(2013) 32, 3254–3262 & 2013 Macmillan Publishers Limited All rights reserved 0950-9232/13 www.nature.com/onc

ORIGINAL ARTICLE The mTOR inhibitor rapamycin opposes carcinogenic changes to epidermal Akt1/PKBa isoform signaling

K Sully1, O Akinduro1, MP Philpott1, AS Naeem2, CA Harwood1, VE Reeve3, RF O’Shaughnessy2 and C Byrne1

Epidermal squamous cell carcinoma (SCC) is the most aggressive non-melanoma skin cancer and is dramatically increased in patients undergoing immunosuppression following solid organ transplantation, contributing substantially to morbidity and mortality. Recent clinical studies show that use of the mammalian target of rapamycin (mTOR) inhibitor rapamycin as a post- transplantation immunosuppressive significantly reduces SCC occurrence compared with other immunosuppressives, though the mechanism is not fully understood. We show that rapamycin selectively upregulates epidermal Akt1, while failing to upregulate epidermal Akt2. Rapamycin increases epidermal Akt1 phosphorylation via inhibition of the mTOR complex 1-dependent regulation of substrate-1. Epidermal Akt1 is commonly downregulated in SCC while Akt2 is upregulated. We now demonstrate similar Akt1 downregulation and Akt2 upregulation by ultraviolet (UV) radiation, the most important skin carcinogen. Hence, rapamycin’s upregulation of Akt1 signaling could potentially oppose the effects of UV radiation and/or tumor-associated changes on Akt1 signaling. We show in skin culture that rapamycin does enhance restoration of Akt1 phosphorylation in skin recovering from UV radiation, suggesting a mechanism for rapamycin’s antitumor activity in epidermis in spite of its efficient immunosuppressive properties.

Oncogene (2013) 32, 3254–3262; doi:10.1038/onc.2012.338; published online 13 August 2012 Keywords: mTOR; rapamycin; epidermis; Akt/PKB; squamous cell carcinoma; ultraviolet radiation

INTRODUCTION Rapamycin inhibits mTORC1 downstream of the PI3K/Akt path- 22 A serious problem associated with long-term immunosuppression way (reviewed in Zoncu et al. ). However, in many cells negative following solid organ transplantation is development of skin feedback loops exist between mTORC1 signaling and the receptor cancer, in particular squamous cell carcinoma (SCC) which shows accessory protein, insulin receptor substrate-1 (IRS-1) causing 23–25 dramatically increased incidence.1 Clinical studies2,3 show that the rapamycin-mediated upregulation of Akt signaling. For mammalian target of rapamycin (mTOR) inhibitor rapamycin example, mTORC1 can directly phosphorylate IRS-1 at S636/639, substantially reduces cancer incidence, with findings confirmed in interrupting the docking of IRS-1 to PI3K and downregulating the 26 mouse.4,5 The rapamycin-mediated anticancer mechanism is not PI3K/Akt pathway in response to growth factor stimulation. In fully understood but is probably partly via inhibition of tumor tumors where this mTORC1-dependent negative feedback loop is 27,28 vascularization (Guba et al.,6 reviewed by Geissler and Schlitt7). active, rapamycin treatment can result in poor clinical outcome. Epidermal SCC is associated with changes in Akt (Protein It is unclear whether mTORC1-dependent negative feedback 29 B) signaling.8,9 Akt1 and Akt2 isoforms are expressed in loops result in epidermal Akt activation. In epidermis, Akt1 and epidermis.10–12 Akt2, associated with less differentiated Akt2 signaling is spatially separated (Figure 3) and the two keratinocytes, is upregulated in SCC.13 In contrast, Akt1, active in isoforms appear to have different functions, leading to the 12,13,30 highly differentiated keratinocytes,12 is downregulated in likelihood of independent regulation. The two isoforms 13 response to high-risk human cutaneous papillomavirus early behave in opposite ways during SCC progression. Rapamycin- expression and in SCC, suggesting a tumor-suppressor role mediated upregulation of epidermal Akt signaling may, therefore, for Akt1.13 have different carcinogenic effects depending on which of the Akt is activated by translocation to the plasma membrane and two isoforms is affected. We therefore sought to find the effects of phosphorylation at two sites. Akt is recruited to the plasma rapamycin on epidermal Akt signaling. membrane by phosphatidylinositol-(3,4,5)triphosphate, produced by phosphatidylinositol 3-kinase (PI3K).14,15 Akt is then phosphory- lated in the activation loop, at T308, by phosphoinositide- RESULTS dependent kinase 1, and also in the hydrophobic pocket, Rapamycin increases Akt phosphorylation in keratinocytes at S473, predominantly by mTOR complex 2 (mTORC2), resulting Rapamycin causes a dose-dependent increase in Akt phosphory- 16,17 in full activation of Akt. Akt activates mTOR complex 1 lation at both S473 and T308 within 1 h of treatment and at (mTORC1), the second mTOR complex, via the tuberous rapamycin concentrations as low as 1 nM in rat epidermal keratino- sclerosis complex18,19 and also via the proline-rich Akt substrate cytes (REKs; Baden and Kubilus31) (Figure 1a; Supplementary 40 kDa.20,21 Figure S1a). The efficacy of rapamcyin action is shown by efficient

1Centre for Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London, UK; 2Immunbiology and Dermatology, Institute of Child Health, UCL, London, UK and 3Faculty of Veterinary Science, University of Sydney, Sydney, NSW, Australia. Correspondence: Professor C Byrne, Centre for Cutaneous Research, Blizard Institute, Barts and The London School of Medicine and Dentistry, Queen Mary University of London, London E1 2AT, UK. E-mail: [email protected] Received 18 October 2011; revised 17 May 2012; accepted 24 June 2012; published online 13 August 2012 mTOR inhibition and epidermal Akt/PKB signaling K Sully et al 3255 0 10 20 30 60 Rap (min) 24 Duration (hr) due to signaling events upstream of PI3K in keratinocytes, which is compatible with rapamycin-induced changes to IRS-1. pAkt S473 - + Rapamycin It was then asked if rapamycin-induced Akt phosphorylation pAkt S473 was due to regulation of keratinocyte IRS proteins. Rapamycin pAkt T308 increased IRS-1 protein levels from 4–6 h to at least 8 h after drug treatment (Figure 2b). IRS-2 demonstrated little change in protein Total Akt level, however, rapamycin increased the electro-mobility of both Total Akt IRS-1 and IRS-2 (marked by *). This mobility change has been pS6 S240 reported to be due to changes in IRS tyrosine and serine/ 37 pS6 S240 /244 threonine phosphorylation. Rapamycin had little effect on IRS-1 /244 and IRS-2 transcript levels (Supplementary Figure S2), supporting Actin the idea that increased IRS-1 protein is due to post-transcriptional Total S6 changes. Previous reports show that IRS-1 is degraded by the 26S proteasome, which can be inhibited by the specific proteasome 1.5 37,38 pAkt S473 pAkt T308 inhibitor lactacystin. To find the consequences of IRS-1 protein accumulation, changes in IRS-1 protein levels were investigated 1.0 following treatment with either rapamycin or lactacystin (Figure 2c). Lactacystin increased IRS-1 and also Akt phosphoryla- tion, though less markedly than rapamycin, demonstrating that 0.5 increased IRS-1 increases Akt serine phosphorylation in keratinocytes. To show that rapamycin increases the stability of the IRS-1 0.0 protein, keratinocytes were treated with rapamycin and lactacystin

