as a predictor of poor prognosis in lung cancer210. Curiously, looking at Kapoor et al. results they also reported that PDK2 downregulation was associated to RKIP loss in ccRCC187. Therefore, the downregulation of PDH could be associated to an upregulation of PDK upon RKIP overexpression. Finally, faced with some unexpected results, we further performed an in silico analysis to understand and validate some of our results, using the TCGA database. Firstly, we confirmed that RKIP expression is inversely correlated with HIF-1α expression, being the later enriched in patients with RKIP downregulated, which is in line with the first in silico analysis that showed us an enrichment on HIF signalling associated genes upon RKIP downregulation. Unfortunately, we were not able to detect HIF-1α expression, due to the actual lack of a good antibody in the Lab, herein in the future it would be interesting to study the effect of RKIP expression on HIF-1α expression, and vice-versa, in the models used in this study. Additionally, we found that RKIP mRNA expression was inversely correlated with glycolysis-related genes and positively correlated OXPHOS-related genes in a significant way, which is not consistent with the negative correlation we found in vitro between RKIP and some TCA cycle proteins, mainly in H292 cell line. At this point, it is important to note that the differences and ambiguities found between RKIP’s effect in the two cell lines, and apparent lower impact of RKIP in HCC827 cell line metabolic modulation, could be somehow explained due to its metabolic heterogeneity and plasticity, that naturally occurs inside the tumours137, but which is even more pronounced and less controlled in “artificial” in vitro models with immortalized cell lines. Despite the limitations of that kind of studies, the differences could be also somehow due to their different genetic backgrounds. HCC827 is EGFR mutated, and it was already demonstrated that EGFR mutations can lead to metabolic rewiring towards aerobic glycolysis, due to the 65 15 Ahrendt, S. A. et al. Cigarette smoking is strongly associated with mutation of the K-ras gene in patients with primary adenocarcinoma of the lung. Cancer 92, 1525-1530, doi:10.1002/10970142(20010915)92:6<1525::aid-cncr1478>3.0.co;2-h (2001). 16 Ciuffreda, L. et al. Signaling intermediates (MAPK and PI3K) as therapeutic targets in NSCLC. Current pharmaceutical design 20, 3944-3957, doi:10.2174/13816128113196660763 (2014). 17 Lemmon, M. A. & Schlessinger, J. Cell signaling by receptor tyrosine kinases. Cell 141, 11171134, doi:10.1016/j.cell.2010.06.011 (2010). 18 Hirsch, F. R. et al. Lung cancer: current therapies and new targeted treatments. Lancet (London, England) 389, 299-311, doi:10.1016/s0140-6736(16)30958-8 (2017). 19 Kris, M. G., Hellmann, M. D. & Chaft, J. E. Chemotherapy for lung cancers: here to stay. American Society of Clinical Oncology educational book. American Society of Clinical Oncology. Annual Meeting , e375-380, doi:10.14694/EdBook_AM.2014.34.e375 (2014). 20 Politi, K. & Herbst, R. S. Lung cancer in the era of precision medicine. Clinical cancer research : an official journal of the American Association for Cancer Research 21, 2213-2220, doi:10.1158/1078-0432.Ccr-14-2748 (2015). 21 Network, C. G. A. R. Comprehensive genomic characterization of squamous cell lung cancers. Nature 489, 519-525, doi:10.1038/nature11404 (2012). 22 Network, C. G. A. R. Comprehensive molecular profiling of lung adenocarcinoma. Nature 511, 543550, doi:10.1038/nature13385 (2014). 23 Travis, W. D. et al. The 2015 World Health Organization Classification of Lung Tumors: Impact of Genetic, Clinical and Radiologic Advances Since the 2004 Classification. Journal of thoracic oncology 10, 1243-1260, doi:10.1097/jto.0000000000000630 (2015). 24 Zhou, C. et al. Erlotinib versus chemotherapy as first-line treatment for patients with advanced EGFR mutation-positive non-small-cell lung cancer (OPTIMAL, CTONG-0802): a multicentre, openlabel, randomised, phase 3 study. The Lancet. Oncology 12, 735-742, doi:10.1016/s14702045(11)70184-x (2011). 