International Journal of Molecular Sciences

Article HIF-1β Positively Regulates NF-κB Activity via Direct Control of TRAF6

Laura D’Ignazio 1,2, Dilem Shakir 3, Michael Batie 3 , H. Arno Muller 4 and Sonia Rocha 3,*

1 Centre for Gene Regulation and Expression, College of Life Sciences, University of Dundee, Dundee DD1 5EH, UK; [email protected] 2 The Lieber Institute for Brain Development, Department of Neurology, Johns Hopkins School of Medicine, Baltimore, MD 21205, USA 3 Department of Biochemistry, Institute of Integrative Biology, University of Liverpool, Liverpool L69 7ZB, UK; [email protected] (D.S.); [email protected] (M.B.) 4 Developmental Genetics Unit, Institute of Biology, University of Kassel, 34132 Kassel, Germany; [email protected] * Correspondence: [email protected]; Tel.: +44-(0)151-794-9084



Received: 9 March 2020; Accepted: 20 April 2020; Published: 24 April 2020


Abstract: NF-κB signalling is crucial for cellular responses to inflammation but is also associated with the hypoxia response. NF-κB and hypoxia inducible factor (HIF) transcription factors possess an intense molecular crosstalk. Although it is known that HIF-1α modulates NF-κB transcriptional response, very little is understood regarding how HIF-1β contributes to NF-κB signalling. Here, we demonstrate that HIF-1β is required for full NF-κB activation in cells following canonical and non-canonical stimuli. We found that HIF-1β specifically controls TRAF6 expression in human cells but also in Drosophila melanogaster. HIF-1β binds to the TRAF6 gene and controls its expression independently of HIF-1α. Furthermore, exogenous TRAF6 expression is able to rescue all of the cellular phenotypes observed in the absence of HIF-1β. These results indicate that HIF-1β is an important regulator of NF-κB with consequences for homeostasis and human disease.

Keywords: NF-κB TRAF6; HIF; ARNT; Drosophila; TNF

1. Introduction The transcription factor family NF-κB (Nuclear Factor kappa-light-chain-enhancer of activated B cells) is associated with a variety of stress responses in cells and organisms from response to pathogens to heat shock and chemotherapeutic drugs [1]. Activation of NF-κB can be achieved by three main signalling pathways, canonical, non-canonical and atypical [2], depending or not on the involvement of the Inhibitor of κB Kinase (IKK). Although some of the major components of these pathways have been identified, it is currently unclear how independent these truly are. As such, crosstalk and sharing of key molecules in these seemingly independent pathways are now being investigated. Canonical and non-canonical NF-κB activation relies on the engagement of ligands to specific receptors on the membrane. These include the Tumour Necrosis Factor (TNF) superfamily of ligands and receptors, Toll and Interleukin-1β (IL-1β) and their receptors, as well as CD40/BAFF/LTR [1]. Upon receptor binding, recruitment of specific adaptors initiates intracellular signalling cascades. These adaptors include TNF Receptor Type 1-Associated Death Domain (TRADD), Myeloid Differentiation Primary Response gene 88 (MYD88), and TNF Receptor Associated Factors (TRAFs). TRAFs (1–6) are important in multiple steps in the NF-κB signalling cascade, and their biological relevance is exemplified by a variety of phenotypes when they are deleted in mice [3]. As such, these pathways are crucially important for immune function and the cell and organism response to inflammation [1,3].

Int. J. Mol. Sci. 2020, 21, 3000; doi:10.3390/ijms21083000 www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2020, 21, 3000 2 of 19

Inflammation is also intimately connected to hypoxia, or decreased oxygen levels [4,5]. The cellular response to hypoxia is best known for the activation of the Hypoxia Inducible Factor (HIF) family of transcription factors. HIFs are heterodimers incorporating an oxygen labile HIF-α (HIF-1α, HIF-2α, HIF-3α) and a stably expressed HIF-1β (also known by its gene name Aryl Hydrocarbon Nuclear Translocator, ARNT)[6]. Oxygen control is achieved via a post-translational mechanism relying on the inactivation of Prolyl-Hydroxylases (PHDs) and von Hippel Lindau (VHL) tumor suppressor’s inability to ubiquitinate HIF-α in the absence of oxygen [7]. Published work demonstrated that hypoxia and inflammation have an intense molecular crosstalk in cells [4,5,8]. As such, NF-κB was shown to directly control HIF-1α, HIF-2α and HIF-1β in response to different cytokines [9–14]. Also, NF-κB interacts with HIF-1α and HIF-1β under specific conditions [15,16]. On the other hand, PHDs are reported to control NF-κB activity both in hydroxylase dependent and independent mechanisms in response to hypoxia [17], and VHL was also shown to regulate NF-κB signalling [18]. Furthermore, HIF-1α can restrict NF-κB activation in conditions of inflammation or infection [13]. In addition, and in specific immune cells, HIF-1α and HIF-2α were shown to be involved in normal immunological functions [8]. HIF-1β, although not altered in oxygen deprivation conditions, is essential for HIF activity in cells and organisms in conditions of hypoxia [6,19]. Furthermore, we had previously demonstrated that NF-κB can induce HIF-1β mRNA in response to TNF-α in human cells and in response to bacterial infection in Drosophila melanogaster [12,13]. Interestingly, HIF-1β was shown to bind RelB in response to CD30 stimulation of the non-canonical pathway, and control RelB transcriptional output [15]. However, whether HIF-1β is involved in broader NF-κB signalling or if this is important at the level of an organism has not been explored thus far. In this report, we demonstrate that HIF-1β is required for full NF-κB activation in cells following canonical and non-canonical stimulation. We found that HIF-1β is required for cell survival under basal and stimulated conditions. In addition, loss-of-function of Drosophila HIF-1β (known as tango, tgo) results in reduced viability following infection. Mechanistically, we identified TRAF6 as a gene specifically regulated by HIF-1β not only in human cancer cells but also in Drosophila melanogaster. Finally, exogenous expression of TRAF6 in the absence of HIF-1β was able to restore NF-κB signalling and prevent cell death. These results indicate that HIF-1β is required for full NF-κB signalling in cells, with implication in normal and disease settings.

2. Results

2.1. HIF-1β Is Required for Full NF-κB Activation in Response to Canonical and Non-Canonical Stimuli We had previously described a role for HIF-1α in the modulation of NF-κB activation in cells and Drosophila melanogaster [13]. In addition, published work had indicated that HIF-1β is direct target of NF-κB[12,13], and that HIF-1β is involved in non-canonical NF-κB signalling [15]. However, whether HIF-1β is a general regulator of NF-κB or this action is restricted to CD30-mediated NF-κB activation is not known. To address this question we performed efficient siRNA mediated depletion of HIF-1β (Figure1A, Figure S1A,B) in multiple NF- κB-luciferase reporter cell lines and assessed activity following non-canonical (LIGHT) and canonical (TNF-α) stimulation (Figure1B–D, Figure S1C). We could determine that in response to non-canonical NF-κB stimulus, HIF-1β depletion resulted in significantly less reporter gene activity (Figure1B). Interestingly, this was also the case when canonical signalling was initiated with TNF-α (Figure1C,D), the opposite e ffect that is known for HIF-1α [13]. Int. J. Mol. Sci. 2020, 21, 3000 3 of 19 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 3 of 20

A siRNA HIF-1β: - + - + TNF-α (hrs): 0 0.5 2 4 0 0.5 2 4 0 0.5 2 4 0 0.5 2 4 98 kDa HIF-1β 50 kDa Actin HeLa A549

B

180

y

t i

v 160 HeLa-kB

i t

c 140 ***

A

e s 120 ***

a ***

r e

f 100 Control

i c

u 80 siRNA HIF-1β_#1

L

B siRNA HIF-1β_#2

k 60

e v

i 40

t

a l

e 20 R 0 LIGHT: LIG0H Th 0 h LIG2H4T h 24 h

C D

180

180

y y

t t i

i 160 v

v 160 HeLa-kB A549-kB i

i ** t

t *** c

c 140 140 ***

A A

**

e e s

s 120 120 a

a *** ***

r r

e e f

f 100 100

i i

c c u

u 80 80

L L

B

B 60 60

k k

e e

v v i

i 40 40

t t

a a

l l e

e 20 20

R R 0 0 TNF-α: TN0F h0 h TN6F h6 h TNF-α: 0 h 6 h

Control siRNA HIF-1β Control siRNA HIF-1β 1 2 FigureFigure 1.1. HIFHIF-1-1ββ isis required required forfor full full NF- NFκB-κ reporterB reporter gene gene activity activity following following non-canonical non-canonical and canonical and 3 canonicalstimuli. ( stimuli.A) HeLa A. and HeLa A549 and cells A549 were transfected cells were with transfected control orwith HIF-1 controlβ siRNA or HIF oligonucleotides-1β siRNA 4 oligonucleotidesprior to treatment prior with to treatment 10 ng/mL with TNF- 10α ng/mLfor the TNF indicated-α for the periods indicated of time periods before of time lysis before and Western lysis 5 andblot Western analysis. blot Actin analysis. was used Actin as awas loading used control. as a loading (B) HeLa- control.κB luciferase B. HeLa- cellsκB luciferase were transfected cells were with 6 transfectedcontrol and with HIF-1 controlβ siRNAs and HIF for 24-1β h siRNAs prior to treatmentfor 24 h prior with to 100 treatment ng/mL LIGHT with 100 for anotherng/mL LIGHT 24 h prior for to 7 anotherlysis and 24 luciferaseh prior to activity lysis and measurements. luciferase activity All values measurements. were normalised All values to untreated were normalised sample. Graph to 8 untreateddepicts meansample. and Graph SEM fromdepicts four mean independent and SEM biological from four experiments. independent One biological way ANOVA experiments. analysis One was 9 wayperformed ANOVA and analysis levels was of significance performed wasand determined levels of significance as follows: was *** pdetermined< 0.001. (C )as HeLa- follows:κB luciferase *** p < 10 0.001.cells C. were HeLa transfected-κB luciferase as in cellsA, but were 10 transfected ng/mL TNF- asα inwas A, addedbut 10 forng/mL 6 h priorTNF-α to was lysis added and luciferase for 6 h 11 priormeasurements. to lysis and Graphluciferase depicts measurements. mean andSEM Graph from depicts four independentmean and SEM biological from four experiments. independent One 12 biologicalway ANOVA experiments. analysis Onewas performed way ANOVA and analysis levels of was significance performed are andindicated levels as of follows: significance ** p < are0.01, 13 indicated*** p < 0.001. as follows: (D) A549- ** p <κ 0.01,B luciferase *** p < 0.001. cells wereD. A549 treated- κB luciferase as in 1C. Graphcells were depicts treated mean as andin 1C SEM. Graph from 14 depictsthree independentmean and SEM biological from three experiments. independent One biological way ANOVA experiments. analysis One was way performed ANOVA and analysis levels of 15 wassignificance performed determined and levels asof follows:significance ** p determined< 0.01, *** p as< 0.001.follows: ** p < 0.01, ** *p < 0.001.