pAkt relative to total Akt (au) pAkt relative 0406020 in the presence of the translational inhibitor, cycloheximide Time (min) (Figure 2d). Cycloheximide treatment gradually decreased IRS-1 Figure 1. Rapamycin increases epidermal Akt phosphorylation. levels as the protein pool was degraded. Both rapamycin and (a) Akt (S473 and T308) and S6 phosphorylation after 10 nM lactacystin maintained IRS-1 levels in the presence of cyclohex- rapamycin treatment of REKs for 0–60 min. Decrease in S6 imide, confirming that the rapamycin-mediated increase in IRS-1 phosphorylation at S240/244 demonstrates the efficacy of rapamy- was due to increased protein stability. cin activity. (b) Quantification of phosphorylation relative to total The effect of rapamycin on IRS-I phosphorylation was examined. Akt showing similar kinetics of S473 and T308 phosphorylation. IRS-1 becomes phosphorylated at multiple tyrosine residues (c) Long-term (24 h) rapamycin treatment maintains epidermal Akt within its C-terminus by activated receptor tyrosine .39 phosphorylation in REKs. These phosphorylated tyrosine sites interact with SH2 domains of downstream substrates to increase signaling. For example, Y612 and Y632 are the major docking sites for PI3K (Figure 2e; Esposito downregulation of S6 phosphorylation at serine 240/244 (S240/ et al.40). Rapamycin treatment increases tyrosine phosphorylation 244; Figures 1a and c; Supplementary Figures S1a–c, plus later of IRS-1 in keratinocytes (Figures 2e and g), demonstrating that figures). There is little difference between the kinetics rapamycin promotes IRS-1 signaling. In contrast, phosphorylation of phosphorylation at S473 and T308 (Figure 1b), suggesting a of serine residues in the C-terminus of IRS-1 can inhibit IRS-1 common activation mechanism. Rapamycin elevates Akt phos- signaling in a number of ways, including disruption of the docking phorylation for at least 24 h (Figure 1c). of PI3K (reviewed by Boura-Halfon and Zick41). IRS-1 is directly Rapamycin also induces Akt phosphorylation in human primary phosphorylated at S636/639 by mTORC1, which lies close to the keratinocytes (Supplementary Figures S1b and c), validating the tyrosine phosphorylation site Y632 (Figure 2e).26,37 Changes in use of the REKs. Primary human keratinocytes differentiate in phosphorylation at the S636/639 site could not be detected in response to calcium, eventually expressing the late terminal basal keratinocytes (not shown); however, rapamycin inhibited differentiation marker profilaggrin (Supplementary Figure S1b). IRS-1 phosphorylation at S636/639 in response to insulin Rapamycin increased Akt phosphorylation in differentiating (Figure 2g). This demonstrates that the IRS-1 S636/639 site is keratinocytes, with differentiation-dependent biphasic activation regulated by mTORC1 in keratinocytes, which probably facilitates of Akt as reported by others,32 and robustly increased Akt the phosphorylation of Y632 and contributes to an increase in the phosphorylation in highly differentiated keratinocytes (Supple- stability of IRS-1 in response to rapamycin. mentary Figure S1c). In summary, rapamycin inhibited the mTORC1-dependent negative feedback loop to IRS-1 in keratinocytes, leading to increased Akt phosphorylation. This feedback loop was activated Rapamycin-induced Akt phosphorylation is due to inhibition of in response to insulin. Rapamycin increased both the stability and the mTORC1-dependent negative feedback loop to IRS-1 in signaling of IRS-1, leading to activation of the PI3K/Akt pathway keratinocytes (Figure 2h). In some cell types, rapamycin induces Akt phosphorylation by inhibition of the mTORC1-dependent negative feedback loop to IRS-1 (reviewed by Harrington et al.33). This feedback loop can Epidermal Akt isoform responses to ultraviolet and tumorigenesis regulate both IRS-1 and IRS-2, depending on cell type.34–36 We previously used a combination of Akt isoform knockout Therefore, it was asked whether this feedback loop was present in animals and specific antibody labeling to show that epidermal keratinocytes and if induction of Akt phosphorylation by Akt1 is expressed and phosphorylated in the epidermal granular rapamycin was due to a loss of IRS regulation. layer, while Akt2 is expressed ubiquitously in all epidermal layers The PI3K inhibitor, LY294002, inhibits Akt phosphorylation at but phosphorylated as cells leave the basal layer (Figure 3a; both S473 and T308 and blocks the rapamycin-mediated increase O’Shaughnessy et al.12). Developmentally epidermal Akt1 is first in Akt S473 and T308 phosphorylation (Figure 2a). This expressed and phosphorylated in the granular layer of fetal demonstrates that rapamycin-induced Akt phosphorylation is murine epidermis as it terminally differentiates, shown by

& 2013 Macmillan Publishers Limited Oncogene (2013) 3254 – 3262 mTOR inhibition and epidermal Akt/PKB signaling K Sully et al 3256 2 4 6 8 Duration (hr) -+-+LY-294002 --++Rapamycin -+-+-+-+Rapamycin * * IRS-1 pAkt S473