25 Mitsudomi, T. et al. Gefitinib versus cisplatin plus docetaxel in patients with non-small-cell lung cancer harbouring mutations of the epidermal growth factor receptor (WJTOG3405): an open label, randomised phase 3 trial. The Lancet. Oncology 11, 121-128, doi:10.1016/s14702045(09)70364-x (2010). 66 26 Yuan, M., Huang, L. L., Chen, J. H., Wu, J. & Xu, Q. The emerging treatment landscape of targeted therapy in non-small-cell lung cancer. Signal transduction and targeted therapy 4, 61, doi:10.1038/s41392-019-0099-9 (2019). 27 Antonia, S. J. et al. Nivolumab alone and nivolumab plus ipilimumab in recurrent small-cell lung cancer (CheckMate 032): a multicentre, open-label, phase 1/2 trial. The Lancet. Oncology 17, 883-895, doi:10.1016/s1470-2045(16)30098-5 (2016). 28 Herbst, R. S. et al. Pembrolizumab versus docetaxel for previously treated, PD-L1-positive, advanced non-small-cell lung cancer (KEYNOTE-010): a randomised controlled trial. Lancet (London, England) 387, 1540-1550, doi:10.1016/s0140-6736(15)01281-7 (2016). 29 Rittmeyer, A. et al. Atezolizumab versus docetaxel in patients with previously treated non-small-cell lung cancer (OAK): a phase 3, open-label, multicentre randomised controlled trial. Lancet (London, England) 389, 255-265, doi:10.1016/s0140-6736(16)32517-x (2017). 30 Al-Mulla, F., Bitar, M. S., Taqi, Z. & Yeung, K. C. RKIP: much more than Raf kinase inhibitory protein. Journal of cellular physiology 228, 1688-1702, doi:10.1002/jcp.24335 (2013). 31 Keller, E. T., Fu, Z. & Brennan, M. The role of Raf kinase inhibitor protein (RKIP) in health and disease. Biochemical pharmacology 68, 1049-1053, doi:10.1016/j.bcp.2004.04.024 (2004). 32 Klysik, J., Theroux, S. J., Sedivy, J. M., Moffit, J. S. & Boekelheide, K. Signaling crossroads: the function of Raf kinase inhibitory protein in cancer, the central nervous system and reproduction. Cellular signalling 20, 1-9, doi:10.1016/j.cellsig.2007.07.003 (2008). 33 Gabriela-Freitas, M., Pinheiro, J., Raquel-Cunha, A., Cardoso-Carneiro, D. & Martinho, O. RKIP as an Inflammatory and Immune System Modulator: Implications in Cancer. Biomolecules 9, doi:10.3390/biom9120769 (2019). 34 Farooqi, A. A., Li, Y. & Sarkar, F. H. The biological complexity of RKIP signaling in human cancers. Experimental & molecular medicine 47, e185, doi:10.1038/emm.2015.70 (2015). 35 Lorenz, K., Schmid, E. & Deiss, K. RKIP: a governor of intracellular signaling. Critical reviews in oncogenesis 19, 489-496, doi:10.1615/critrevoncog.2014011923 (2014). 36 Yesilkanal, A. E. & Rosner, M. R. Targeting Raf Kinase Inhibitory Protein Regulation and Function. Cancers 10, doi:10.3390/cancers10090306 (2018). 37 Yeung, K. et al. Suppression of Raf-1 kinase activity and MAP kinase signalling by RKIP. Nature 401, 173-177, doi:10.1038/43686 (1999). 67 38 Vandamme, D., Herrero, A., Al-Mulla, F. & Kolch, W. Regulation of the MAPK pathway by raf kinase inhibitory protein. Critical reviews in oncogenesis 19, 405-415, doi:10.1615/critrevoncog.2014011922 (2014). 39 Yeung, K. et al. Mechanism of suppression of the Raf/MEK/extracellular signal-regulated kinase pathway by the raf kinase inhibitor protein. Molecular and cellular biology 20, 3079-3085, doi:10.1128/mcb.20.9.3079-3085.2000 (2000). 40 Granovsky, A. E. et al. Raf kinase inhibitory protein function is regulated via a flexible pocket and novel phosphorylation-dependent mechanism. Molecular and cellular biology 29, 1306-1320, doi:10.1128/mcb.01271-08 (2009). 41 Lorenz, K., Lohse, M. J. & Quitterer, U. Protein kinase C switches the Raf kinase inhibitor from Raf1 to GRK-2. Nature 426, 574-579, doi:10.1038/nature02158 (2003). 42 Corbit, K. C. et al. Activation of Raf-1 signaling by protein kinase C through a mechanism involving Raf kinase inhibitory protein. The Journal of biological chemistry 278, 13061-13068, doi:10.1074/jbc.M210015200 (2003). 