To determine whether our luciferase reporter results were also reflected at the level of endogenous 16 To determine whether our luciferase reporter results were also reflected at the level of targets, we investigated NF-κB subunits expression and some of their target genes in the presence or 17 endogenous targets, we investigated NF-κB subunits expression and some of their target genes in the absence of HIF-1β following treatment with LIGHT (Figure2) or TNF- α (Figure3). Levels of RelB 18 presence or absence of HIF-1β following treatment with LIGHT (Figure 2) or TNF-α (Figure 3). Levels and p100 were significantly reduced in the absence of HIF-1β in response to both stimuli (Figure2A; 19 of RelB and p100 were significantly reduced in the absence of HIF-1β in response to both stimuli Figure3A). This reduction in p100 was also observed in another cellular background, A549 lung cancer 20 (Figure 2A; Figure 3A). This reduction in p100 was also observed in another cellular background, cells (Figure S2A). In response to LIGHT, we could observe significant reduction in the mRNA levels of 21 A549 lung cancer cells (Figure S2A). In response to LIGHT, we could observe significant reduction in p100 and RelB, as well as in Rantes, a target of the non-canonical NF-κB pathway (Figure2B), when 22 the mRNA levels of p100 and RelB, as well as in Rantes, a target of the non-canonical NF-κB pathway HIF-1β was depleted. Furthermore, other p52-dependent target genes [20,21] were also reduced at 23 (Figure 2B), when HIF-1β was depleted. Furthermore, other p52-dependent target genes [20,21] were protein level in the absence of HIF-1β (Figure2C). 24 also reduced at protein level in the absence of HIF-1β (Figure 2C).

Int. J. Mol. Sci. 2020, 21, 3000 4 of 19 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 4 of 20

A B 160 p100 140

siRNA HIF-1β: - + 120 A

LIGHT (hrs): 0 4 24 0 4 24 N 100

98 kDa HIF-1β R m

80

e v p100 i 60 ***

98 kDa t

a l p52 e 40 50 kDa R 20 64 kDa RelB 0 50 kDa Actin siRNA: contro2l4 hHIF-1β

160 C 140 RelB 120

siRNA HIF-1β: - + A N

LIGHT (hrs): 0 4 24 0 4 24 R 100

m

98 kDa e 80

HIF-1β v i

t ***

a 60 l

50 kDa SKP2 e

R 40 36 kDa Cyclin D1 20 50 kDa Actin 0 siRNA: contro2l4 hHIF-1β

160 140 Rantes

120

A N

R 100

m

e 80

v * i

t 60

a

l e

R 40 20 0 siRNA: contro2l4 Hh IF-1β 1 Figure 2. HIF-1β is required for full NF-κB activity in response to LIGHT. (A) HeLa cells were 2 Figure 2. HIF-1β is required for full NF-κB activity in response to LIGHT. A. HeLa cells were transfected with control or HIF-1β siRNA oligonucleotides. Where indicated, cells were treated with 3 transfected with control or HIF-1β siRNA oligonucleotides. Where indicated, cells were treated with 100 ng/mL LIGHT for 4 or 24 h prior to whole cell lysis and Western blot analysis. Actin was used as a 4 100 ng/mL LIGHT for 4 or 24 h prior to whole cell lysis and Western blot analysis. Actin was used as loading control. (B) HeLa cells were transfected with control or HIF-1β siRNA oligonucleotides for 24 h 5 a loading control. B. HeLa cells were transfected with control or HIF-1β siRNA oligonucleotides for prior to treatment with 100 ng/mL LIGHT for other 24 h prior to lysis and RNA extraction. Following 6 24 h prior to treatment with 100 ng/mL LIGHT for other 24 h prior to lysis and RNA extraction. cDNA synthesis, qPCR was performed using primers for the indicated targets. Graphs depict mean 7 Following cDNA synthesis, qPCR was performed using primers for the indicated targets. Graphs and SEM from three independent biological experiments. Student t-test analysis was performed for 8 depict mean and SEM from three independent biological experiments. Student t-test analysis was each gene and levels of significance determined as: * p < 0.05, *** p < 0.001. (C) HeLa cells were treated 9 performed for each gene and levels of significance determined as: * p < 0.05, ***p < 0.001. C. HeLa cells and analysed as in 2A. 10 were treated and analysed as in 2A. In response to TNF-α, p105 levels were also reduced (Figure3A) in the absence of HIF-1 β, a 11 Inpredicted response result to TNF since-α, these p105 NF- levelsκB subunitswere also are reduced under the (Figure direct 3A control) in the of RelAabsence [22]. of Also, HIF as-1β expected,, a 12 predictedand given result that since it is these not anNF NF--κBκ Bsubunits target, RelA are levelsunder were the direct unaffected control by HIF-1 of RelAβ depletion [22]. Also in, responseas 13 expected,to TNF- andα given(Figure that3A). it is Importantly, not an NF-κB NF- target,κB targetsRelA levels such were as p100 unaffected and IL-8 by mRNA HIF-1β levels depletion were also 14 in responsereduced to inTNF the-α absence (Figure of 3A HIF-1). Importantly,β following NF TNF--κB αtargetstreatment such (Figure as p1003B,C). and IL Furthermore,-8 mRNA levels when we 15 were alsooverexpressed reduced in HIF-1 the abβ,sence levels of of HIF p100-1β and following IL-8 mRNA TNF were-α treatment also significant (Figure increased3B–C). Furthermore, following TNF- α 16 when treatment we overexpressed (Figure S2C,D), HIF-1 suggestingβ, levels of that p100 indeed and HIF-1IL-8 mRNAβ levels were have an also impact significant on NF- increasedκB signaling in 17 followingthe cells TNF we-α treatment investigated. (Figure S2C–D), suggesting that indeed HIF-1β levels have an impact on 18 NF-κB signaling in the cells we investigated.

Int. J. Mol. Sci. 2020, 21, 3000 5 of 19 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 5 of 20

A siRNA HIF-1β: - + TNF-α (hrs): 0 0.5 2 4 0 0.5 2 4 98 kDa HIF-1β 64 kDa RelA 64 kDa RelB

98 kDa p100 50 kDa p52 p105 98 kDa 50 kDa p50 50 kDa Actin

B C 1.4 1.4 1.2 p100 IL-8

*** *** 1.2 *** *** A

1 A 1

N

N R 0.8 R

m 0.8

m

e

e

v v i * 0.6 i

t 0.6

t

a

a

l

l e 0.4 e

R 0.4 R ** 0.2 0.2 0 0 TNF-α: 0 h 4 h TNF-α: 0 h 4 h Control siRNA HIF-1β Control siRNA HIF-1β 1 2 FigureFigure 3. HIF-1 3.β HIFis required-1β is required for full NF- for κ fullB activity NF-κB inactivity response in to response TNF-α.( toA TNF) HeLa-α. cells A. HeLa were cells were 3 transfectedtransfected with control with orcontrol HIF-1 orβ siRNA HIF-1βoligonucleotides siRNA oligonucleotides for 48 h. for Where 48 h indicated,. Where indicated, cells were cells were 4 treatedtreated with 10 with ng/mL 10 ng/mL TNF-α TNFfor 0.5,-α for 2 or 0.5, 4 h2 or prior 4 h toprior lysis to andlysis Western and Western blot analysis. blot analysis. Actin Actin was was used 5 used asas a loadinga loading control. control. (B B.) HeLa HeLa cells cells were were transfected transfected as as in in3A 3,A but, but also also treated treated where where indicated indicated with 10 6 with 10ng/mL ng/mL TNF TNF--α forfor 4 4 hh prior prior to to lysis lysis and and RNA RNA extraction. extraction. Following Following cDNA cDNA synthesis, synthesis, qPCR qPCR was 7 was performedperformed using using p100 p100 primers. primers. Graphs Graphs depict depict mean and mean SEM and from SEM four from independent four independent biological biological experiments. Student t-test analysis was performed and levels of significance determined as: * p < 0.05, 8 experiments. Student t-test analysis was performed and levels of significance determined as: * p < 0.05, *** p < 0.001. (C) HeLa cells were transfected as in 3A, but also treated where indicated with 10 ng/mL 9 *** p < 0.001. C. HeLa cells were transfected as in 3A, but also treated where indicated with 10 ng/mL TNF-α for 4 h prior to lysis and RNA extraction. Following cDNA synthesis, qPCR was performed 10 TNF-α for 4 h prior to lysis and RNA extraction. Following cDNA synthesis, qPCR was performed using IL-8 primers. Graph depicts mean and SEM from three independent biological experiments. 11 using IL-8 primers. Graph depicts mean and SEM from three independent biological experiments. One way ANOVA analysis was performed and levels of significance indicated as follows: ** p < 0.01, 12 One way ANOVA analysis was performed and levels of significance indicated as follows: ** p < 0.01, *** p < 0.001. 13 *** p < 0.001. It is well known that NF-κB signalling in response to TNF-α is used to prevent cell death [23]. 14To understandIt is the well biological known impactthat NF of-κ HIF-1B signallingβ, we investigated in response cellular to TNF apoptosis-α is used markers to prevent as well cell asdeath [23]. 15organism To understand survival using the the biological model system impactDrosophila of HIF-1β melanogaster, we investigated(Figure cellular4). Our apo analysisptosis markers revealed as well as 16that HIF-1organismβ depletion survival led tousing increased the model levels ofsystem apoptotic Drosophila markers melanogaster in control, LIGHT (Figure and 4).TNF- Our αanalysistreated revealed 17cells (Figurethat HIF4A,B).-1β Increased depletion levels led of to cleaved increased Poly-ADP levels of Ribose apoptotic Polymerase markers (PARP), in control cleaved, LIGHT caspase-3 and TNF-α 18and cleavedtreated caspase-7 cells (Figure were 4A observed–B). Increased when levels HIF-1 ofβ cleavedlevels were Poly reduced-ADP Ribose by siRNA Polymerase (Figure (4PARPA,B).), cleaved 19In Drosophilacaspase,- NF-3 andκB cleaved mediates caspase innate-7 immunitywere observed responses when toHIF infection-1β levels [24 were]. We reduced had previously by siRNA (Figure 20demonstrated 4A–B). In that Drosophila loss of NF-, NFκB-κ inB Drosophilamediates innateresults immunity in reduced responses viability in to response infection to [24] infection. We had with previously 21bacterial demonstrated strains such asthatSerratia loss of marcescens NF-κB in[13 Drosophila]. Importantly, results at in the reduced level of theviability whole in organism, response loss to infection 22of functionwith ofbacterialDrosophila strainsHIF-1 suchβ, Tangoas Serratia (Figure marcescens S3A,B), [13] resulted. Importantly, in reduced at the viability level inof responsethe whole to organism, 23infection loss with of functionSerratia marcescensof Drosophila(Figure HIF-41C).β, Tango (Figure S3A–B), resulted in reduced viability in response 24 Theseto infection results arewith intriguing, Serratia marcescens as deletion (Figure of Drosophila 4C). HIF-1α, Sima, produces similar results but 25due to overThese active results NF-κB[ are13]. intriguing, As such, we as analysed deletion levelsof Drosophila of an NF- HIFκB- target,1α, Sima, Drosomycin, produces assimilar wellas results but 26one ofdue the NF-to overκB subunits active NF in-DrosophilaκB [13]. As, Dorsal, such, we with analy or withoutsed levels Tango of an in NF response-κB target, to infection Drosomycin with , as well 27E. coli (Figureas one 4ofD). the While NF- theκB absencesubunits of in Tango Drosophila resulted, Dorsal in reduced, withlevels or without of Drosomycin, Tango inwe response observed to infection 28increases with in theE. coli levels (Figure of Dorsal 4D). (FigureWhile 4theD) andabsence Attacin of ATango (Figure resulted S3C), inin a reduced manner analogouslevels of Drosomycin to loss , we 29of Simaobserved [13]. These increases results suggest in the levels that deletion of Dorsal of Tango(Figure in Drosophila4D) and Attacinimpacts A on (Figure NF-κB activityS3C), in in a manner 30 analogous to loss of Sima [13]. These results suggest that deletion of Tango in Drosophila impacts on 31 NF-κB activity in mechanisms that are dependent and independent of its partner Sima. Taken 32 together, these findings indicate that HIF-1β is required for an appropriate NF-κB signalling in cells 33 and Drosophila melanogaster.