* * IRS-2 pAkt T308

pAkt S473 Total Akt

Total Akt pS6 S240/244

pS6 S240/244 Total S6

Actin

Vehicle Rapamycin Lactacystin - + - Rapamycin - - + Lactacystin 02468 02468 02468 Duration (hr) IRS-1 IRS-1

pAkt S473 pAkt S473

Total Akt Total Akt

pS6 S240/244 pS6 S240/244

Total S6 Total S6

Actin

PI3K mTORC1

Y632 S636/9 1 12 113 160 263Y612 1242 N PH PTB C-terminus C

RTK

- + Rapamycin -+Rapamycin ++Insulin p-Tyr In IP: IRS-1 p-Tyr IP: IRS-1 IRS-1 IRS-1 IR IRS-1 pIRS-1 PDK1 mTORC2 IRS-1 PI3K S636/9 pAkt S473 S636/639 PIP2 PIP3 IRS-1 T308 S473 Akt Total Akt Lysate pAkt S473 Negative pS6 S240/ Lysate feedback 244 Total Akt mTORC1 Rapamycin Total S6 pS6 S240 /244

Total S6

Figure 2. Rapamycin-induced Akt phosphorylation is due to inhibition of the mTORC1-dependent negative feedback loop to IRS-1 in keratinocytes. (a) Rapamycin-induced Akt phosphorylation occurs upstream of PI3K signaling and is accompanied by changes in IRS-1 and IRS-2. Inhibition of PI3K by LY-294002 blocks rapamycin-mediated upregulation of Akt at S473 and T308. Rapamycin efficacy is shown by inhibition of S6 S240/244 phosphorylation. REK cells were given 1 h pretreatment with either vehicle or 10 mM LY-294002, followed by 1 h treatment with either vehicle or 10 nM rapamycin. (b) Rapamycin causes changes in levels and mobility of IRS-1 and 2 (shown by *). Cells were treated with vehicle or 10 nM rapamycin treatment for 2, 4, 6 and 8 h. (c) Rapamycin increases the stability of IRS-1 promoting Akt phosphorylation. Inhibition of IRS-1 degradation by the proteosome inhibitor lactacystin increases S473 Akt phosphorylation, as does rapamycin. REKs were treated with vehicle, 10 nM rapamycin or 10 mM lactacystin for 4 h. (d) Rapamycin and lactacystin can maintain IRS levels, despite depletion of the IRS pool by cycloheximide-mediated translation inhibition. REK cells were treated with 25 mg/ml cycloheximide and vehicle, 10 nM rapamycin or 10 mM lactacystin for 0, 2, 4, 6 or 8 h. (e) Rapamycin increases tyrosine phosphorylation and reduces S636/639 phosphorylation of IRS-1. Human IRS-1 protein domain structure illustrating phosphorylation sites. Phosphorylation of IRS-1 at Y612 and Y632 promotes PI3K binding. Phosphorylation of S636/9 interrupts the docking of PI3K and inhibits PI3K activity associated with IRS-1. (f) Rapamycin increases tyrosine phosphorylation of immunoprecipitated endogenous IRS-1. Lower is the immunoblot of the lysates from REK cells treated with vehicle or 10 nM rapamycin for 1 h, showing increased phospho-Akt (S473) and decreased phospho-S6 (S240/244). (g) Rapamycin decreases IRS-1 phosphorylation at S636/639. Here, lysates were prepared from serum-starved REK cells treated with either vehicle or 10 nM rapamycin for 1 h followed by 1 mg/ml insulin for 15 min. Under these conditions rapamycin also increases IRS-1 phosphotyrosine levels, increases Akt (S473) and inhibits S6 phosphorylation. (h) Summary of findings. Rapamycin increases Akt phosphorylation by inhibition of the mTORC1-dependent negative feedback loop to IRS-1. Rapamycin inhibits mTORC1, preventing phosphorylation of IRS-1 at S636/639 and increasing IRS-1 signaling and Akt phosphorylation.

Oncogene (2013) 3254 – 3262 & 2013 Macmillan Publishers Limited mTOR inhibition and epidermal Akt/PKB signaling K Sully et al 3257 Fetal skin (E18.5) Adult skin post UV Akt1 pAkt S473 pAkt1 (days) pAkt1 pAkt1 pAkt S473 0 pAkt2 pAkt S473

Akt1

Akt1 1 pAkt2 Akt2 Akt1

2 Filaggrin

H&EAkt1 Akt2 pAkt S473 3

normal skin 4

5 tumour

Figure 3. Akt1 and Akt2 expression in epidermis in response to UVB and tumorigenic changes. (a) During late fetal murine development (E18.5) Akt1 is expressed and phosphorylated (pSerAkt) in the epidermal granular layer as the epidermis undergoes terminal differentiation, shown by filaggrin expression. Akt2, expressed ubiquitously, is phosphorylated in lower suprabasal cells (pSerAkt panel) and at E18.5 is still strongly expressed. However, phosphorylation levels fall postnatally and are undetectable in adult skin (right panels). In adult skin, Akt1 is still strongly expressed and phosphorylated in the granular layer. (b) In human skin tumors, Akt1 levels are usually low but phospho-Akt (S473) levels are high, comprising phosphorylated Akt2. (c) Solar-simulated UV radiation causes Akt1 levels to fall, while phosphorylated Akt2 levels rise. Bar ¼ 50 mM. Dotted line ¼ basement membrane. induction of the terminal differentiation marker filaggrin Rapamycin selectively increases Akt1 phosphorylation 12 (Figure 3a; O’Shaughnessy et al. ). Akt1 regulates keratinocyte These findings show different responses of Akt1 and Akt2 to terminal differentiation and expression and phosphorylation is carcinogens and different changes during tumorigenesis. There- 12 maintained in adult skin. In contrast Akt2, which is ubiquitously fore, it was important to determine which of the two epidermal expressed but only phosphorylated in the lower suprabasal layers Akt isoforms was affected by rapamycin. Isoform-specific phos- 12 during late development (Figure 3a; O’Shaughnessy et al. ), is pho-Akt (S473) antibodies are unavailable so phosphorylated Akt downregulated in normal adult skin (Figure 3a, note that pSerAkt, was immunoprecipitated from vehicle and rapamycin-treated which detects epidermal Akt1 and Akt2, is exclusively associated keratinocytes and changes in Akt isoforms monitored. Rapamycin with Akt1 in the adult granular layer). increases the amount of immunoprecipitated phosphorylated Both the expression and phosphorylation of Akt isoforms are Akt1 compared with levels of immunoprecipitated phosphory- altered in SCC. Akt1 is downregulated in human SCCs (Figure 3a; lated Akt1 in vehicle-treated cells, but has little effect on relative 13 O’Shaughnessy et al. ), suggesting a tumor-suppressor role. levels of phosphorylated Akt2 and Akt3 precipitated from drug In contrast, pAkt2 is strongly upregulated (Figure 3a; O’Shaugh- and vehicle-treated cells (Figure 4a). This demonstrates that 13 nessy et al. ). Similar changes in Akt1 and Akt2 were also seen in rapamycin selectively increases Akt1 phosphorylation (Figure 4a). murine SCCs (not shown). We extended this analysis by showing The epidermal Akt isoform responsive to rapamycin was also that ultraviolet (UV) radiation, the most important skin carcinogen, investigated using HA-tagged transfected Akt1 and 2 constructs affects epidermal Akt isoform activity in murine epidermis. (Figure 4b). Only HA-immunoprecipitated Akt1 shows enhanced Two days after acute UV treatment Akt1 phosphorylation and serine 473 phosphorylation in response to rapamycin, while total levels have fallen substantially (Figure 3c), mirroring immunoprecipitated HA-Akt2 levels remain constant. Akt3 is not tumorigenic changes. In tandem, Akt2 phosphorylation levels rise expressed in epidermis though present in cultured keratino- strongly (Figure 3c), again reflecting tumorigenic changes. cytes.11,12 Since Akt1 and Akt2 are the two Akt isoforms present in Akt1 levels start to recover by 3 days followed by Akt1 epidermis, this finding confirms that rapamycin selectively phosphorylation. By 5 days, Akt1 levels and phosphorylation are increases Akt1 phosphorylation. restored and pAkt2 levels are reduced (Figure 3c). We have shown Akt phosphorylation in response to rapamycin was also similar Akt1 downregulation and pAkt2 upregulation in response monitored using a mouse fetal skin explant model, which permits to chronic UV radiation (not shown). Hence, Akt1 and Akt2 detection of spatially separated Akt1 and 2 isoforms (Figures 3a respond in opposite ways to UV, reflecting the changes during and 4c; O’Shaughnessy et al.12). In this model, rapamycin tumorigenesis. treatment had no significant effect on epidermal proliferation or