43 Zaravinos, A., Bonavida, B., Chatzaki, E. & Baritaki, S. RKIP: A Key Regulator in Tumor Metastasis Initiation and Resistance to Apoptosis: Therapeutic Targeting and Impact. Cancers 10, doi:10.3390/cancers10090287 (2018). 44 Yeung, K. C. et al. Raf kinase inhibitor protein interacts with NF-kappaB-inducing kinase and TAK1 and inhibits NF-kappaB activation. Molecular and cellular biology 21, 7207-7217, doi:10.1128/mcb.21.21.7207-7217.2001 (2001). 45 Tang, H. et al. RKIP inhibits NF-kappaB in cancer cells by regulating upstream signaling components of the IkappaB kinase complex. FEBS letters 584, 662-668, doi:10.1016/j.febslet.2009.12.051 (2010). 46 Wang, A. et al. Reduced RKIP expression levels are associated with frequent non-small cell lung cancer metastasis and STAT3 phosphorylation and activation. Oncology letters 13, 3039-3045, doi:10.3892/ol.2017.5846 (2017). 47 Yousuf, S. et al. Raf kinase inhibitor protein (RKIP) blocks signal transducer and activator of transcription 3 (STAT3) activation in breast and prostate cancer. PloS one 9, e92478, doi:10.1371/journal.pone.0092478 (2014). 48 Xie, S. Y., Li, G., Han, C., Yu, Y. Y. & Li, N. RKIP reduction enhances radioresistance by activating the Shh signaling pathway in non-small-cell lung cancer. OncoTargets and therapy 10, 5605-5619, doi:10.2147/ott.S149200 (2017). 68 49 Noh, H. S. et al. Regulation of the epithelial to mesenchymal transition and metastasis by Raf kinase inhibitory protein-dependent Notch1 activity. Oncotarget 7, 4632-4646, doi:10.18632/oncotarget.6728 (2016). 50 Al-Mulla, F. et al. Raf kinase inhibitor protein RKIP enhances signaling by glycogen synthase kinase3β. Cancer research 71, 1334-1343, doi:10.1158/0008-5472.Can-10-3102 (2011). 51 Cohen, P. & Frame, S. The renaissance of GSK3. Nature reviews. Molecular cell biology 2, 769776, doi:10.1038/35096075 (2001). 52 Cardoso-Carneiro, D., Raquel-Cunha A., Pinheiro J., Gabriela-Freitas M., Fontão P., Reis R. M., Martinho, O. in Prognostic and Therapeutic Applications of RKIP in Cancer (ed Stavroula Baritaki Benjamin Bonavida) Ch. 20, 389-414 (Elsevier, 2020). 53 Fu, Z. et al. Effects of raf kinase inhibitor protein expression on suppression of prostate cancer metastasis. Journal of the National Cancer Institute 95, 878-889, doi:10.1093/jnci/95.12.878 (2003). 54 Hagan, S. et al. Reduction of Raf-1 kinase inhibitor protein expression correlates with breast cancer metastasis. Clinical cancer research : an official journal of the American Association for Cancer Research 11, 7392-7397, doi:10.1158/1078-0432.Ccr-05-0283 (2005). 55 Minoo, P. et al. Loss of raf-1 kinase inhibitor protein expression is associated with tumor progression and metastasis in colorectal cancer. American journal of clinical pathology 127, 820827, doi:10.1309/5d7mm22davgdt1r8 (2007). 56 Afonso, J. et al. Low RKIP expression associates with poor prognosis in bladder cancer patients. Virchows Archiv : an international journal of pathology 462, 445-453, doi:10.1007/s00428-0131388-2 (2013). 57 Martinho, O. et al. Loss of RKIP expression during the carcinogenic evolution of endometrial cancer. Journal of clinical pathology 65, 122-128, doi:10.1136/jclinpath-2011-200358 (2012). 58 Martinho, O. et al. Downregulation of RKIP is associated with poor outcome and malignant progression in gliomas. PloS one 7, e30769, doi:10.1371/journal.pone.0030769 (2012). 59 Martinho, O. et al. Loss of RKIP expression is associated with poor survival in GISTs. Virchows Archiv : an international journal of pathology 455, 277-284, doi:10.1007/s00428-009-0821-z (2009). 60 Martinho, O. et al. Raf Kinase Inhibitor Protein Expression and Prognostic Value in Soft Tissue Sarcomas. Pathobiology : journal of immunopathology, molecular and cellular biology 83, 41-46, doi:10.1159/000441227 (2016). 69 61 Martinho, O. et al. RKIP inhibition in cervical cancer is associated with higher tumor aggressive behavior and resistance to cisplatin therapy. PloS one 8, e59104, doi:10.1371/journal.pone.0059104 (2013). 