Int. J. Mol. Sci. 2020, 21, 3000 6 of 19

mechanisms that are dependent and independent of its partner Sima. Taken together, these findings Int. J. Mol.indicate Sci. 20 that20, 21 HIF-1, x FORβ PEERis required REVIEW for an appropriate NF-κB signalling in cells and Drosophila melanogaster6 of 20 .

A B siRNA HIF-1β: - + TNF-α (hrs): 0 4 LIGHT (hrs): 0 4 24 0 4 24 siRNA HIF-1β: - + - + 98 kDa HIF-1β 98 kDa HIF-1β PARP 98 kDa 98 kDa PARP Cleaved PARP 98 kDa Cleaved PARP 98 kDa Cleaved PARP 36 kDa Caspase 3 36 kDa Caspase 3 16 kDa Cleaved Caspase 3 22 kDa Cleaved Caspase 3 Caspase 7 36 kDa 36 kDa Caspase 7 Cleaved Caspase 7 16 kDa 22 kDa Cleaved Caspase 7 50 kDa Actin 50 kDa Actin

C

1.0

0.8 l

a 0.6 w1118 M ock (PBS)

v

i v

r Tango LF M ock (PBS) u

S 0.4 w1118 Serratia Tango LF Serratia *

0.2

0.0 8 10 12 14 16 18 20 22 Time (hours)

D 2 3 1.8 Drosomycin Dorsal 1.6 2.5

A *

A N

1.4 N

R 2

1.2 *** *** R m

m w1118

***

e 1

e 1.5

v

v i

0.8 i t

ns t ns Tango LF

a a

l 1

0.6 l

e

e R

0.4 R 0.2 0.5 0 0 Mock 2h E.coli Mock 2h E.coli 1 Figure 4. HIF-1β is required for viability of cells and Drosophila melanogaster following infection. 2 Figure 4. HIF-1β is required for viability of cells and Drosophila melanogaster following infection. (A) HeLa cells were transfected with control or HIF-1β siRNA oligonucleotides. Where indicated, 3 A. HeLa cells were transfected with control or HIF-1β siRNA oligonucleotides. Where indicated, cells cells were treated with 100 ng/mL LIGHT for 4 or 24 h prior to whole cell lysis and Western blot 4 were treatedanalysis. with Actin 100 was ng/mL used asLIGHT a loading for control.4 or 24 h (B prior) HeLa to cellswhole were cell transfected lysis and asWestern in 4A, but blot treated analysis. with 5 Actin 10was n/ mLused TNF- as aα loadingfor 4 h priorcontrol. to lysisB. HeLa and cells Western were blot transfected analysis. Actinas in 4 wasA, but used treated as a loading with 10 control. n/mL 6 TNF-α(C for) Wild-type 4 h prior adultto lysis flies and (w1118 Western) and blot HIF-1 analysis.β (Tango Actin) loss-of-function was used as fliesa loading (Tango control. LF) were C. pricked Wild-

7 type adultusing flies a thin (w1118 needle) and dipped HIF in-1 aβ diluted(Tango) overnight loss-of-function culture offliesSerratia (Tango marcescens LF) wereDb10 pricked (OD600 using= 0.2) a thin or in 8 needlea salinedipped solution in a diluted (PBS, Mock).overnight Groups culture of 100 of fliesSerratia were marcescens used and keptDb10 at (OD room600 temperature. = 0.2) or in a Survival saline 9 solutionwas (PBS, monitored Mock). and Groups expressed of 100 as theflies ‘estimated were used probability and kept ofatsurvival’. room temperature. The p-value Survival was obtained was 10 monitoredfrom log-rankand expressed statistical as the analysis: ‘estimated * p < probability0.05. (D) Wild-type of survival’. adult The flies P- (valuew1118 )was and obtained HIF-1β ( Tangofrom ) 11 log-rankloss-of-function statistical analysis: flies (Tango * p < LF) 0.05. were D. pricked Wild-type using adult a thin flies needle (w1118 dipped) and in aHIF culture-1β ( ofTangoEscherichia) loss-of coli- 12 functionor inflies a saline (Tango solution LF) were (PBS, pricked Mock). using RNA a was thin extracted needle dipped 2 h later. in Followinga culture of cDNA Escherichia synthesis, coli qPCRor in 13 a salinewas solution performed (PBS, using Mock). the indicatedRNA was primers. extracted Graphs 2 h later. depict Following mean and cDNA SEM from synthesis, threeindependent qPCR was 14 performedbiological using experiments. the indicated Student primers.t-test analysis Graphs was depict performed mean for and each SEM gene from and three levels independent of significance determined as: ns = not significant, * p < 0.05; *** p < 0.001. 15 biological experiments. Student t-test analysis was performed for each gene and levels of significance 16 determined as: ns = not significant, * p < 0.05; *** p < 0.001.

17 2.2. HIF-1β Is Required for TRAF6 Expression 18 To investigate the mechanism behind HIF-1β involvement in NF-κB signalling, we next analysed 19 levels of IKK activation markers such as phosphorylated IKK and IκB-α (Figure 5A). In the absence 20 of HIF-1β, we observed reduced levels of basal IκB-α, as expected, due to it being a NF-κB target. 21 Furthermore, levels of p-IKK and p-IκB-α were also reduced. Similarly, following LIGHT stimulation, 22 level of NF-κB inducing Kinase (NIK) were reduced (Figure 5B). These results suggested that HIF-1β

Int. J. Mol. Sci. 2020, 21, 3000 7 of 19

2.2. HIF-1β Is Required for TRAF6 Expression To investigate the mechanism behind HIF-1β involvement in NF-κB signalling, we next analysed levels of IKK activation markers such as phosphorylated IKK and IκB-α (Figure5A). In the absence of HIF-1β, we observed reduced levels of basal IκB-α, as expected, due to it being a NF-κB target. Furthermore, levels of p-IKK and p-IκB-α were also reduced. Similarly, following LIGHT stimulation, level of NF-κB inducing Kinase (NIK) were reduced (Figure5B). These results suggested that HIF-1 β is regulating a component upstream of the IKK complex. Directly upstream of IKK, are the Transforming Growth Factor (TGF)-β Kinase (TAK), and TAK Binding Protein (TAB) complex [25]. We thus analysed levels of the TAK-TAB complex, TRAFs but also Inhibitor of Apoptosis 1 (IAP1) and Receptor Interacting Serine/Threonine Kinase 1 (RIP1) (Figure5C), proteins involved in di fferent parts of the TNF-α signalling cascade to NF-κB[1]. Unexpectedly, we observed increases in the protein levels of IAP1, TRAF1 and TRAF5 (Figure5C). No changes were consistently seen for RIP1, TAK1 and TAB1 (Figure5C, Figure S4A). However, we observed reduced levels of TRAF6 in both cellular backgrounds we investigated (HeLa and A549), in the presence or absence of TNF-α, when HIF-1β was depleted (Figure5C, Figure S4A). The reduction of TRAF6 was also seen at transcriptional level when HIF-1 β was depleted (Figure3C), but the opposite was observed for TRAF3 (Figure S4B), suggesting a pleotropic effect for HIF-1β. Importantly, loss of function for HIF-1β in Drosophila, also resulted in reduced levels of Drosophila TRAF6 (dTRAF6) mRNA (Figure5E), implying that this is a conserved function for HIF-1β across organisms.

2.3. HIF-1β Binds to the TRAF6 Promoter and Controls TRAF6 Expression Independently of HIF-1α Since HIF-1β is part of several transcription factor complexes including HIF and Aryl Hydrocarbon Receptor (AHR), we interrogated the publicly available datasets for reports of HIF-1β binding at genomic areas controlling the TRAF6 gene. Our analysis revealed that in the breast cancer cell line T47D, HIF-1β binds at the proximal promoter of TRAF6, 178 bp upstream to 366 bp downstream of the transcription start site (TSS) (TSS -178/+366) (Figure6A) [ 26]. In combination with the analysis of other ENCODE datasets, we also performed a bioinformatic analysis for potential HIF-1β binding sites (HREs and AHR) in TRAF6 (Figure S5A) using the ALGGEN PROMO software tool. This revealed several potential hypoxia response elements (HRE) sites and AHR binding sites located both upstream and downstream to TRAF6 TSS (Figure S5A). We thus investigated if in our cell model, the site identified in T47D dataset and the AHR sites identified by our bioinformatic analysis were bona fide sites occupied by HIF-1β by Chromatin ImmunoPrecipitation (ChIP)-qPCR (Figure6B). Our analysis revealed that both the TSS -178/+366 site and a putative AHR site mapping 524 bp downstream of TRAF6 TSS (TSS +524) were significantly enriched for HIF-1β above control antibody levels (Figure6B), while control regions of TRAF6 genes did not contain any HIF-1β binding (Figure S5B). Interestingly, and also as predicted by our western blot analysis, treatment with TNF-α did not change HIF-1β levels present at the TRAF6 gene in either of the regions we analysed (Figure6B). As mentioned above, HIF-1β is a known component of several transcription factor complexes [27]. As we had previously implicated HIF-1α in the control of NF-κB signalling [13], we determined if HIF-1α was equally important for the control of TRAF6 levels. Western blot analysis indicated that HIF-1α depletion results in slightly increased levels of TRAF6 (Figure6C), the opposite results obtained following HIF-1β siRNA-mediated knockdown. These findings suggest that HIF-1α plays an inhibitory role for TRAF6 expression. Interestingly, although ChIP-qPCR analysis revealed a significant level of HIF-1α binding in control conditions to the TSS -178/+366 site of the TRAF6 gene, treatment with TNF-α resulted in a significant decrease of HIF-1α binding to this site (Figure6D). As expected, control regions of TRAF6 had no significant HIF-1α binding (Figure S5C). These data suggest that HIF-1β control of TRAF6 gene occurs independently of HIF-1α. We also investigated the potential involvement of HIF-2α (Figure6E; Figure S5D) and AHR (Figure6F; Figure S5E) in the control of TRAF6 expression. As observed with HIF-1α, HIF-2α depletion results in increased levels of TRAF6. However, depletion of AHR, results in reduced TRAF6 protein levels but not mRNA (Figure6F; Figure S5E). These results Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 7 of 20