& 2013 Macmillan Publishers Limited Oncogene (2013) 3254 – 3262 mTOR inhibition and epidermal Akt/PKB signaling K Sully et al 3258 HA- HA- - + Rapamycin - + Rapamycin Akt1 Akt2 pcDNA

pAkt S473 pAkt S473 -+-+ -+Rapamycin 1 2.3±0.6 pAkt S473 Akt1 Akt1 Akt1 IP: HA Akt2 IP: pAkt 1 2.1±0.7 Akt2 HA (S473) Akt2 Akt1 Lysate Akt3 Akt2 1 0.9±0.2 pS6 HA Akt3 S240/244 Lysate pS6 1 0.8±0.2 Total S6 S240/244 Total S6

Actin Actin

Rapamycin (µM) 0 1 500 pAkt1 pS6 400 S240/244 300

Akt1 200

pAkt S473 100

Mean pAkt1 S473 (au) 0 Akt2 015 Rapamycin (M) ANOVA for effect of rapamycin on pAkt1 S473: Filaggrin P < 0.005

Figure 4. Rapamycin treatment selectively increases Akt1 isoform phosphorylation. (a) Left panel: rapamycin causes a selective increase in the Akt1 component of immunoprecipitated endogenous Akt. REK cells were treated with either vehicle or 10 nM rapamycin for 2 h. Mean densitometry, þ s.d. (n ¼ 3), shown for each isoform relative to vehicle control. Right panel: immunoblot of lysates used for immunoprecipitation showing expected rapamycin-mediated increase in pAkt (S473) and fall in pS6. (b) Rapamycin causes a selective increase in Akt1 phosphorylation when HA-Akt1 and HA-Akt2 transfected constructs are immunoprecipitated with HA antibody from REK lysates. Lower panel demonstrates HA-Akt1 and HA-Akt2 levels in transfected cells treated with rapamycin or vehicle. REK cells were treated with either vehicle or 10 nM rapamycin for 2 h. (c) Left shows rapamycin selectively increasing phospho-Akt1 levels in a murine fetal explant culture model. Skin was grown for 72 h in the presence of vehicle, 1 or 5 mM rapamycin. Note that 1 mM rapamycin is the minimum dermally supplied dose which can downregulate epidermal S6 kinase activity. Bar ¼ 50 mM. Dotted line ¼ basement membrane. Right panel shows quantification of pAkt1 levels in (c). Mean Akt1 (S473) phosphorylation, n ¼ 3, with  3 fields taken per sample. Rapamycin significantly effects Akt1 phosphorylation, ANOVA Po0.005.

differentiation (Supplementary Figure S3) but it significantly and downregulates pSerAkt1 and total Akt1, with upregulation of dose dependently increased Akt1 phosphorylation levels pSerAkt2 within 24 h, modeling skin’s in-vivo response to UV (Figure 4c, quantification shown adjacently) without upregulating (Supplementary Figures S5a and b). At the experimental UVB dose Akt2 phosphorylation. Hence, rapamycin-induced Akt phosphor- (6 mJ/cm2), we find UVB-induced thymidine dimer (cyclobutane ylation was specific to epidermal Akt1. pyrimidine dimer) formation (Supplementary Figure S5c; reviewed Rapamycin treatment also increased total IRS-1 levels in the in Pfeifer and Besaratinia42). Although epidermal morphology is epidermal explants (Supplementary Figure S4, quantification unaffected (Supplementary Figure S5d) we find modest upregula- shown adjacently), consistent with rapamycin-mediated upregula- tion of activated caspase three, a marker of UV-induced apoptosis tion of keratinocyte IRS-1 (Figures 2b and c). This supports the (Supplementary Figure S5e) by 24 h post radiation. By 3 days post finding that rapamycin-induced epidermal pAkt1 is due to UV the thymidine dimers and levels of activated caspase three are inhibition of the mTORC1-dependent negative feedback loop to indistinguishable from mock-irradiated controls (data not shown), IRS-1 (Figure 2h). indicating that repair has occurred. Hence, this experimental system mimics an authentic response to a physiologically mild dose of UVB. Rapamycin’s selective increase in Akt1 phosphorylation does not We use this experimental model to show that rapamycin cannot rescue UV-mediated decrease in Akt1 phosphorylation but affects rescue the UV-induced downregulation of phosphorylated serine Akt1 phosphorylation in skin recovering from UV 473 Akt1 (Figure 5a). UV also downregulates total Akt1 levels Since rapamycin treatment improves SCC incidence and progres- (Supplementary Figures S5b and 3c), providing a partial explana- sion in patients and rapamycin can selectively increase phos- tion for rapamycin’s inability to oppose UV’s effects on pSerAkt1 phorylated Akt1 levels we speculated that rapamycin could act to levels. prevent or ameliorate the UV-mediated downregulation of Akt1 An alternative possibility is that rapamycin affects phosphoryla- phosphorylation (Figure 3c), providing a potential antitumorigenic tion and activation of Akt1 in epidermis recovering from UV. We mechanism. find that 3 days after acute UV radiation Akt1 levels begin to be We showed first that the mouse fetal explant culture responds restored and are subsequently phosphorylated during days 4 to a physiological ultraviolet B (UVB) dose appropriately and and 5 (Figure 3c). Therefore, we monitored UV-radiated explant