62 Raquel-Cunha, A., Cardoso-Carneiro, D., Reis, R. M. & Martinho, O. Current Status of Raf Kinase Inhibitor Protein (RKIP) in Lung Cancer: Behind RTK Signaling. Cells 8, doi:10.3390/cells8050442 (2019). 63 Lamiman, K., Keller, J. M., Mizokami, A., Zhang, J. & Keller, E. T. Survey of Raf kinase inhibitor protein (RKIP) in multiple cancer types. Critical reviews in oncogenesis 19, 455-468, doi:10.1615/critrevoncog.2014011987 (2014). 64 Al-Mulla, F. et al. Raf kinase inhibitor protein: mechanism of loss of expression and association with genomic instability. Journal of clinical pathology 61, 524-529, doi:10.1136/jcp.2007.046987 (2008). 65 Wei, H. et al. Promoter methylation and expression of Raf kinase inhibitory protein in esophageal squamous cell carcinoma. Oncology letters 13, 1866-1872, doi:10.3892/ol.2017.5617 (2017). 66 Kim, G. E., Kim, N. I., Lee, J. S., Park, M. H. & Yoon, J. H. Reduced RKIP Expression is Associated With Breast Neoplastic Progression and is Correlated With Poor Outcomes and Aberrant Methylation in Breast Carcinoma. Applied immunohistochemistry & molecular morphology : AIMM 25, 467-474, doi:10.1097/pai.0000000000000323 (2017). 67 Beach, S. et al. Snail is a repressor of RKIP transcription in metastatic prostate cancer cells. Oncogene 27, 2243-2248, doi:10.1038/sj.onc.1210860 (2008). 68 Lee, J. et al. Network of mutually repressive metastasis regulators can promote cell heterogeneity and metastatic transitions. Proceedings of the National Academy of Sciences of the United States of America 111, E364-373, doi:10.1073/pnas.1304840111 (2014). 69 Li, J., Wang, Y., Song, Y., Fu, Z. & Yu, W. miR-27a regulates cisplatin resistance and metastasis by targeting RKIP in human lung adenocarcinoma cells. Molecular cancer 13, 193, doi:10.1186/1476-4598-13-193 (2014). 70 Du, Y. et al. MiR-543 Promotes Proliferation and Epithelial-Mesenchymal Transition in Prostate Cancer via Targeting RKIP. Cellular physiology and biochemistry : international journal of experimental cellular physiology, biochemistry, and pharmacology 41, 1135-1146, doi:10.1159/000464120 (2017). 70 71 Hatzl, S. et al. Increased Expression of miR-23a Mediates a Loss of Expression in the RAF Kinase Inhibitor Protein RKIP. Cancer research 76, 3644-3654, doi:10.1158/0008-5472.Can-15-3049 (2016). 72 Huang, L. et al. MicroRNA-224 targets RKIP to control cell invasion and expression of metastasis genes in human breast cancer cells. Biochemical and biophysical research communications 425, 127-133, doi:10.1016/j.bbrc.2012.07.025 (2012). 73 Cross-Knorr, S. et al. RKIP phosphorylation and STAT3 activation is inhibited by oxaliplatin and camptothecin and are associated with poor prognosis in stage II colon cancer patients. BMC cancer 13, 463, doi:10.1186/1471-2407-13-463 (2013). 74 Dangi-Garimella, S. et al. Raf kinase inhibitory protein suppresses a metastasis signalling cascade involving LIN28 and let-7. The EMBO journal 28, 347-358, doi:10.1038/emboj.2008.294 (2009). 75 Davudian, S., Mansoori, B., Shajari, N., Mohammadi, A. & Baradaran, B. BACH1, the master regulator gene: A novel candidate target for cancer therapy. Gene 588, 30-37, doi:10.1016/j.gene.2016.04.040 (2016). 76 Yun, J. et al. Signalling pathway for RKIP and Let-7 regulates and predicts metastatic breast cancer. The EMBO journal 30, 4500-4514, doi:10.1038/emboj.2011.312 (2011). 77 Beshir, A. B. et al. Raf kinase inhibitor protein suppresses nuclear factor-κB-dependent cancer cell invasion through negative regulation of matrix metalloproteinase expression. Cancer letters 299, 137-149, doi:10.1016/j.canlet.2010.08.012 (2010). 78 Lin, K. et al. The Role of B-RAF Mutations in Melanoma and the Induction of EMT via Dysregulation of the NF-κB/Snail/RKIP/PTEN Circuit. Genes & cancer 1, 409-420, doi:10.1177/1947601910373795 (2010). 79 Baritaki, S. et al. Mechanisms of nitric oxide-mediated inhibition of EMT in cancer: inhibition of the metastasis-inducer Snail and induction of the metastasis-suppressor RKIP. Cell cycle (Georgetown, Tex.) 9, 4931-4940, doi:10.4161/cc.9.24.14229 (2010). 