1 is regulating a component upstream of the IKK complex. Directly upstream of IKK, are the 2 Transforming Growth Factor (TGF)-β Kinase (TAK), and TAK Binding Protein (TAB) complex [25]. 3 We thus analysed levels of the TAK-TAB complex, TRAFs but also Inhibitor of Apoptosis 1 (IAP1) 4 and Receptor Interacting Serine/Threonine Kinase 1 (RIP1) (Figure 5C), proteins involved in different 5 parts of the TNF-α signalling cascade to NF-κB [1]. Unexpectedly, we observed increases in the 6 protein levels of IAP1, TRAF1 and TRAF5 (Figure 5C). No changes were consistently seen for RIP1, 7 TAK1 and TAB1 (Figure 5C, Figure S4A). However, we observed reduced levels of TRAF6 in both 8 cellular backgrounds we investigated (HeLa and A549), in the presence or absence of TNF-α, when Int. J. Mol. Sci. 2020, 21, 3000 8 of 19 9 HIF-1β was depleted (Figure 5C, Figure S4A). The reduction of TRAF6 was also seen at 10 transcriptional level when HIF-1β was depleted (Figure 3C), but the opposite was observed for 11indicate TRAF3 that (Figure HIF-2α S4B),and suggesting AHR are not a pleotropic involved effect in TRAF6 for HIF control-1β. Importantly, mediated byloss HIF-1 of functionβ. In particular,for HIF- 12AHR 1 altersβ in Drosophila TRAF6 protein, also resulted in a manner in reduced independent levels of ofDrosophila transcriptional TRAF6 ( regulation.dTRAF6) mRNA (Figure 5E), 13 implying that this is a conserved function for HIF-1β across organisms. A B

siRNA HIF-1β: - + siRNA HIF-1β: - + TNF-α (hrs): 0 0.5 2 4 0 0.5 2 4 LIGHT (hrs): 0 4 24 0 4 24 98 kDa HIF-1β 98 kDa HIF-1β 98 kDa IKK-a 64 kDa TRAF3 98 kDa p-IKK 98 kDa NIK 98 kDa 36 kDa IkB-a IKK-α 50 kDa p-IkB-a 50 kDa Actin 50 kDa Actin C TNF-α (hrs): 0 4 siRNA HIF-1β: - + - + 98 kDa HIF-1β

64 kDa IAP1 50 kDa TRAF1 64 kDa TRAF2

50 kDa TRAF5

98 kDa RIP 64 kDa TRAF6 64 kDa TAB1

64 kDa TAK1 50 kDa Actin D E

1.4 TRAF6 1.4 dTRAF6

1.2 1.2

A A N

N 1.0 1.0

R R

m m

0.8 0.8

e e v

v ***

i i t

t 0.6 0.6

a a

l l

e e R R 0.4 0.4 *** 0.2 0.2 0.0 0.0 siRNA: Control HIF-1β w1118 Tango LF 14 15 FigureFigure 5. HIF-1 5. HIFβ -is1β required is required to controlto control TRAF6 TRAF6 levels. levels. ( AA.) HeLa HeLa cells cells were were transfected transfected with with control control or 16 or HIF-1HIFβ-1βsiRNA siRNA oligonucleotides. oligonucleotides. Where Where indicated, indicated, cells cells where where treated treated with with 10 ng/ 10mL ng /TNFmL- TNF-α for αthefor the periods of time depicted prior to whole cell lysis and Western blot analysis. Actin was used as a loading control. (B) HeLa cells were transfected as in 5A, but where indicated, cells were treated with 100 ng/mL LIGHT prior to whole cell lysis and Western blot analysis. Actin was used as a loading control. (C) HeLa cells were transfected as in 5A, but treated with 10 ng/mL TNF-α for 4 h prior to cell lysis and Western blot analysis. Actin was used as a loading control. (D) HeLa cells were transfected with control or HIF-1β siRNA oligonucleotides for 48 h prior to lysis and RNA extraction. Following cDNA synthesis, qPCR was performed using specific primers for TRAF6. Graphs depict mean and SEM from three independent biological experiments. Student t-test analysis was performed and levels of significance determined as follows: *** p < 0.001. (E) Wild-type adult flies (w1118) and HIF-1β loss-of-function flies (Tango LF) were used for RNA extraction and cDNA synthesis. qPCR analysis was performed for the levels of dTRAF6. Graph depicts mean and SEM from three independent biological experiments. Student t-test analysis was performed and levels of significance as follows: *** p < 0.001. Int. J. Mol. Sci. 2020, 21, 3000 9 of 19 Int. J. Mol. Sci. 2020, 21, x FOR PEER REVIEW 9 of 20

A

6

F

A hg19 36.53 mb 36.52 mb R

T chr11 36.52 mb

P I

h Binding site

C TSS -178 to +366

β

1

-

F

I H

B

ChIP: TRAF6 (TSS -178/+366) ChIP: TRAF6 (TSS +524)

t

t

u u

p ns

2.0 p 2.0

n

n i

i ns

o

*** o

t

t

1.5 1.5

e

e v

v *** i

*** i ***

t

t a

1.0 a 1.0

l

l

e

e

r

r

l

l a

0.5 a 0.5

n

n

g

g

i

i S 0.0 S 0.0 β 1β β β G -1 G - G 1 G 1 g F g F g - g - I I I I I IF I IF :H :H H H P P : : I I IP IP TNF-α: 0 h 4 h TNF-α: 0 h 4 h

C D ChIP: TRAF6 (TSS -178/+366)

TNF-α (hrs): 0 4 t u

p 2.0

- + - + n

siRNA HIF-1α: i

o *** t

1.5

98 kDa HIF-1α e v

i *** t

a 1.0 l

64 kDa e

TRAF6 r

ns l

a 0.5

50 kDa n g

Actin i S 0.0 α 1α 1 G - gG - Ig IF I IF :H :H IP IP TNF-α: 0 h 4 h E F TNF-α (hrs): 0 4 siRNA AHR: - #1 #2 siRNA HIF-2α: - + - + 64 kDa TRAF6 64 kDa TRAF6 50 kDa Actin 50 kDa Actin 1

2 FigureFigure 6. 6.HIF-1 HIF-β1βbinds binds toto TRAF6 gene gene and and regulates regulates its its expression expression independently independently of ofHIF HIF-1-1/2α/.2 A.α. 3 (A)Coverage Coverage tracks tracks of of HIF HIF-1-1ββ ChIPChIP-seq-seq at at the the TRAF6TRAF6 genegene in in T47D T47D cells. cells. TSS TSS = transcription= transcription start start site. site. B. 4 (B)ChIP ChIP for for HIF HIF-1-1ββ waswas performed performed in inlysates lysates derived derived from from HeLa HeLa cells cells treated treated or not or with not 10 with ng/mL 10 ng TNF/mL- 5 TNF-α forα for 4 h 4prior h prior to cross to cross-linking-linking and lysis. and lysis.Occupancy Occupancy at the indicated at the indicated sites was sites analysed was analysedby qPCR. byIn 6 qPCR.all cases, In all cases,Rabbit Rabbit IgG was IgG used was as used antibody as antibody control. control. Graphs Graphs depict depict mean mean and SEM and from SEM fromthree 7 threeindependent independent biological biological experiments. experiments. One One way way ANOVA ANOVA analysis analysis was was performed performed and and levels of of 8 significancesignificance indicated indicated as follows:as follows: ns ns= not = not significant, significant, *** ***p < p 0.001.< 0.001. (C C.) HeLaHeLa cellscells werewere transfected with 9 controlcontrol or HIF-1or HIFα-1siRNAα siRNA oligonucleotide, oligonucleotide, but but also, also, where where indicated, indicated, treated treated with with 1010 ngng/mL/mL TNF-TNF-αα for 10 4 h4 prior h prior to lysisto lysis and and Western Western blot blo analysis.t analysis. Actin Actin was was used used as as a loadinga loading control. control. (D D.) ChIPChIP forfor HIF-1HIF-1α 11 waswas performed performed in lysatesin lysates derived derived from from HeLa HeLa cells cells treated treated or or not not with with 10 10 ng ng/mL/mL TNF-TNF-αα forfor 44 hh priorprior 12 cross-linking and lysis. Occupancy at the indicated site was analysed by qPCR. Rabbit IgG was used cross-linking and lysis. Occupancy at the indicated site was analysed by qPCR. Rabbit IgG was used 13 as antibody control. Graph depicts mean and SEM from three independent biological experiments. as antibody control. Graph depicts mean and SEM from three independent biological experiments. 14 One way ANOVA analysis was performed and levels of significance determined as follows: ns = not One way ANOVA analysis was performed and levels of significance determined as follows: ns = not 15 significant, *** p < 0.001. E. HeLa cells were transfected with control or HIF-2α siRNA significant, *** p < 0.001. (E) HeLa cells were transfected with control or HIF-2α siRNA oligonucleotides, 16 oligonucleotides, but also, where indicated, treated with 10 ng/mL TNF-α for 4 h prior to lysis and but also, where indicated, treated with 10 ng/mL TNF-α for 4 h prior to lysis and Western blot analysis. 17 Western blot analysis. Actin was used as a loading control. F. HeLa cells were transfected with control Actin was used as a loading control. (F) HeLa cells were transfected with control or AHR siRNA oligonucleotides prior to lysis and Western blot analysis. Actin was used as a loading control. Int. J. Mol. Sci. 2020, 21, 3000 10 of 19

2.4. Exogenous TRAF6 Rescues NF-κB Signalling Defect in Cells Depleted of HIF-1β Our analysis revealed that TRAF6 is directly regulated by HIF-1β, with a potential impact on NF-κB signalling. Therefore, we hypothesised that if TRAF6 is important for HIF-1β control over NF-κB, restoration of TRAF6 would rescue the phenotypes we observed when HIF-1β was depleted. As such, we started this analysis using our reporter cell line. Here, exogenous TRAF6 was able to fully rescue NF-κB luciferase activity in the absence of HIF-1β (Figure S6) following TNF-α stimulation. Levels of endogenous p100 were also fully restored when TRAF6 overexpression was combined with HIF-1β depletion upon treatment with TNF-α or LIGHT (Figure7A,B). We also investigated IKK and IκB-α phosphorylation and levels. Here, overexpression of TRAF6 partially rescued levels of IKK phosphorylation but completely rescued levels of p-IκB-α (Figure7C). These findings strongly suggest that, indeed,Int. J. TRAF6Mol. Sci. 20 regulation20, 21, x FOR PEER by REVIEW HIF-1 β significantly impacts NF-κB signalling following11 of TNF-20 α.

A B LIGHT (hrs): 0 4 TNF-α (hrs): 0 4 Flag-TRAF6 (µg): 0 1 0 1 Flag-TRAF6 (µg): 0 1 0 1 siRNA HIF-1β: - + - + - + - + siRNA HIF-1β: - + - + - + - + 64 kDa TRAF6 98 kDa HIF-1β 98 kDa HIF-1β 64 kDa TRAF6 (short exp.)

98 kDa p100 64 kDa TRAF6 (long exp.)