Oncogene (2013) 3254 – 3262 & 2013 Macmillan Publishers Limited mTOR inhibition and epidermal Akt/PKB signaling K Sully et al 3259

- UV + UV 10 - UV 24 hr + UV 24 hr 0101Rap (µM) 8 pAkt1 pAkt S473 6 * pAkt2 4 2 Akt1

Mean pAkt1 S473 (au) 0 015 Rapamycin (µM)

- UV + UV − UV 66 hr + UV 66 hr 050 5Rap (µM) 12 pAkt1 4 pAkt S473 8 pAkt2 2 Akt1 4 Akt1 Mean Akt1 (au) 0 0 Mean pAkt1 S473 (au) 05 0 5 Rapamycin (µM) Rapamycin (µM) Figure 5. Rapamycin’s ability to selectively increase Akt1 phosphorylation does not rescue UV-mediated decrease in Akt1 phosphorylation but can affect Akt1 phosphorylation in skin recovering from UV. (a) Twenty-four hours after UV radiation at 6 mJ/cm2 there is significant downregulation of pSerAkt1 shown immunohistochemically on the left with quantification by Image J on the right. Rapamycin cannot rescue UV-mediated Akt1 downregulation at 1 or 5 mM. This may be partly because UV downregulates total Akt1. (b) Sixty-six hours after acute UV radiation at 6 mJ/cm2 Akt1 levels are recovering (see quantification of Akt1 on right). Rapamycin enhances Ser473 phosphorylation of recovering Akt, shown on the left immunohistochemically with quantification by Image J on the right. Bar ¼ 50 mM. skin cultures for effects of rapamycin on Akt1 phosphorylation B3 skin include rapamycin-induced inhibition of tumor-associated ion days (66 h) after an acute UV dose when Akt1 levels are recovering channel activity,44 and rapamycin-mediated inhibition of mTOR- but not yet substantially phosphorylated. We find significant mediated promotion of CD8 þ effector versus memory cell fate.45 rapamycin-mediated increases in phosphorylation of Akt1 during Rapamycin-induced Akt phosphorylation at both S473 and recovery from UV (Figure 5b, quantification on the right). We T308, the two sites required for full activation of Akt, implying suggest that this effect of rapamycin on Akt1 activation in skin increased Akt activity.17 The similar kinetics of S473 and T308 recovering from UV radiation could potentially be antitumorigenic. phosphorylation (Figure 1b) suggest a common activation mechanism. T308 is phosphorylated by phosphoinositide- dependent kinase 1, downstream of PI3K14,16 while S473 is phosphorylated by mTORC2, also downstream of PI3K.17,46,47 DISCUSSION Therefore, the increased phosphorylation at both T308 and S473 can be explained by rapamycin increasing PI3K activity by It was shown here that rapamycin increases epidermal Akt1 inhibiting the mTORC1-dependent negative feedback loop to phosphorylation by inhibition of the mTORC1-dependent negative IRS-1. feedback loop to IRS-1 in keratinocytes and epidermis (Figure 2h). Rapamycin treatment selectively increased Akt1 rather than Rapamycin affects stabilization of IRS-1 with consequent increased Akt2 phosphorylation. Possible mechanisms for selectivity include Akt phosphorylation. Rapamycin inhibited phosphorylation of restricted Akt isoform expression and differential regulation of IRS-1 at S636/639, one of the sites of mTORC1 regulation that Akt1 and 2 by extracellular stimuli (reviewed by Gonzalez and causes downregulation of the PI3K/Akt pathway,41 while McGraw48). It is unlikely that the selectivity of rapamycin is due to promoting IRS-1 tyrosine phosphorylation that increases the restricted expression patterns of the Akt isoforms. mTORC1 downstream signaling. Akt1 isoform phosphorylation was activity, shown by phosphorylation of its substrate S6, is present in specifically increased. Akt1 is the isoform downregulated during the suprabasal layers (Figure 4c). Therefore, mTORC1 activity tumorigenesis13 and after expression of high-risk cutaneous overlaps with the expression of both Akt1 in the granular layer human papillomavirus early .13 We now report that acute and Akt2, expressed in all the epidermal layers. and chronic UV radiation also downregulate Akt1 and Another possibility is that the Akt isoforms are regulated phosphoserine Akt1, and that in skin culture rapamycin can differently. In epidermis, we predict that Akt1 is downstream of promote the phosphorylation of Akt1 as epidermis recovers from IRS-1 and the insulin receptor, and that Akt2 is regulated by UV radiation. These findings show that rapamycin is opposing the alternative stimuli. This is because rapamycin increases IRS-1 effects of the most significant epidermal carcinogen, UV radiation, signaling in response to insulin in keratinocytes (Figure 2) and on Akt1 activity and suggest another mechanism for rapamycin’s rapamycin treatment selectively phosphorylates Akt1 isoform antitumor activity in epidermis. (Figure 4). Regulation of Akt1 by IRS-1 and the insulin receptor Our finding of rapamycin-mediated increase in epidermal is supported by both IRS-1 and insulin receptor knockout mice, Akt1 phosphorylation is consistent with a reported increase in since keratinocytes from insulin receptor null mice have reduced Akt phosphorylation in skin of rapamycin-treated patients,43 IRS-1 tyrosine phosphorylation. Both models show disrupted suggesting that our findings are relevant to the decreased epidermal differentiation.49,50 In contrast, IRS-2 knockout mice do tumors in rapamycin-treated patients. not have an epidermal phenotype,49 suggesting a less important It is thought that rapamycin reduces tumors in immunosup- role for IRS-2 in epidermal homeostasis. IGF-1R null mice show pressed patients by a number of mechanisms.7 First, compared accelerated keratinocyte differentiation and reduced proliferation, with immunosuppressives such as calcineurin inhibitors, suggesting that Akt1 is not downstream from IGF-1R.51 rapamycin may lack pro-tumorigenic properties (reviewed in Recently, it was shown that an induced, suprabasal Akt activity Geissler and Schlitt7). Rapamycin inhibits tumor angiogenesis6 and associated with epidermal hyperplasticity, which probably corre- has been recently reported to alter UV-signature mutations in sponds to the UV-induced Akt2 activity reported here, is regulated skin.5 Additional antitumor mechanisms that could be relevant to by epidermal Rac1 and suprabasal keratinocyte Pac2.52 The