80 Baritaki, S. et al. Inhibition of epithelial to mesenchymal transition in metastatic prostate cancer cells by the novel proteasome inhibitor, NPI-0052: pivotal roles of Snail repression and RKIP induction. Oncogene 28, 3573-3585, doi:10.1038/onc.2009.214 (2009). 81 Datar, I. et al. RKIP regulates CCL5 expression to inhibit breast cancer invasion and metastasis by controlling macrophage infiltration. Oncotarget 6, 39050-39061, doi:10.18632/oncotarget.5176 (2015). 71 82 Frankenberger, C. et al. Metastasis Suppressors Regulate the Tumor Microenvironment by Blocking Recruitment of Prometastatic Tumor-Associated Macrophages. Cancer research 75, 4063-4073, doi:10.1158/0008-5472.Can-14-3394 (2015). 83 Bainer, R. et al. Gene expression in local stroma reflects breast tumor states and predicts patient outcome. Scientific reports 6, 39240, doi:10.1038/srep39240 (2016). 84 Al-Mulla, F., Bitar, M. S., Feng, J., Park, S. & Yeung, K. C. A new model for raf kinase inhibitory protein induced chemotherapeutic resistance. PloS one 7, e29532, doi:10.1371/journal.pone.0029532 (2012). 85 Yang, K. et al. KRAS promotes tumor metastasis and chemoresistance by repressing RKIP via the MAPK-ERK pathway in pancreatic cancer. International journal of cancer 142, 2323-2334, doi:10.1002/ijc.31248 (2018). 86 Chatterjee, D. et al. RKIP sensitizes prostate and breast cancer cells to drug-induced apoptosis. The Journal of biological chemistry 279, 17515-17523, doi:10.1074/jbc.M313816200 (2004). 87 Baritaki, S., Yeung, K., Palladino, M., Berenson, J. & Bonavida, B. Pivotal roles of snail inhibition and RKIP induction by the proteasome inhibitor NPI-0052 in tumor cell chemoimmunosensitization. Cancer research 69, 8376-8385, doi:10.1158/0008-5472.Can-091069 (2009). 88 Odabaei, G. et al. Raf-1 kinase inhibitor protein: structure, function, regulation of cell signaling, and pivotal role in apoptosis. Advances in cancer research 91, 169-200, doi:10.1016/s0065230x(04)91005-6 (2004). 89 Jazirehi, A. R., Vega, M. I., Chatterjee, D., Goodglick, L. & Bonavida, B. Inhibition of the RafMEK1/2-ERK1/2 signaling pathway, Bcl-xL down-regulation, and chemosensitization of nonHodgkin's lymphoma B cells by Rituximab. Cancer research 64, 7117-7126, doi:10.1158/00085472.Can-03-3500 (2004). 90 Wei, J. et al. Didymin induces apoptosis through mitochondrial dysfunction and up-regulation of RKIP in human hepatoma cells. Chemico-biological interactions 261, 118-126, doi:10.1016/j.cbi.2016.11.026 (2017). 91 Li, H. et al. MiR-150 promotes cellular metastasis in non-small cell lung cancer by targeting FOXO4. Scientific reports 6, 39001, doi:10.1038/srep39001 (2016). 92 Cho, J. H. et al. NF2 blocks Snail-mediated p53 suppression in mesothelioma. Oncotarget 6, 10073-10085, doi:10.18632/oncotarget.3543 (2015). 72 93 Lee, S. J., Lee, S. H., Yoon, M. H. & Park, B. J. A new p53 target gene, RKIP, is essential for DNA damage-induced cellular senescence and suppression of ERK activation. Neoplasia (New York, N.Y.) 15, 727-737, doi:10.1593/neo.121862 (2013). 94 Giovannetti, E. et al. Molecular mechanisms and modulation of key pathways underlying the synergistic interaction of sorafenib with erlotinib in non-small-cell-lung cancer (NSCLC) cells. Current pharmaceutical design 19, 927-939 (2013). 95 Pavlova, N. N. & Thompson, C. B. The Emerging Hallmarks of Cancer Metabolism. Cell metabolism 23, 27-47, doi:10.1016/j.cmet.2015.12.006 (2016). 96 Wishart, D. S. Is Cancer a Genetic Disease or a Metabolic Disease? EBioMedicine 2, 478-479, doi:10.1016/j.ebiom.2015.05.022 (2015). 97 Warburg, O. On the origin of cancer cells. Science (New York, N.Y.) 123, 309-314, doi:10.1126/science.123.3191.309 (1956). 98 Vander Heiden, M. G., Cantley, L. C. & Thompson, C. B. Understanding the Warburg effect: the metabolic requirements of cell proliferation. Science (New York, N.Y.) 324, 1029-1033, doi:10.1126/science.1160809 (2009). 