50 kDa p52 98 kDa p100 50 kDa Actin 50 kDa p52 50 kDa Actin

C Control plasmid Flag TRAF6 siRNA HIF-1β: - + - + TNF-α (min): 0 10 30 0 10 30 0 10 30 0 10 30 98 kDa HIF-1β 64 kDa TRAF6

98 kDa p-IKK

98 kDa IKK-α 50 kDa p-IkB-α 50 kDa IkB-α

50 kDa Actin

D

Flag-TRAF6 (µg): 0 1 TNF-α (hrs): 0 4 0 4 siRNA HIF-1β: - + - + - + - + 98 kDa HIF-1β 64 kDa TRAF6 PARP 98 kDa Cleaved PARP 98 kDa Cleaved PARP 36 kDa Caspase 3 22 kDa Cleaved Caspase 3 36 kDa Caspase 7

22 kDa Cleaved Caspase 7 50 kDa 1 Actin

Figure2 7. ExogenousFigure 7. Exogenous TRAF6 TRAF6 is able is to able rescue to rescue HIF-1 HIFβ loss-1β inloss terms in terms of NF- of κNFB- signallingκB signalling and and cell cell survival. (3A ) HeLa cellssurvival. were A. transfectedHeLa cells were with transfected control with and control HIF-1 βandsiRNA HIF-1β oligonucleotides siRNA oligonucleotides as well as well as with as 1 µg of4 controlwith or TRAF6 1 μg of plasmid,control or TRAF6 prior toplasmid, treatment prior withto treatment 10 ng/ mLwith of10 TNF-ng/mLα offor TNF 0- orα for 4 h. 0 or Whole 4 h. Whole cell lysates 5 cell lysates were collected, and Western blot analysis for the indicated proteins was performed. Actin were collected, and Western blot analysis for the indicated proteins was performed. Actin was used 6 was used as a loading control. B. HeLa cells were transfected as in 7A, but, where indicated, also as a loading control. (B) HeLa cells were transfected as in 7A, but, where indicated, also treated with 7 treated with 100 g/mL LIGHT for 4 h prior to lysis and Western blot analysis. Actin was used as a 1008 g/mL LIGHTloading control. for 4 h C. prior HeLa to cells lysis were and transfected Western as blot in 7A analysis., and treated Actin where was indicated used aswith a loading10 ng/mLcontrol. (9C ) HeLaTNF cells-α were for the transfected depicted periods as in of7A time, and prior treated to lysis whereand Western indicated blot analysis. with 10 Actin ng /wasmL used TNF- asα a for the 10depicted periodsloading control. of time D. prior HeLa tocells lysis were and transfected, Western treated blot analysis. and analyzed Actin as in was 7A. used as a loading control. (D) HeLa cells were transfected, treated and analyzed as in 7A. 11 3. Discussion 12 In this report, we identified that HIF-1β, via the control of TRAF6, is required for full activation 13 of NF-κB signalling in human cancer cell models. The functional importance of HIF-1β for controlling 14 TRAF6 levels, and survival, was also observed in the model organism Drosophila melanogaster. Our 15 previous published work indicated that loss of the HIF-1α homologue in Drosophila, Sima, also 16 resulted in reduced viability of flies when infected with bacteria [13]. Mechanistically, this was due

Int. J. Mol. Sci. 2020, 21, 3000 11 of 19

Given that we had found that HIF-1β depletion resulted in increased levels of apoptotic markers in cells, we also investigated the importance of TRAF6 in this phenotype. Exogenous expression of TRAF6 strongly reduced the increase in apoptotic markers in cells depleted of HIF-1β (Figure7D). This was observed in diminished levels of cleaved PARP, cleaved caspase-3 and cleaved caspase-7. Taken together, these results imply that HIF-1β-mediated control of TRAF6 is required for full NF-κB activity in cells, and, as such, this regulatory mechanism controls cell survival.

3. Discussion In this report, we identified that HIF-1β, via the control of TRAF6, is required for full activation of NF-κB signalling in human cancer cell models. The functional importance of HIF-1β for controlling TRAF6 levels, and survival, was also observed in the model organism Drosophila melanogaster. Our previous published work indicated that loss of the HIF-1α homologue in Drosophila, Sima, also resulted in reduced viability of flies when infected with bacteria [13]. Mechanistically, this was due to elevated levels of NF-κB dependent target genes. Here, we show that HIF-1β depletion results in reduced levels of NF-κB activation in cells. In Drosophila, depletion of Tango (HIF-1β homologue) results in both reduced and increased levels of NF-κB targets. This implies that Tango shares some overlapping functions in response to infection with Sima, but also suggests that it possesses Sima independent functions. TRAF6 is an adaptor protein able to bridge signalling pathways derived from TNFR and IL-1β/Toll Receptors (reviewed in [28]). It possesses a TRAF domain, used for protein-protein interaction and a ring ubiquitin ligase domain, needed for poly-ubiquitination of substrates as well as auto-ubiquitination [3]. While the absolute requirement for the TRAF domain is well established, the requirement for the ubiquitinase function of the protein for signalling is controversial and context dependent [29]. TRAF6 knockout mice are prenatal and postnatal lethal [30]. Conditional knockout models revealed essential roles for TRAF6 in a variety of immune cellular backgrounds but also in epithelial cell types [31–34], indicating how important this factor is for normal cellular homeostasis. More recently, TRAF6 was also implicated in cancer signalling, by regulating epithelial to mesenchymal transition (EMT) in colon cancer [35], DNA damage in breast cancer backgrounds [36], and involvement in autophagy responses [37]. Our results demonstrate the importance of TRAF6 in cell survival, following NF-κB stimulation, in cancer cells. This is in line with reduced NF-κB transcriptional activity required for the activation of pro-survival genes following TNF-α [23]. Interestingly, inhibition of TRAF6 was also associated with increased apoptosis when cells were treated with cinchonine, a natural occurring chemical compound [38]. Although much is known about TRAF6 regulation at the post-translational level, very little is known about how its promoter is controlled. TRAF6 mRNA was found to be over expressed in cancer [39], but the mechanisms underlying this increased expression have not been determined yet. Our data identified TRAF6 as a gene specifically down-regulated when HIF-1β, and its homologue in Drosophila melanogaster, Tango, are depleted. Analysis, using Ominer [40,41], of public datasets for ChIP-seq revealed that several transcription factors occupy genomic regions within the TRAF6 promoter, including Basic Helix-Loop-Helix ones such as HIFs. Furthermore, analysis of publicly released datasets for HIF-1β ChIP-seq identified a specific binding site in the promoter region of TRAF6 in T47D breast cancer cells [26]. We could validate this occurrence in the cell lines used in this study, and we also identified another binding site in the vicinity of TRAF6 transcriptional start site. These results suggest a potential direct control of HIF-1β over the TRAF6 gene. Interestingly, although we could detect HIF-1α binding to one of these sites, TRAF6 expression increased following depletion of HIF-1α, suggesting that HIF-1β control over TRAF6 occurs independently of HIF, potentially via another of its binding partners. AHR depletion, while reducing TRAF6 protein did not reduce TRAF6 mRNA levels, suggesting that AHR controls TRAF6 translation or protein stability indirectly. Another possibility would be that HIF-1β impacts on NF-κB-dependent regulation of TRAF6. In fact, NF-κB also possesses several binding sites in the TRAF6 promoter, and given that we found that HIF-1β Int. J. Mol. Sci. 2020, 21, 3000 12 of 19 depletion significantly reduces NF-κB transcriptional activity, this would be a potential mechanism behind TRAF6 mRNA decrease. Analysis of RelA ChIP-seq datasets available on ENCODE [42,43], indicated that RelA binds to one of the sites we identified as HIF-1β binding site to the TRAF6 gene (TSS -178/+366). It is possible that a cooperative binding between HIF-1β and RelA occurs. Future studies depleting NF-κB could help answer these questions. The implications of our findings are relevant for situations of homeostatic control but also pathological deregulation such as inflammation and cancer. It would be predicted that a strong correlation between HIF-1β and TRAF6 would occur in tissues both in normal conditions as well as in disease situations. Analysis of TGCA data and expression correlation analysis using GEPIA 2 [44] revealed that HIF-1β and TRAF6 mRNA are very strongly correlated in normal tissues such as pancreas and breast (Figure S7A). In cancer situations, this correlation in mRNA levels was found to be also strong in pancreatic cancer, breast cancer, kidney renal cell carcinoma (RCC), uveal melanoma, and diffuse B-cell lymphoma (Figure S7B). These analyses are remarkably supportive of our cellular findings. TRAF6 function in Drosophila melanogaster was also shown to connect NF-κB and survival pathways [45,46]. In, addition, dTRAF6 was shown to control the Notch signalling in these organisms [47]. Our results demonstrated that dHIF-1β is required for dTRAF6 levels, and its loss of function also resulted in reduced viability in response to infection, suggesting a conservation of the responses we observed in cancer cells. It would be interesting to determine if the dTRAF6 promoter also contains dHIF-1β binding sites. Further studies are needed also to establish a direct regulatory mechanisms in Drosophila. Taken together our analyses revealed an unexpected link between HIF-1β and NF-κB signalling, suggesting that these transcription factors are strongly linked in signal transduction in homeostasis and disease.

4. Materials and Methods

4.1. Cell Lines and Growth Conditions All cells, with the exception of 786-O, were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Lonza, Slough, UK) supplemented with 10% fetal bovine serum (FBS, Invitrogen/ThermoFisher, Paisley, UK), 1% penicillin-streptomycin (Lonza, Slough, UK), and 1% L-glutamine (Lonza, Slough, UK) at 5% CO2 and 37 ◦C for no more than 30 passages. 786-O cells were maintained in RPMI (Gibco/ThermoFisher, Paisley, UK), supplemented as above. Cells were routinely tested for mycoplasma contamination using MycoAlert Mycoplasma Detection Kit (Lonza, Slough, UK). Human cervix carcinoma HeLa, and human lung carcinoma A549 were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA). HeLa-κB Luciferase cells, containing an integrated copy of the 3x-κB ConA luciferase reporter plasmid and described in [48], were kindly gifted by Prof. Ron Hay (School of Life Sciences, University of Dundee, UK). A549-κB Luciferase cells were previously created in the lab, by transfecting the NF-κB-reporter construct pGL4.32(luc2P/NF-κB-RE/Hygro), selecting individual clones on the basis of their response to TNF-α, and maintaining the derived cell lines with 150 µg/mL Hygromycin B (Sigma, Gillinham, UK) [49].

4.2. Cell Transfection Small interfering RNA oligonucleotides were purchased from Eurofins Genomics (Ebersberg, Germany) and used in a final concentration of 27 nM. siRNA transfections in all cell lines were performed using Interferin (Polyplus, Illkirch, France) according to manufacturer’s instructions. Oligonucleotide sequences used for siRNA knockdown are as follows: Control- AACAGUCGCGUUUGCGACU; HIF-1β_#1- GGUCAGCAGUCUUCCAUGA; HIF-1β_#2- GAAAGAAACAUGUGAGUAA; HIF-1α - GCAUAUAUCUAGAAGGUAU; HIF-2α_#1- CAGCAUCUUUGAUAGCAGUTT; Int. J. Mol. Sci. 2020, 21, 3000 13 of 19

HIF-2α _#2- GGCAGAACUUGAAGGGUUA; AHR_#1- UACUUCCACCUCAGUUGGCTT; AHR_#2- GGACAAACUUUCAGUUCUU. To perform DNA plasmid overexpression in all cell lines, TurboFect (ThermoFisher, Paisley, UK) or GeneJuice (Novagen/ThermoFisher, Paisley, UK) were used according to manufacturer’s instructions. The following expression plasmids were used in this study: GFP-C3 (Clontech/Takara, Montain View, CA, USA); Flag-pcDNA3.1 (a gift from Stephen Smale, Addgene plasmid #20011, Watertown, MD, USA); pEBB-HIF-1β GFP (kind gift from Colin Duckett, Ann Habour, MI, USA); pCMV5-FLAG − TRAF6 (MRC-PPU reagents, Dundee, UK).