& 2013 Macmillan Publishers Limited Oncogene (2013) 3254 – 3262 mTOR inhibition and epidermal Akt/PKB signaling K Sully et al 3260

separate regulation of epidermal Akt1 and Akt2 activity is the differentiation and 1.2 mM CaCl2 used to induce differentiation. Inhibitors subject of ongoing investigation. (all from Sigma-Aldrich, St Louis, MO, USA) in DMSO were diluted to We show for the first time that Akt1 is downregulated and give final medium DMSO concentrations of 0.1%, and inhibitor final pAkt2 upregulated within 1 day of UV radiation (solar-simulated concentrations: cycloheximide, 25 mg/ml; insulin, 1 mg/ml; lactacystin, radiation comprising environmentally relevant UVA and UVB 10 mM; LY-294002, 10 mM; rapamycin, 10 nM. doses) with recovery over 5 days. Akt1 is expressed in granular keratinocytes synthesizing the stratum corneum barrier, and Akt1 Fetal skin explant culture null mice have mild stratum corneum defects.12 Double Akt1/Akt2 Skin (epidermis and dermis) was removed from the dorsal torso region of knockout produces lethal skin barrier loss,53 showing Akt2 can E15.5 fetal CD1 mice, rinsed in culture medium, then cultured dermal side compensate for Akt1 activity. The transient downregulation of down on Millipore filters at the air–liquid interface for 72 h. Culture epidermal Akt1 within days of an acute UV dose correlates with medium was Williams Medium E supplemented with 10 mg/ml insulin, the well-documented transient decrease in skin barrier within a 10 ng/ml hydrocortisone, 2 mM glutamine, 100 IU/L penicillin, 100 mg/L few days after UV exposure (for example, Haratake et al.54 and streptomycin and 25 mg/ml amphotericin (all reagents from Invitrogen). 55 Inhibitors in DMSO were supplied dermally to produce a final DMSO Yamamoto et al. ). It is possible that the UV-induced increase in concentration of 0.1%. Rapamycin was used at final concentrations of Akt2 activity compensates for Akt1 loss, preventing severe post- 1 and 5 mM, with 1 mM the minimum dermally supplied dose which UV barrier dysfunction. downregulates epidermal S6 kinase activity. Explants were fixed in Bouin’s It has also been shown in fibroblasts, endothelial cells and solution (Sigma-Aldrich) for 2 h before processing for paraffin embedding. ovarian epithelial cells that Akt2 but not Akt1 can protect cells UVB radiation of explants was with a UV Crosslinker CX-2000 ( Ultra-Violet from UV-induced cell death,56,57 suggesting a crucial role for UV- Products, Cambridge, UK) at a dosage of 6 mJ/cm2. Explants were cultured induced epidermal Akt2 activity. UV-induced Akt2 inhibits for 48 h in media containing rapamycin or DMSO vehicle before UV apoptosis by inhibiting activation of UV-induced JNK and p38.56 radiation. Control samples were mock irradiated. Akt2 enhanced survival of mutated cells could provide mechanism for Akt2-mediated tumorigenesis. Immunohistochemistry Several groups have previously shown a relationship between Clinical material was obtained with informed written consent from patients UVB and Akt signaling in keratinocytes or skin, without specifying attending dermatology clinics at Barts and The London NHS Trust. Ethical isoform (reviewed by Cao and Wan58). Akt is activated as part of approval was granted by the East London and City local research ethics skin’s immediate response to acute UVB radiation, though activity committee. Immunohistochemistry on paraffin or frozen sections was by subsides within 30 min.59 This early response is distinct from the standard technique. Antibodies to the following proteins were used at later (1–5 days) response to acute UV reported in this work. specified dilutions: phospho-Akt S473 (9271), 1:10; Akt1 2H10 (2967), 1:250; phospho-S6 S240/244 (2215), 1:100; cleaved caspase 3 (Asp175), 1:100—all Additional groups have reported UV-induced Akt upregulation from Cell Signaling Technology, Danvers, MA, USA: Akt2 (ab13991; Abcam, within hours and days after exposure, which probably corresponds Cambridge, UK), 1:50: filaggrin (#219-234; Zymed Laboratories, San to the UV-induced Akt2 upregulation reported here (reviewed by Francisco, CA, USA), 1:400, thymidine dimer (clone KTM53; Kamiya 58 59 60 Cao and Wan , Wang et al. and Meeran et al. ). Biomedical Company, Seattle, WA, USA) 1:400. For immunofluorescent In this work, we show that B3 days after an acute UV dose Akt1 antibody detection, Alexa 488 and 546 conjugated secondary antibodies expression is recovering to be followed later by Akt1 phosphor- (Molecular Probes, Eugene, OR, USA) or FITC anti-chicken (ab6873; Abcam) ylation (Figure 3). Using a culture model that mimics many was used at 1:800 dilution. Counterstaining was with 4’,6-diamidino-2- features of skin’s response to UV we show that that Akt1 phenylindole. Thymidine dimer and some Akt1 immunohistology were by phosphorylation during UV recovery can be enhanced by immunoperoxidase staining using biotinylated anti-mouse secondary antibody (1:200; Vector Laboratories, Burlingame, CA, USA) and Vectastain rapamycin and speculate that this may be one means for avidin-biotin-peroxidase (Vector Laboratories) amplification. Images were anticancer action of rapamycin. taken with a Nikon Eclipse E600 microscope (Nikon, Tokyo, Japan) with We have as yet no direct evidence that the UV-mediated either  20 (NA 0.4) or  60 Oil Immersion (NA 1.40) objectives, using a downregulation of Akt1 (and the HPV-mediated downregulation SPOT digital camera (Diagnostic Instruments Inc., Sterling Heights, MI, USA) of Akt1)13 is tumorigenic or renders skin susceptible to further with Spot RT Software v3.0 (Diagnostic Instruments). tumors, nor do we know whether Akt1 downregulation in tumors affects tumor progression or malignancy. However, because we Western analysis, immunoprecipitation and PCR analysis and others have shown a link between Akt1 activity and barrier 12,53 Western analysis was on RIPA cultured cell lysates or from epidermal integrity we suggest that UV-mediated Akt1 downregulation lysates prepared after separation from dermis as described.13 Antibodies will compromise epidermal barrier activity. Mild epidermal barrier were used at the specified dilutions: Phospho-Akt T308 (4056), 1:500; defects have previously been linked to skin carcinogenesis phospho-AktS473 (9271), 1:500; total Akt (9272), 1:1000; Akt1 2H10 (2967), through a mechanism involving interactions between the 1:1000; Akt2 5B5 (2964), 1:500; Akt3 (4059), 1:1000; phospho-IRS-1 S636/ barrier-defective epidermis and its underlying stroma.61 In 639 (2388), 1:500; IRS-2 L1326 (3089), 1:1000; phospho-S6 S240/244 (2215), addition, the barrier is the primary defense against further 1:1000, S6 5G10 (2217), 1:1000—all from Cell Signaling Technology: carcinogen (for example, further UV) entry. filaggrin (#219-234; Zymed Laboratories), 1:1000: b-actin AC-15 (A5441; In conclusion, rapamycin treatment selectively increases epi- Sigma-Aldrich), 1:2000: HA clone HA-7 (H3663; Sigma-Aldrich), 1:5000: IRS-1 (06-248; Merck Millipore, Billerica, MA, USA), 1:1000: keratin-14 AF64 (PRB- dermal Akt1 phosphorylation, via inhibition of the mTORC-1- 155; Covance, Princeton, NJ, USA), 1:2000: phospho-Tyr PY20 (610007; BD dependent negative feedback loop (Figure 2h). Rapamycin- Transduction Labs, Lexinton, KY, USA), 1:200. Secondary antibodies were mediated changes in Akt1 signaling oppose the changes in Akt1 goat anti-mouse (ab6789; Abcam) and anti-rabbit (111-035-144; Jackson induced by the most important skin carcinogen, UV, and also Immunoresearch, Westgrove, PA, USA), both used at 1:5000. Transfection oppose Akt1 changes that occur during SCC progression, of keratinocytes with HA-tagged Akt1, Akt2 constructs and pcDNA was suggesting another potential antiskin tumor mechanism for using Lipofectamine 2000 (Invitrogen) according to the manufacturer rapamycin. instructions. Akt HA construct were Addgene Plasmid 9021: 1477 pcDNA3 flag HA Akt1 from William J Sellers, Dana-Farber Cancer Institute, Boston, MA, USA,62 and Addgene plasmid 16000: pcDNA3 Hygro HA Akt2 from Morris J Birnbaum, University of Pennsylvania, PA, USA. Immuno- MATERIALS AND METHODS precipitations of IRS1 and 2 with 2.5 mg of anti-IRS-1 or anti-IRS-2 were Cell culture incubated with 400 mg of REK protein lysate overnight followed by 2 h with REK cells31 were cultured as described.12 Neonatal human epidermal Protein-A sepharose beads (Sigma-Aldrich). Immunoprecipitation of keratinocytes (Invitrogen, Paisley, UK) were grown as advised by the phospho-S473 Akt was with 10 ml of immobilized phospho-Akt (S473) manufacturers, with low calcium levels of 0.2 mM used to suppress antibody (4070; Cell Signaling Technology) from 200 mg of REK protein