99 Liberti, M. V. & Locasale, J. W. The Warburg Effect: How Does it Benefit Cancer Cells? Trends in biochemical sciences 41, 211-218, doi:10.1016/j.tibs.2015.12.001 (2016). 100 Gatenby, R. A. & Gillies, R. J. Why do cancers have high aerobic glycolysis? Nature reviews. Cancer 4, 891-899, doi:10.1038/nrc1478 (2004). 101 Pereira-Nunes, A., Afonso, J., Granja, S. & Baltazar, F. Lactate and Lactate Transporters as Key Players in the Maintenance of the Warburg Effect. Advances in experimental medicine and biology 1219, 51-74, doi:10.1007/978-3-030-34025-4_3 (2020). 102 Ibrahim-Hashim, A. & Estrella, V. Acidosis and cancer: from mechanism to neutralization. Cancer metastasis reviews 38, 149-155, doi:10.1007/s10555-019-09787-4 (2019). 103 Cairns, R. A., Harris, I. S. & Mak, T. W. Regulation of cancer cell metabolism. Nature reviews. Cancer 11, 85-95, doi:10.1038/nrc2981 (2011). 104 Sanderson, S. M. & Locasale, J. W. Revisiting the Warburg Effect: Some Tumors Hold Their Breath. Cell metabolism 28, 669-670, doi:10.1016/j.cmet.2018.10.011 (2018). 105 Bui, T. & Thompson, C. B. Cancer's sweet tooth. Cancer cell 9, 419-420, doi:10.1016/j.ccr.2006.05.012 (2006). 106 Vander Heiden, M. G. & DeBerardinis, R. J. Understanding the Intersections between Metabolism and Cancer Biology. Cell 168, 657-669, doi:10.1016/j.cell.2016.12.039 (2017). 73 107 DeBerardinis, R. J. & Chandel, N. S. Fundamentals of cancer metabolism. Science advances 2, e1600200, doi:10.1126/sciadv.1600200 (2016). 108 Boroughs, L. K. & DeBerardinis, R. J. Metabolic pathways promoting cancer cell survival and growth. Nature cell biology 17, 351-359, doi:10.1038/ncb3124 (2015). 109 Min, H. Y. & Lee, H. Y. Oncogene-Driven Metabolic Alterations in Cancer. Biomolecules & therapeutics 26, 45-56, doi:10.4062/biomolther.2017.211 (2018). 110 Papa, S., Choy, P. M. & Bubici, C. The ERK and JNK pathways in the regulation of metabolic reprogramming. Oncogene 38, 2223-2240, doi:10.1038/s41388-018-0582-8 (2019). 111 Nagao, A., Kobayashi, M., Koyasu, S., Chow, C. C. T. & Harada, H. HIF-1-Dependent Reprogramming of Glucose Metabolic Pathway of Cancer Cells and Its Therapeutic Significance. International journal of molecular sciences 20, doi:10.3390/ijms20020238 (2019). 112 Denko, N. C. Hypoxia, HIF1 and glucose metabolism in the solid tumour. Nature reviews. Cancer 8, 705-713, doi:10.1038/nrc2468 (2008). 113 Hay, N. Reprogramming glucose metabolism in cancer: can it be exploited for cancer therapy? Nature reviews. Cancer 16, 635-649, doi:10.1038/nrc.2016.77 (2016). 114 Park, J. H., Pyun, W. Y. & Park, H. W. Cancer Metabolism: Phenotype, Signaling and Therapeutic Targets. Cells 9, doi:10.3390/cells9102308 (2020). 115 Adekola, K., Rosen, S. T. & Shanmugam, M. Glucose transporters in cancer metabolism. Current opinion in oncology 24, 650-654, doi:10.1097/CCO.0b013e328356da72 (2012). 116 Szablewski, L. Expression of glucose transporters in cancers. Biochimica et biophysica acta 1835, 164-169, doi:10.1016/j.bbcan.2012.12.004 (2013). 117 Pinheiro, C. et al. GLUT1 expression in pediatric adrenocortical tumors: a promising candidate to predict clinical behavior. Oncotarget 8, 63835-63845, doi:10.18632/oncotarget.19135 (2017). 118 Hong, S. Y., Yu, F. X., Luo, Y. & Hagen, T. Oncogenic activation of the PI3K/Akt pathway promotes cellular glucose uptake by downregulating the expression of thioredoxin-interacting protein. Cellular signalling 28, 377-383, doi:10.1016/j.cellsig.2016.01.011 (2016). 119 Sasaki, H. et al. Overexpression of GLUT1 correlates with Kras mutations in lung carcinomas. Molecular medicine reports 5, 599-602, doi:10.3892/mmr.2011.736 (2012). 120 Dong, Y., Tu, R., Liu, H. & Qing, G. Regulation of cancer cell metabolism: oncogenic MYC in the driver's seat. Signal transduction and targeted therapy 5, 124, doi:10.1038/s41392-020-002352 (2020). 74 121 Kruiswijk, F., Labuschagne, C. F. & Vousden, K. H. p53 in survival, death and metabolic health: a lifeguard with a licence to kill. Nature reviews. Molecular cell biology 16, 393-405, doi:10.1038/nrm4007 (2015). 