4.3. Cell Treatments To stimulate the inflammatory response, human recombinant TNF-α (Peprotech, London, UK) and human recombinant LIGHT (TNFSF14, Peprotech, London, UK) were dissolved in sterile PBS and used at a final concentration of 10 ng/mL and 100 ng/mL, respectively.

4.4. Luciferase Assay 2 105 cells stably transfected with a luciferase reporter gene were seeded in 6-well plates, and × transfected with appropriate siRNA or DNA plasmid, according to procedures previously described above. Then, cells were stimulated with TNF-α or LIGHT (κB luciferase reporter), for the indicated times, and harvested with 400 µL of Passive Lysis Buffer (Promega, Southampton, UK). Luciferase activity was measured according to manufacturer’s instructions (Luciferase Assay System, Promega, Southampton, UK), and normalised to protein concentration (Bradford, BioRad, Watford, UK). All experiments were performed a minimum of three times before calculating means and standard error of the means.

4.5. RNA Extraction cDNA Synthesis and Real Time Quantitative PCR Analysis PeqGOLD total RNA kit (Peqlab, Bishop’s Waltham, UK) or PureLink RNA Mini Kit (Life Technologies/ThermoFisher, Paisley, UK) were used to extract total RNA from cells according to the manufacturer’s instructions. RNA was converted to cDNA using Quantitect Reverse Transcription Kit (Qiagen, Manchester, UK) or First Strand cDNA Synthesis kit (ThermoFisher, Paisley, UK). For quantitative PCR, Brilliant II Sybr green kit (Stratagene/Agilent, Stockport, UK), and recommended MX3005P 96-well skirted plates were used to analyse samples on the Mx3005P qPCR platform (Stratagene/Agilent, Stockport, UK). Alternatively, PerfeCTa Sybr Green FastMix (Quanta Bio, Bervely, MA, USA) with ROX dye added in 1:250 ratio, and recommended Microamp Optical 96-well reaction plates were used to analyse samples on the QuantStudio 6 Flex qPCR platform (Applied Biosystem/ThermoFisher, Paisley, UK). Actin or 18S were used as normalising genes. RT-PCR results were analysed by the ∆∆Ct method. The primers used for gene expression analysis by RT-PCR are: 18S, For: 50- AAACGGCTACCACATCCAAG-30, Rev: 50-CGCTCCCAAGATCCAACTAC-30. Actin, For: 50-CTGGGAGTGGGTGGAGGC-30, Rev: 50-TCAACTGGTCTCAAGTCAGTG-30. HIF-1β, For: 50-CAAGCCCCTTGAGAAGTCAG-30, Rev: 50-GAGGGGCTAGGCCACTATTC-30. IL-8, For: 50-CCAGGAAGAAACCACCGGA-30, Rev: 50-GAAATCAGGAAGGCTGCCAAG-30. p100, For: 50-AGCCTGGTAGACACGTACCG-30, Rev: 50-CCGTACGCACTGTCTTCCTT-30. TRAF3, For: 50-CTCACAAGTGCAGCGTCCAG-30, Rev: 50-GCTCCACTCCTTCAGCAGGTT-30. TRAF6, For: 50-CCTTTGGCAAATGTCATCTGTG-30, Rev: 50-CTCTGCATCTTTTCATGGCAAC-30. AHR, For: 50-ACTCCACTTCAGCCACCATC-30, Rev: 50-ATGGGACTCGGCACAATAAA-30. Rantes, For: 50-GTCGTCTTTGTCACCCGAAAG-30, Rev: 50-TCCCGAACCCATTTCTTCTCT-30. RelB, For: 50-TCCCAACCAGGATGTCTAGC-30, Rev: 50-AGCCATGTCCCTTTTCCTCT-30.

4.6. Protein Lysis and Western Blotting Cells were lysed using 100 µL of whole cell protein lysis buffer (20 mM Tris pH 7.6, 150 mM NaCl, 0.75% NP-40, 5 mM NaF, 500 µM Na3VO4, and 1 Pierce Protease Inhibitor Mini Tablet (EDTA Free, Int. J. Mol. Sci. 2020, 21, 3000 14 of 19

ThermoFisher, Paisley, UK) per 10 mL of buffer) or RIPA buffer (50 mM Tris pH 8, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 10 mM NaF, 1 mM Na3VO4, and 1 Pierce Protease Inhibitor Mini Tablet (EDTA Free, ThermoFisher, Paisley, UK) per 10 mL of buffer). Upon collection, cells were kept on ice for 30 min before centrifugation at 16,060 g at 4 C for 15 min. The supernatant × ◦ was collected and stored at 80 C. Protein concentration was determined using Bradford (BioRad, − ◦ Watford, UK) method. 20–30 µg of protein was prepared in 2 SDS loading buffer (100 mM Tris-HCl × pH 6.8, 20% glycerol, 4% SDS, 200 mM DTT, and Bromophenol Blue), and incubated for 5–10 min at 105 ◦C. Western blotting was performed as described in [12]. Briefly, samples were loaded into an SDS-page gel (Tris-HCl poly-acrylamide gel) previously prepared and run at 80–120 volts in Running Buffer (25 mM Tris, 0.195 M glycine, and 0.1% SDS). The gel was then transferred in a semi-dry transfer (BioRad, Watford, UK) into a PVDF membrane (Millipore, Feltham, UK) for 1.5–2 h at 15 volts/0.80 mA in Transfer Buffer (50 mM Tris, 40 mM glycine, 0.001% SDS and 10% methanol). Then, the membrane was blocked with 10% Milk in TBS-tween buffer (20 mM Tris pH 7.6, 150 mM NaCl, 0.1% Tween) for 10 min, followed by three 5 min washes with TBS-tween buffer. Membranes were incubated with primary antibodies for 1 h at room temperature or overnight at 4 ◦C, in accordance with primary antibodies’ manufacturer instructions. Primary antibodies purchased from Cell signalling (Leiden, The Netherlands) were: caspase 3 (#9665S), caspase 7 (#128275), cleaved caspase 3 (#9664S), cleaved caspase 7 (#8438S), cleaved PARP (#5625), HIF-1β (#5537S), IAP1 (#4952), IKK-α (#2682), IκB-α (#4821), PARP (#9532S), phospho-IKK-α (#2681S), phospho-IκB-α (#9246S), RelB (#10544), RIP (#4926), TAB1 (#3225), TAK1 (#4505), TRAF1 (#4715), TRAF2 (#4724), TRAF5 (#41658), TRAF6 (#8028), HIF-2α (#7096), SKP2 (#4358), NIK (#4994). Primary antibodies against HIF-1α (#610958) were from BD Biosciences (Workingham, UK), anti-p100/p52 (#05-361) from Millipore (Feltham, UK), Cyclin B1 (#ab137875) from Abcam (Cambridge, UK), p105/p50 (#sc-7178) and RelA (#sc-372) from Santa Cruz (Dallas, TX, USA), and anti-b-Actin (#66009-1-1g) from Proteintech (Manchester, UK). Membranes were then washed three times with TBS-tween and incubated with the appropriate secondary HRP antibody (anti-mouse IgG, HRP-linked, Cell Signalling #7076, Leiden, Holland; anti-rabbit IgG, HRP-linked, Cell Signalling #7074, Leiden, The Netherlands). After washes, membranes were developed using ECL solution (Pierce/ThermoFisher, Paisley, UK).

4.7. Chromatin Immunoprecipitation Chromatin Immunoprecipitation (ChIP) was performed adapting the protocol described in [21]. Cells were plated and grown to 70–80% confluency on 150 mm plates in 16 mL of appropriate culturing media. When indicated cells were treated with TNF-α for 4 h. Then, proteins and chromatin were cross-linked with 1% formaldehyde at 37 ◦C for 10 min. To quench the cross-linking, glycine was added to a final concentration of 0.125 M for 5 min at 37 ◦C. Cells were washed twice with ice-cold PBS, then scraped and centrifuged at 1000 rpm in a Beckman Coulter’s Allegra X-12 benchtop centrifuge for 5 min. The supernatant was removed and the pellet resuspended in 400 µL of ChIP lysis buffer (1% SDS, 10 mM EDTA, 50 mM Tris-HCl pH 8.1, and 1 Pierce Protease Inhibitor Mini Tablet (EDTA Free, ThermoFisher, Paisley, UK), up to 10 mL of buffer), before being snap-frozen in dry ice and stored at 80 C. Once thawed on ice, 200 µL aliquots of each sample were transferred into 1.5 mL TPX − ◦ Polymethylpentene (PMP) tubes (Diagenode, Seraing, Belgium) to improve sonication and shearing efficiency. Samples were sonicated in Bioruptor NGS (Diagenode, Seraing, Belgium) at 4 ◦C for five cycles of 30 s ON/30 s OFF, at high intensity amplitude. This sonication procedure was repeated four times. Supernatants were recovered by centrifugation at maximum speed in a benchtop centrifuge for 10 min at 4 ◦C prior storage of 10% of each sample as input. Remaining samples were split into 120 µL aliquots before being diluted 10 fold in ChIP Dilution Buffer (1% Triton X-100, 2 mM EDTA, 150 mM NaCl, 20 mM Tris-HCl pH 8.1). Diluted samples were pre-cleared for 2 h at 4 ◦C by incubation with 2 µg of sheared salmon sperm DNA and 20 µL of protein G-Sepharose (50% slurry), previously washed in cold PBS. Immunoprecipitation was performed overnight on the remaining samples with addition of appropriate antibodies (5 µL of anti-HIF-1α (Active Motif, #39665, La Hupe, Belgium), Int. J. Mol. Sci. 2020, 21, 3000 15 of 19