Oncogene (2013) 3254 – 3262 & 2013 Macmillan Publishers Limited mTOR inhibition and epidermal Akt/PKB signaling K Sully et al 3261 lysate. Immunoprecipitation of HA antigen was with 200 mg REK cell lysates 11 Thrash BR, Menges CW, Pierce RH, McCance DJ. AKT1 provides an essential incubated with 40 ml of monoclonal anti-HA agar monoclonal conjugated survival signal required for differentiation and stratification of primary human beads (Clone HA-7, A2095; Sigma-Aldrich) in a total volume of 300 ml for 2 h o keratinocytes. J Biol Chem 2006; 281: 12155–12162. at 4 C. RNA was isolated by standard technique and PCR performed for 25 12 O’Shaughnessy RF, Welti JC, Cooke JC, Avilion AA, Monks B, Birnbaum MJ et al. cycles (IRS-1 and IRS-2), or 20 cycles (actin). Primers were to rat IRS-1, AKT-dependent HspB1 (Hsp27) activity in epidermal differentiation. J Biol Chem CTGACATTGGAGGTGGGTCT, GCAAGGCAAGAGCTTACCAC; to IRS-2, CCACA 2007; 282: 17297–17305. CACCTGTCCT, CTTATGAAGGCTGACGAGGCCATTG; and to b-actin, GAGCGG 13 O’Shaughnessy RF, Akgul B, Storey A, Pfister H, Harwood CA, Byrne C. Cutaneous TTCCGATGCCCTGAG, GCCGGACTCATCGTACTCCTG. Quantification of human papillomaviruses down-regulate AKT1, whereas AKT2 up-regulation and densitometry data and gel images was using Image J software (National activation associates with tumors. Cancer Res 2007; 67: 8207–8215. Institutes of Health, Bethesda, MA, USA, http://imagej.nih.gov/ij/). 14 Andjelkovic M, Alessi DR, Meier R, Fernandez A, Lamb NJ, Frech M et al. Role of translocation in the activation and function of . J Biol Chem 1997; UV radiation of mouse skin 272: 31515–31524. Inbred female Skh:hr-1 (hairless) mice were obtained from the Veterinary 15 Frech M, Andjelkovic M, Ingley E, Reddy KK, Falck JR, Hemmings BA. High affinity Science breeding colony, University of Sydney. For acute UV irradiation binding of inositol phosphates and phosphoinositides to the pleckstrin homology experiments, Skh:hr-1 mice 8–10 weeks old were irradiated with a single domain of RAC/protein kinase B and their influence on kinase activity. J Biol Chem exposure of 3 Â the Minimal Edematous Dose of solar-simulated UV 1997; 272: 8474–8481. radiation from a fluorescent UVA and UVB tube source on the dorsum, and 16 Alessi DR, James SR, Downes CP, Holmes AB, Gaffney PR, Reese CB et al. mid-dorsal skin samples were taken before and at 1–5 days after. The Characterization of a 3-phosphoinositide-dependent protein kinase which Minimal Edematous Dose of solar-simulated UV radiation in the Skh:hr-1 phosphorylates and activates protein kinase Balpha. Curr Biol 1997; 7: mouse, previously established as 1.33 kJ/m2 of UVB and 21.3 kJ/m2 of 261–269. UVA63 is equivalent to a minimal erythemal dose in humans. All skin 17 Sarbassov DD, Guertin DA, Ali SM, Sabatini DM. Phosphorylation and regulation of samples were fixed for 6 h in HistoChoice (Amresco, Solon, OH, USA), Akt/PKB by the rictor-mTOR complex. Science 2005; 307: 1098–1101. processed in an automated ethanol-formalin system, wax-embedded, and 18 Inoki K, Li Y, Zhu T, Wu J, Guan KL. TSC2 is phosphorylated and inhibited by Akt sections of 4 mm were cut for immunohistochemical staining. Procedures and suppresses mTOR signalling. Nat Cell Biol 2002; 4: 648–657. were approved by the University of Sydney Animal Ethics Committee and 19 Manning BD, Tee AR, Logsdon MN, Blenis J, Cantley LC. Identification of the complied with the NSW Animal Research Act 1985. tuberous sclerosis complex-2 tumor suppressor gene product tuberin as a target of the phosphoinositide 3-kinase/akt pathway. Mol Cell 2002; 10: 151–162. 20 Kovacina KS, Park GY, Bae SS, Guzzetta AW, Schaefer E, Birnbaum MJ et al. CONFLICT OF INTEREST Identification of a proline-rich Akt substrate as a 14-3-3 binding partner. J Biol Chem 2003; 278: 10189–10194. The authors declare no conflict of interest. 21 Vander Haar E, Lee SI, Bandhakavi S, Griffin TJ, Kim DH. Insulin signalling to mTOR mediated by the Akt/PKB substrate PRAS40. Nat Cell Biol 2007; 9: 316–323. 