122 Patra, K. C. et al. Hexokinase 2 is required for tumor initiation and maintenance and its systemic deletion is therapeutic in mouse models of cancer. Cancer cell 24, 213-228, doi:10.1016/j.ccr.2013.06.014 (2013). 123 Hoxhaj, G. & Manning, B. D. The PI3K-AKT network at the interface of oncogenic signalling and cancer metabolism. Nature reviews. Cancer 20, 74-88, doi:10.1038/s41568-019-0216-7 (2020). 124 Wong, N., Ojo, D., Yan, J. & Tang, D. PKM2 contributes to cancer metabolism. Cancer letters 356, 184-191, doi:10.1016/j.canlet.2014.01.031 (2015). 125 Christofk, H. R. et al. The M2 splice isoform of pyruvate kinase is important for cancer metabolism and tumour growth. Nature 452, 230-233, doi:10.1038/nature06734 (2008). 126 Yang, W. et al. ERK1/2-dependent phosphorylation and nuclear translocation of PKM2 promotes the Warburg effect. Nature cell biology 14, 1295-1304, doi:10.1038/ncb2629 (2012). 127 Ždralević, M. et al. Double genetic disruption of lactate dehydrogenases A and B is required to ablate the "Warburg effect" restricting tumor growth to oxidative metabolism. The Journal of biological chemistry 293, 15947-15961, doi:10.1074/jbc.RA118.004180 (2018). 128 Granja, S. et al. Disruption of BASIGIN decreases lactic acid export and sensitizes non-small cell lung cancer to biguanides independently of the LKB1 status. Oncotarget 6, 6708-6721, doi:10.18632/oncotarget.2862 (2015). 129 Miranda-Gonçalves, V. et al. Hypoxia-mediated upregulation of MCT1 expression supports the glycolytic phenotype of glioblastomas. Oncotarget 7, 46335-46353, doi:10.18632/oncotarget.10114 (2016). 130 Pinheiro, C. et al. Role of monocarboxylate transporters in human cancers: state of the art. Journal of bioenergetics and biomembranes 44, 127-139, doi:10.1007/s10863-012-9428-1 (2012). 131 Lin, X., Xiao, Z., Chen, T., Liang, S. H. & Guo, H. Glucose Metabolism on Tumor Plasticity, Diagnosis, and Treatment. Frontiers in oncology 10, 317, doi:10.3389/fonc.2020.00317 (2020). 132 Elf, S. E. & Chen, J. Targeting glucose metabolism in patients with cancer. Cancer 120, 774-780, doi:10.1002/cncr.28501 (2014). 133 Martinez-Outschoorn, U. E., Peiris-Pagés, M., Pestell, R. G., Sotgia, F. & Lisanti, M. P. Cancer metabolism: a therapeutic perspective. Nature reviews. Clinical oncology 14, 11-31, doi:10.1038/nrclinonc.2016.60 (2017). 81 207 Andrzejewski, S., Siegel, P. M. & St-Pierre, J. Metabolic Profiles Associated With Metformin Efficacy in Cancer. Frontiers in endocrinology 9, 372, doi:10.3389/fendo.2018.00372 (2018). 208 Leithner, K. et al. PCK2 activation mediates an adaptive response to glucose depletion in lung cancer. Oncogene 34, 1044-1050, doi:10.1038/onc.2014.47 (2015). 209 Gwinn, D. M. et al. Oncogenic KRAS Regulates Amino Acid Homeostasis and Asparagine Biosynthesis via ATF4 and Alters Sensitivity to L-Asparaginase. Cancer cell 33, 91-107.e106, doi:10.1016/j.ccell.2017.12.003 (2018). 210 Sun, Y. et al. Metabolic and transcriptional profiling reveals pyruvate dehydrogenase kinase 4 as a mediator of epithelial-mesenchymal transition and drug resistance in tumor cells. Cancer & metabolism 2, 20, doi:10.1186/2049-3002-2-20 (2014). 211 Shulman, R. G. & Rothman, D. L. The Glycogen Shunt Maintains Glycolytic Homeostasis and the Warburg Effect in Cancer. Trends in cancer 3, 761-767, doi:10.1016/j.trecan.2017.09.007 (2017). 212 Rapozzi, V. & Xodo, L. E. in Prognostic and Therapeutic Applications of RKIP in Cancer (eds Benjamin Bonavida & Stavroula Baritaki) Ch. 4, 77-93 (Elsevier, 2020). 213 Zhu, J. & Thompson, C. B. Metabolic regulation of cell growth and proliferation. Nature reviews. Molecular cell biology 20, 436-450, doi:10.1038/s41580-019-0123-5 (2019). 214 Valle-Mendiola, A. & Soto-Cruz, I. Energy Metabolism in Cancer: The Roles of STAT3 and STAT5 in the Regulation of Metabolism-Related Genes. Cancers 12, doi:10.3390/cancers12010124 (2020). 