4 µg of anti-HIF-1β (Bethyl, #A302-765A, Montegomery, TX, USA) or Normal Rabbit IgG (#I5381, Sigma, Gillinham, UK) as negative control, and 0.1% of BRIJ-35 detergent. The following day, immune complexes were captured by incubation with 30 µL of protein G-Sepharose (50% slurry, previously washed with cold PBS) and 2 µg of salmon sperm DNA for 2.5 h at 4 ◦C. The immunoprecipitates were washed sequentially for 5 min at 4 ◦C in 1 mL of cold Wash Buffer 1 (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl pH 8.1, 150 mM NaCl), 1mL of cold Wash Buffer 2 (0.1% SDS, 1% Triton X-100, 2 mM EDTA, 20 mM Tris-HCl pH 8.1, 500 mM NaCl), and 1 mL of cold Wash Buffer 3 (0.25 M LiCl, 1% Nonidet P-40, 1% deoxycholate, 1 mM EDTA, 10 mM Tris-HCl pH 8.1). Next, beads were washed twice with 500 µL of Tris-EDTA (TE) buffer. Chromatin reverse cross-linking and DNA elution were performed using the IPure kit v2 (Diagenode, Seraing, Belgiumfollowing manufacturer’s instructions. Briefly, 10% inputs and beads were resuspended in 90 µL and 100 µL of Elution Buffer mix, respectively. The elution buffer mix was prepared with 115.4 µL of Buffer A and 4.6 µL of Buffer B per sample. All samples were incubated overnight at 65 ◦C on a thermomixer with continuous shaking at 300 rpm. The following day, supernatants were recovered by centrifugation at 1000 rpm for 1 min in a benchtop centrifuge. 2 µL of carrier, 100 µL of 100% isopropanol and 10 µL of magnetic beads were added to each input and immunoprecipitate sample, then incubated on a rotating wheel (40 rpm) for 10 min at room temperature. After a quick centrifugation, tubes were placed into a magnetic rack for separation of buffer, then discarded, and beads. Captured beads were gently mixed to 100 µL of Wash Buffer 1 (previously diluted with 100% isopropanol with 1:1 ratio), prior incubation on rotating wheel (40 rpm) for 5 min at room temperature. Following another step of buffer separation, beads were gently mixed to 100 µL of Wash Buffer 2 (previously diluted with 100% isopropanol with 1:1 ratio), and incubated on rotating wheel (40 rpm) for 5 min at room temperature. Finally, DNA elution was performed adding to captured beads 25 µL of Buffer C and incubating tubes on a rotating wheel (40 rpm) for 15 min at room temperature. After separation on magnetic rack, the first fraction of eluted DNA was transferred into a new storage tube, while captured beads were subjected to a second step of DNA elution by adding other 25 µL of Buffer C, to obtain a final volume of 50 µL of purified DNA. 3 µL DNA were used for RT-PCR analysis with primers specifically designed on TRAF6 promoter regions: TRAF6 (TSS -178/+366), For: 50-AAGGAGACTCACCGTTCTA-30, Rev: 50-TCTGTGTCCGTCCTCTAC-30. TRAF6 (TSS +524), For: 50-GGTCGAGGACACCGTTC-30, Rev: 50-GTGGAATGAGCGAGGAAGA-30. When performed quantitative PCR on ChIP samples, primers designed on coding regions were used as negative control.

4.8. Chromatin Immunoprecipitation Sequencing Analysis T47D cell HIF-1β ChIP-seq data was downloaded from the NCBI SRA (SRX666557) and GEO (GSM1462476). Coverage tracks were generated using the R Bioconductor package, Gviz [50]. RelA ChIP-seq datasets were downloaded from the ENCODE portal [46](https://www.encodeproject.org/) with the following identifiers: ENCSR000EBM/ENCFF002CQN, ENCSR000EAG/ENCFF002CPA, ENCSR000EAN/ENCFF002CQB, ENCSR000EBD/ENCFF002CQJ, ENCSR772EEN/ENCFF580QGA.

4.9. Statistical Analysis Means, standard deviations and standard error means were calculated on a minimum of three independent biological experiments. Student t tests was used when comparing two conditions only; one way ANOVA analysis followed by Dunnett test was used for comparing multiple conditions to a control condition only; one way ANOVA analysis followed by Tukey test was used for multiple pair-wise condition comparisons. In all cases p-values were calculated as follows: * p < 0.050; ** p < 0.010 and *** p < 0.001. Int. J. Mol. Sci. 2020, 21, 3000 16 of 19

4.10. Drosophila Melanogaster Fly culture and husbandry were performed according to standard protocols. The tgo allele P{EPgy2}tgoEY03802 was used in all experiments. represents a P-Element insertion within the first exon of the tgo gene. For the experiments, we used heterozygous tgoEY03802 flies, which exhibit reduced levels of tgo mRNA. white1118 flies were used as control. For any details about fruit flies used in this study refer to the content on FlyBase website (http://flybase.org/). For RNA extraction and RT-PCR analysis, ~70 adult flies were placed in a 1.5 mL tube, frozen in liquid nitrogen and mechanically disrupted by vortexing. Drosophila heads were collected and homogenised with 250 µL of appropriate Lysis Buffer. RNA extraction was performed using PeqGOLD total RNA kit (Peqlab, Bishop’s Waltham, UK) or PureLink RNA Mini Kit (Life Technologies/ThermoFisher, Paisley, UK) according to manufacturer’s instructions. RNA was converted to cDNA using Quantitect Reverse Transcription Kit (Qiagen, Manchester, UK) or First Strand cDNA Synthesis kit (ThermoFisher, Paisley, UK). For quantitative PCR, Brilliant II Sybr green kit (Stratagene/Agilent, Stockport, UK), and recommended MX3005P 96-well skirted plates were used to analyse samples on the Mx3005P qPCR platform (Stratagene/Agilent, Stockport, UK). Alternatively, PerfeCTa Sybr Green FastMix (Quanta Bio) with ROX dye added in 1:250 ratio, and recommended Microamp Optical 96-well reaction plates were used to analyse samples on the QuantStudio 6 Flex qPCR platform (Applied Biosystem/ThermoFisher, Paisley, UK). dActin was used as normalising gene in all experiments, and RT-PCR results were analysed by the ∆∆Ct method. Primers used for qPCR are: dActin, For: 50-GCGTTTTGTACAATTC GTCAGCAACC-30, Rev: 50-GCACGCGAAACTGCAGCCAA-30. dTraf6, For: 50-TCAGAACCATC ACTATGCACCC-30, Rev: 50-CAAATGGCGCACTCGTATCG-3. Tgo, For: 50-CGGCTGCTCA TACGCCCGAG-30, Rev: 50-GGCCAGCATGTGCGTCTGGT-30. Dorsal, For: 50-TGTTCAA ATCGCGGGCGTCGA-30, Rev: 50-TCGGACACCTTCGAGCTCCAGAA-30. Drosomycin, For: 50-GTTCGCCCTCTTCGCTGTCCTGA-30, Rev: 50-CCTCCTCCTTGCACACACGACG-30. Attacin A, For: 50-AGGTTCCTTAACCTCCAATC-30, Rev: 50-CATGACCAGCATTGTTGTAG-30. Protein concentration was determined using Pierce BCA Protein Assay Kit (ThermoFisher, Paisley, UK) according to the manufacturer’s instructions. 20 µg of protein were prepared in 2 SDS loading × buffer and processed for Western Blot analysis as described in the previous section. Primary antibody against Tango was purchased from DSHB (Iowa City, IA, USA). To measure survival rate upon bacterial infection, adult flies were pricked with a Tungsten needle inoculated with a concentrated bacterial solution (Serratia marcescens (OD600 < 0.20)) and incubated for 24 h (Survival analysis) at 25 ◦C. In general, means, standard deviations and standard error means were calculated prior performing Student t tests on a minimum of three independent biological experiments and calculating p-values. * p < 0.050; ** p < 0.010 and *** p < 0.001. For Drosophila melanogaster survival studies, p-values were obtained using Log-Rank statistical analysis. * p < 0.050; ** p < 0.010 and *** p < 0.001.

Supplementary Materials: The following are available online at http://www.mdpi.com/1422-0067/21/8/3000/s1. Author Contributions: Conceptualisation, L.D., H.A.M., and S.R.; investigation, L.D., D.S., M.B., S.R.; data curation, L.D., D.S., M.B., S.R.; writing—original draft preparation, L.D., S.R.; writing—review and editing, L.D., D.S., M.B., H.A.M., S.R.; supervision, S.R.; project administration, S.R.; funding acquisition, S.R., H.A.M. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Wellcome Trust, grant number 105307/Z/14/Z (L.D.; S.R.) 206293/Z/17/Z (S.R.); CRUK grant C99667/A12918 (S.R.), Medical Research Council grant KO1853/1 (H.A.M.), and the University of Liverpool (D.S.; M.B.; S.R.). The APC was funded by the University of Liverpool. Acknowledgments: We would like to thank Collin Duckett, Philip Cohen, and Ron Hay for providing us with reagents. Int. J. Mol. Sci. 2020, 21, 3000 17 of 19

Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.

Abbreviations

NF-κB Nuclear Factor kappa-light-chain-enhancer of activated B cells IKK Inhibitor of κB Kinase TNF Tumour Necrosis Factor IL-1β Interleukin-1β TRADD TNF Receptor Type 1-Associated Death Domain MYD88 Myeloid Differentiation Primary Response gene 88 TRAFs TNF Receptor Associated Factors HIF Hypoxia Inducible Factor ARNT Aryl Hydrocarbon Receptor Nuclear Translocator PHDs Prolyl-Hydroxylases VHL von Hippel Lindau PARP Poly-ADP Ribose Polymerase TAK Transforming Growth Factor (TGF)- Kinase TAB TAK Binding Protein NIK NF-κB Inducing Kinase IAP1 Inhibitor of Apoptosis 1 RIP1 Receptor Interacting Serine/Threonine Kinase 1 AHR Aryl Hydrocarbon Receptor TSS Transcription Start Site ChIP Chromatin ImmunoPrecipitation

References

1. Zhang, Q.; Lenardo, M.J.; Baltimore, D. 30 years of NF-kappab: A blossoming of relevance to human pathobiology. Cell 2017, 168, 37–57. [CrossRef] 2. Perkins, N.D. Integrating cell-signalling pathways with NF-kappab and IKK function. Nat. Rev. Mol. Cell Biol. 2007, 8, 49–62. [CrossRef] 3. Shi, J.H.; Sun, S.C. Tumor necrosis factor receptor-associated factor regulation of nuclear factor kappab and mitogen-activated protein kinase pathways. Front. Immunol. 2018, 9, 1849. [CrossRef][PubMed] 4. Biddlestone, J.; Bandarra, D.; Rocha, S. The role of hypoxia in inflammatory disease (Review). Int. J. Mol. Med. 2015, 35, 859–869. [CrossRef][PubMed] 5. D’Ignazio, L.; Batie, M.; Rocha, S. Hypoxia and inflammation in cancer, focus on HIF and NF-kappab. Biomedicines 2017, 5, 21. [CrossRef] 6. Rocha, S. Gene regulation under low oxygen: Holding your breath for transcription. Trends Biochem. Sci. 2007, 32, 389–397. [CrossRef] 7. Fandrey, J.; Gorr, T.A.; Gassmann, M. Regulating cellular oxygen sensing by hydroxylation. Cardiovasc. Res. 2006, 71, 642–651. [CrossRef] 8. Taylor, C.T.; Colgan, S.P. Regulation of immunity and inflammation by hypoxia in immunological niches. Nat. Rev. Immunol. 2017, 17, 774–785. [CrossRef] 9. Frede, S.; Stockmann, C.; Freitag, P.; Fandrey, J. Bacterial lipopolysaccharide induces HIF-1 activation in human monocytes via p44/42 mapk and nf-kappab. Biochem. J. 2006, 396, 517–527. [CrossRef][PubMed] 10. Bonello, S.; Zahringer, C.; BelAiba, R.S.; Djordjevic, T.; Hess, J.; Michiels, C.; Kietzmann, T.; Gorlach, A. Reactive oxygen species activate the HIF-1alpha promoter via a functional NFkappab site. Arter. Thromb. Vasc. Biol. 2007, 27, 755–761. [CrossRef] 11. Van Uden, P.; Kenneth, N.S.; Rocha, S. Regulation of hypoxia-inducible factor-1alpha by NF-kappab. Biochem. J. 2008, 412, 477–484. [CrossRef][PubMed] 12. Van Uden, P.; Kenneth, N.S.; Webster, R.; Muller, H.A.; Mudie, S.; Rocha, S. Evolutionary conserved regulation of HIF-1beta by NF-kappab. PLoS Genet. 2011, 7, e1001285. [CrossRef][PubMed] Int. J. Mol. Sci. 2020, 21, 3000 18 of 19

13. Bandarra, D.; Biddlestone, J.; Mudie, S.; Muller, H.A.; Rocha, S. HIF-1alpha restricts NF-kappab-dependent gene expression to control innate immunity signals. Dis. Models Mech. 2015, 8, 169–181. [CrossRef] 14. D’Ignazio, L.; Batie, M.; Rocha, S. Tnfsf14/light, a non-canonical NF-kappab stimulus, induces the HIF pathway. Cells 2018, 7, 102. [CrossRef] 15. Wright, C.W.; Duckett, C.S. The aryl hydrocarbon nuclear translocator alters CD30-mediated NF-kappab-dependent transcription. Science 2009, 323, 251–255. [CrossRef] 16. Yang, W.; Xia, Y.; Cao, Y.; Zheng, Y.; Bu, W.; Zhang, L.; You, M.J.; Koh, M.Y.; Cote, G.; Aldape, K.; et al. EGFR-induced and pkcepsilon monoubiquitylation-dependent NF-kappab activation upregulates pkm2 expression and promotes tumorigenesis. Mol. Cell. 2012, 48, 771–784. [CrossRef] 17. Strowitzki, M.J.; Cummins, E.P.; Taylor, C.T. Protein hydroxylation by hypoxia-inducible factor (HIF) hydroxylases: Unique or ubiquitous? Cells 2019, 8, 384. [CrossRef] 18. Morais, C.; Gobe, G.; Johnson, D.W.; Healy, H. The emerging role of nuclear factor kappa b in renal cell carcinoma. Int. J. Biochem. Cell Biol. 2011, 43, 1537–1549. [CrossRef] 19. Kozak, K.R.; Abbott, B.; Hankinson, O. Arnt-deficient mice and placental differentiation. Dev. Biol. 1997, 191, 297–305. [CrossRef][PubMed] 20. Rocha, S.; Martin, A.M.; Meek, D.W.; Perkins, N.D. P53 represses cyclin d1 transcription through down regulation of bcl-3 and inducing increased association of the p52 NF-kappab subunit with histone deacetylase 1. Mol. Cell. Biol. 2003, 23, 4713–4727. [CrossRef] 21. Schumm, K.; Rocha, S.; Caamano, J.; Perkins, N.D. Regulation of p53 tumour suppressor target gene expression by the p52 NF-kappab subunit. EMBO J. 2006, 25, 4820–4832. [CrossRef][PubMed] 22. Perkins, N.D. The diverse and complex roles of NF-kappab subunits in cancer. Nat. Rev. Cancer 2012, 12, 121–132. [CrossRef][PubMed] 23. Annibaldi, A.; Meier, P. Checkpoints in TNF-induced cell death: Implications in inflammation and cancer. Trends Mol. Med. 2018, 24, 49–65. [CrossRef][PubMed] 24. Hetru, C.; Hoffmann, J.A. NF-kappab in the immune response of drosophila. Cold Spring Harb. Perspect. Biol. 2009, 1, a000232. [CrossRef] 25. Aashaq, S.; Batool, A.; Andrabi, K.I. Tak1 mediates convergence of cellular signals for death and survival. Apoptosis 2019, 24, 3–20. [CrossRef] 26. Zhang, J.; Wang, C.; Chen, X.; Takada, M.; Fan, C.; Zheng, X.; Wen, H.; Liu, Y.; Wang, C.; Pestell, R.G.; et al. Egln2 associates with the nrf1-pgc1alpha complex and controls mitochondrial function in breast cancer. EMBO J. 2015, 34, 2953–2970. [CrossRef] 27. Wu, D.; Rastinejad, F. Structural characterization of mammalian BHLH-pas transcription factors. Curr. Opin. Struct. Biol. 2017, 43, 1–9. [CrossRef][PubMed] 28. Walsh, M.C.; Lee, J.; Choi, Y. Tumor necrosis factor receptor-associated factor 6 (TRAF6) regulation of development, function, and homeostasis of the immune system. Immunol. Rev. 2015, 266, 72–92. [CrossRef] 29. Cohen, P.; Strickson, S. The role of hybrid ubiquitin chains in the MYD88 and other innate immune signalling pathways. Cell Death Differ. 2017, 24, 1153–1159. [CrossRef][PubMed] 30. Naito, A.; Azuma, S.; Tanaka, S.; Miyazaki, T.; Takaki, S.; Takatsu, K.; Nakao, K.; Nakamura, K.; Katsuki, M.; Yamamoto, T.; et al. Severe osteopetrosis, defective interleukin-1 signalling and lymph node organogenesis in traf6-deficient mice. Genes Cells 1999, 4, 353–362. [CrossRef] 31. Fang, J.; Muto, T.; Kleppe, M.; Bolanos, L.C.; Hueneman, K.M.; Walker, C.S.; Sampson, L.; Wellendorf, A.M.; Chetal, K.; Choi, K.; et al. Traf6 mediates basal activation of NF-kappab necessary for hematopoietic stem cell homeostasis. Cell Rep. 2018, 22, 1250–1262. [CrossRef][PubMed] 32. Matsumoto, R.; Dainichi, T.; Tsuchiya, S.; Nomura, T.; Kitoh, A.; Hayden, M.S.; Ishii, K.J.; Tanaka, M.; Honda, T.; Egawa, G.; et al. Epithelial traf6 drives IL-17-mediated psoriatic inflammation. JCI Insight 2018, 3, e121175. [CrossRef][PubMed] 33. Kanaya, T.; Sakakibara, S.; Jinnohara, T.; Hachisuka, M.; Tachibana, N.; Hidano, S.; Kobayashi, T.; Kimura, S.; Iwanaga, T.; Nakagawa, T.; et al. Development of intestinal m cells and follicle-associated epithelium is regulated by traf6-mediated NF-kappab signaling. J. Exp. Med. 2018, 215, 501–519. [CrossRef] 34. So, T.; Nagashima, H.; Ishii, N. TNF receptor-associated factor (TRAF) signaling network in CD4(+) T-lymphocytes. Tohoku J. Exp. Med. 2015, 236, 139–154. [CrossRef][PubMed] Int. J. Mol. Sci. 2020, 21, 3000 19 of 19

35. Chen, L.; Li, Y.C.; Wu, L.; Yu, G.T.; Zhang, W.F.; Huang, C.F.; Sun, Z.J. Traf6 regulates tumour metastasis through EMT and CSC phenotypes in head and neck squamous cell carcinoma. J. Cell. Mol. Med. 2018, 22, 1337–1349. [CrossRef] 36. Rezaeian, A.H.; Li, C.F.; Wu, C.Y.; Zhang, X.; Delacerda, J.; You, M.J.; Han, F.; Cai, Z.; Jeong, Y.S.; Jin, G.; et al. A hypoxia-responsive traf6-atm-h2ax signalling axis promotes HIF1alpha activation, tumorigenesis and metastasis. Nat. Cell Biol. 2017, 19, 38–51. [CrossRef] 37. Lee, N.R.; Ban, J.; Lee, N.J.; Yi, C.M.; Choi, J.Y.; Kim, H.; Lee, J.K.; Seong, J.; Cho, N.H.; Jung, J.U.; et al. Activation of RIG-I-mediated antiviral signaling triggers autophagy through the mavs-traf6-beclin-1 signaling axis. Front. Immunol. 2018, 9, 2096. [CrossRef] 38. Qi, Y.; Pradipta, A.R.; Li, M.; Zhao, X.; Lu, L.; Fu, X.; Wei, J.; Hsung, R.P.; Tanaka, K.; Zhou, L. Cinchonine induces apoptosis of Hela and A549 cells through targeting traf6. J. Exp. Clin. Cancer Res. 2017, 36, 35. [CrossRef] 39. Sun, H.; Li, X.; Fan, L.; Wu, G.; Li, M.; Fang, J. Traf6 is upregulated in colon cancer and promotes proliferation of colon cancer cells. Int. J. Biochem. Cell Biol. 2014, 53, 195–201. [CrossRef] 40. Becnel, L.B.; Ochsner, S.A.; Darlington, Y.F.; McOwiti, A.; Kankanamge, W.H.; Dehart, M.; Naumov, A.; McKenna, N.J. Discovering relationships between nuclear receptor signaling pathways, genes, and tissues in transcriptomine. Sci. Signal. 2017, 10, eaah6275. [CrossRef] 41. Ochsner, S.A.; Abraham, D.; Martin, K.; Ding, W.; McOwiti, A.; Kankanamge, W.; Wang, Z.; Andreano, K.; Hamilton, R.A.; Chen, Y.; et al. The signaling pathways project, an integrated ‘omics knowledgebase for mammalian cellular signaling pathways. Sci. Data 2019, 6, 252. [CrossRef] 42. Sloan, C.A.; Chan, E.T.; Davidson, J.M.; Malladi, V.S.; Strattan, J.S.; Hitz, B.C.; Gabdank, I.; Narayanan, A.K.; Ho, M.; Lee, B.T.; et al. Encode data at the encode portal. Nucleic Acids Res. 2016, 44, D726–D732. [CrossRef] [PubMed] 43. Consortium, E.P. An integrated encyclopedia of DNA elements in the human genome. Nature 2012, 489, 57–74. [CrossRef][PubMed] 44. Tang, Z.; Kang, B.; Li, C.; Chen, T.; Zhang, Z. Gepia2: An enhanced web server for large-scale expression profiling and interactive analysis. Nucleic Acids Res. 2019, 47, W556–W560. [CrossRef][PubMed] 45. Shen, B.; Liu, H.; Skolnik, E.Y.; Manley, J.L. Physical and functional interactions between drosophila TRAF2 and pelle kinase contribute to dorsal activation. Proc. Natl. Acad. Sci. USA 2001, 98, 8596–8601. [CrossRef] [PubMed] 46. Tang, H.W.; Liao, H.M.; Peng, W.H.; Lin, H.R.; Chen, C.H.; Chen, G.C. Atg9 interacts with dtraf2/traf6 to regulate oxidative stress-induced JNK activation and autophagy induction. Dev. Cell. 2013, 27, 489–503. [CrossRef] 47. Mishra, A.K.; Sachan, N.; Mutsuddi, M.; Mukherjee, A. Traf6 is a novel regulator of notch signaling in drosophila melanogaster. Cell Signal. 2014, 26, 3016–3026. [CrossRef] 48. Rodriguez, M.S.; Thompson, J.; Hay, R.T.; Dargemont, C. Nuclear retention of ikappabalpha protects it from signal-induced degradation and inhibits nuclear factor kappab transcriptional activation. J. Biol. Chem. 1999, 274, 9108–9115. [CrossRef] 49. Kenneth, N.S.; Mudie, S.; Naron, S.; Rocha, S. Tfr1 interacts with the IKK complex and is involved in IKK-NF-kappab signalling. Biochem. J. 2013, 449, 275–284. [CrossRef] 50. Hahne, F.; Ivanek, R. Visualizing genomic data using gviz and bioconductor. Methods Mol. Biol. 2016, 1418, 335–351.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).