215 Gough, D. J. et al. Mitochondrial STAT3 supports Ras-dependent oncogenic transformation. Science (New York, N.Y.) 324, 1713-1716, doi:10.1126/science.1171721 (2009). 216 Kim, N. H. et al. Snail reprograms glucose metabolism by repressing phosphofructokinase PFKP allowing cancer cell survival under metabolic stress. Nature communications 8, 14374, doi:10.1038/ncomms14374 (2017). APPENDIX I: Supplementary data 83 Supplementary Tables Table S1: Top50 genes positively and negatively co-expressed (at mRNA level) with RKIP mRNA in lung adenocarcinoma. TCGA provisional data (www.cbioportal.org). RKIP upregulation RKIP downregulation Gene Cytoband p-value q-value Gene Cytoband p-value q-value ERP29 12q24.13 3.36e-20 2.24e-16 AVL9 7p14.3 4.86e-22 4.87e-18 MMAB 12q24.11 6.23e-20 2.93e-16 OSBPL3 7p15.3 7.31e-20 2.93e-16 ISCU 12q23.3 1.39e-19 4.64e-16 GNA12 7p22.3-p22.2 8.01e-19 2.00e-15 COQ8A 1q42.13 2.65e-19 7.57e-16 CD109 6q13 2.67e-18 5.95e-15 AMBP 9q32 3.78e-18 7.57e-15 ANLN 7p14.2 6.33e-18 9.74e-15 LDHD 16q23.1 4.54e-18 8.19e-15 GREB1L 18q11.1-q11.2 1.30e-17 1.64e-14 TMED6 16q22.1 4.91e-18 8.19e-15 SLC16A3 17q25.3 1.31e-17 1.64e-14 ADI1 2p25.3 1.21e-17 1.64e-14 C1GALT1 7p22.1-p21.3 2.11e-17 2.34e-14 PRKAB1 12q24.23 2.17e-17 2.34e-14 IL1RAP 3q28 2.29e-17 2.34e-14 SNX22 15q22.31 2.56e-17 2.34e-14 FAM220A 7p22.1 2.45e-17 2.34e-14 SLC48A1 12q13.11 2.58e-17 2.34e-14 ELK3 12q23.1 3.74e-17 3.12e-14 FAM104B Xp11.21 2.82e-17 2.45e-14 ITGAV 2q32.1 6.10e-17 4.88e-14 KLF15 3q21.3 8.59e-17 6.61e-14 ADGRF4 6p12.3 1.08e-16 7.75e-14 RANGRF 17p13.1 1.06e-16 7.75e-14 COLGALT1 19p13.11 1.27e-16 8.47e-14 ACAT1 11q22.3 1.15e-16 7.94e-14 PRDM8 4q21.21 1.63e-16 1.02e-13 DACT2 6q27 1.41e-16 9.07e-14 SPATS2L 2q33.1 1.87e-16 1.10e-13 AIFM1 Xq26.1 1.83e-16 1.10e-13 TTYH3 7p22.3 2.47e-16 1.42e-13 SMDT1 22q13.2 3.18e-16 1.77e-13 MMD 17q22 3.44e-16 1.86e-13 DRAIC 15q23 7.91e-16 4.06e-13 LAMC2 1q25.3 6.00e-16 3.16e-13 GSTA1 6p12.2 9.33e-16 4.56e-13 BCL10 1p22.3 9.03e-16 4.52e-13 MARC1 1q41 1.88e-15 8.19e-13 TPBG 6q14.1 1.01e-15 4.83e-13 NR0B2 1p36.11 2.47e-15 1.03e-12 GPRIN1 5q35.2 1.19e-15 5.54e-13 CDK2AP2 11q13.2 2.82e-15 1.14e-12 CERS6 2q24.3 1.27e-15 5.78e-13 BORCS7 10q24.32 2.84e-15 1.14e-12 BICD1 12p11.21 1.86e-15 8.19e-13 AQP7 9p13.3 3.90e-15 1.53e-12 ERO1A 14q22.1 2.47e-15 1.03e-12 UNC13B 9p13.3 6.40e-15 2.17e-12 PTPRH 19q13.42 3.98e-15 1.53e-12 COX6A1 12q24.31|12q24.2 6.40e-15 2.17e-12 STEAP1 7q21.13 4.28e-15 1.62e-12 COX14 12q13.12 7.14e-15 2.36e-12 GPR87 3q25.1 4.41e-15 1.64e-12 C15ORF61 15q23 1.56e-14 4.88e-12 PPP1R18 6p21.33 4.59e-15 1.67e-12 MTERF2 12q23.3 1.66e-14 5.03e-12 ZYX 7q34 5.23e-15 1.87e-12 POP5 12q24.31 1.79e-14 5.27e-12 CALU 7q32.1 6.39e-15 2.17e-12 GFRA3 5q31.2 1.79e-14 5.27e-12 TGFBI 5q31.1 7.20e-15 2.36e-12 FRAT1 10q24.1 2.28e-14 6.42e-12 STEAP1B 7p15.3 7.50e-15 2.42e-12 COX4I1 16q24.1 2.61e-14 7.16e-12 RASAL2 1q25.2 9.05e-15 2.87e-12 PLA2G4F 15q15.1 4.00e-14 1.08e-11 SPHK1 17q25.1 1.65e-14 5.03e-12 SIRT4 12q24.23-q24.31 5.84e-14 1.52e-11 NCS1 9q34.11 2.00e-14 5.79e-12 SLC25A38 3p22.1 6.44e-14 1.65e-11 NABP1 2q32.3 2.10e-14 6.00e-12 PCBD1 10q22.1 7.87e-14 1.97e-11 PLAUR 19q13.31 2.43e-14 6.76e-12 FBXO25 8p23.3 8.61e-14 2.10e-11 SLC2A1 1p34.2 4.42e-14 1.18e-11 TMEM116 12q24.12-q24.13 1.14e-13 2.71e-11 FKBP14 7p14.3 5.22e-14 1.37e-11 FCSK 16q22.1 1.33e-13 3.13e-11 YWHAG 7q11.23 6.62e-14 1.68e-11 FMC1 7q34 1.67e-13 3.71e-11 ANKIB1 7q21.2 8.08e-14 2.00e-11 ALDH2 12q24.12 1.85e-13 4.07e-11 TWISTNB 7p21.1 8.98e-14 2.17e-11 MRPS36 5q13.2 2.01e-13 4.32e-11 SEPTIN7 7p14.2 1.36e-13 3.17e-11 FRAT2 10q24.1 2.14e-13 4.48e-11 PLIN3 19p13.3 1.39e-13 3.20e-11 LRRC27 10q26.3 2.22e-13 4.59e-11 OSMR 5p13.1 1.54e-13 3.50e-11 PLPPR1 9q31.1 2.36e-13 4.83e-11 PCDH7 4p15.1 1.56e-13 3.52e-11 LYRM9 17q11.2 2.45e-13 4.95e-11 CBLB 3q13.11 1.93e-13 4.21e-11 SEC11C 18q21.32 2.71e-13 5.21e-11 FSCN1 7p22.1 2.10e-13 4.47e-11 GSTP1 11q13.2 2.82e-13 5.35e-11 ZNF267 16p11.2 2.15e-13 4.48e-11 *p-value (derived from Student's t-test); q-value (derived from Benjamini-Hochberg procedure); pand q-value<0.05: significant association.