22 Zoncu R, Efeyan A, Sabatini DM. mTOR: from growth signal integration to cancer, ACKNOWLEDGEMENTS diabetes and ageing. Nat Rev Mol Cell Biol 2011; 12: 21–35. This work was supported by the British Skin Foundation 941S, a Medical Research 23 Harrington LS, Findlay GM, Gray A, Tolkacheva T, Wigfield S, Rebholz H et al. The Council DTA studentship for OA and a Royal Society International Outgoing Short TSC1-2 tumor suppressor controls insulin-PI3K signaling via regulation of IRS Visit VO0872438. proteins. J Cell Biol 2004; 166: 213–223. 24 Shah OJ, Wang Z, Hunter T. Inappropriate activation of the TSC/Rheb/mTOR/S6K cassette induces IRS1/2 depletion, insulin resistance, and cell survival deficiencies. REFERENCES Curr Biol 2004; 14: 1650–1656. 1 Mitchell C. The importance of skin cancer prevention in organ transplant patients 25 Um SH, Frigerio F, Watanabe M, Picard F, Joaquin M. Sticker M, et al. Absence of an editorial to paper by Salgo: ‘switch to sirolimus in long-term renal transplant S6K1 protects against age- and diet-induced obesity while enhancing insulin recipients: reduced premalignancies and nonmelanoma skin cancer in a sensitivity. Nature 2004; 431: 200–205. controlled, prospective, randomized, blinded study’. Am J Transplant 2010; 10: 26 Tzatsos A, Kandror KV. Nutrients suppress phosphatidylinositol 3-kinase/Akt 1343–1344. signaling via raptor-dependent mTOR-mediated insulin receptor substrate 1 2 Alberu J, Pascoe MD, Campistol JM, Schena FP, Rial Mdel C, Polinsky M et al. Lower phosphorylation. Mol Cell Biol 2006; 26: 63–76. malignancy rates in renal allograft recipients converted to sirolimus-based, cal- 27 O’Reilly KE, Rojo F, She QB, Solit D, Mills GB, Smith D et al. mTOR inhibition cineurin inhibitor-free immunotherapy: 24-month results from the CONVERT trial. induces upstream receptor tyrosine kinase signaling and activates Akt. Cancer Res Transplantation 2011; 92: 303–310. 2006; 66: 1500–1508. 3 Salgo R, Gossmann J, Schofer H, Kachel HG, Kuck J, Geiger H et al. Switch to a 28 Zhao L, Vogt PK, Class I. PI3K in oncogenic cellular transformation. Oncogene 2008; sirolimus-based immunosuppression in long-term renal transplant recipients: 27: 5486–5496. reduced rate of (pre-)malignancies and nonmelanoma skin cancer in a 29 Lu ZH, Shvartsman MB, Lee AY, Shao JM, Murray MM, Kladney RD et al. prospective, randomized, assessor-blinded, controlled clinical trial. Am J Trans- Mammalian target of rapamycin activator RHEB is frequently overexpressed in plant 2010; 10: 1385–1393. human carcinomas and is critical and sufficient for skin epithelial carcinogenesis. 4 Wulff BC, Kusewitt DF, VanBuskirk AM, Thomas-Ahner JM, Duncan FJ, Cancer Res 2010; 70: 3287–3298. Oberyszyn TM. Sirolimus reduces the incidence and progression of UVB-induced 30 O’Shaughnessy RF, Welti JC, Sully K, Byrne C. Akt-dependent Pp2a activity is skin cancer in SKH mice even with co-administration of cyclosporine A. J Invest required for epidermal barrier formation during late embryonic development. Dermatol 2008; 128: 2467–2473. Development 2009; 136: 3423–3431. 5 de Gruijl FR, Koehl GE, Voskamp P, Strik A, Rebel HG, Gaumann A et al. Early and 31 Baden HP, Kubilus J. The growth and differentiation of cultured newborn rat late effects of the immunosuppressants rapamycin and mycophenolate mofetil keratinocytes. J Invest Dermatol 1983; 80: 124–130. on UV carcinogenesis. Int J Cancer 2010; 8:8. 32 Janes SM, Ofstad TA, Campbell DH, Watt FM, Prowse DM. Transient activation of 6 Guba M, von Breitenbuch P, Steinbauer M, Koehl G, Flegel S, Hornung M et al. FOXN1 in keratinocytes induces a transcriptional programme that promotes Rapamycin inhibits primary and metastatic tumor growth by antiangiogenesis: terminal differentiation: contrasting roles of FOXN1 and Akt. J Cell Sci 2004; involvement of vascular endothelial growth factor. Nat Med 2002; 8: 128–135. 117(Pt 18): 4157–4168. 7 Geissler EK, Schlitt HJ. Transplantation: can sirolimus prevent skin cancer in 33 Harrington LS, Findlay GM, Lamb RF. Restraining PI3K: mTOR signalling goes back transplant recipients? Nat Rev Nephrol 2010; 6: 639–641. to the membrane. Trends Biochem Sci 2005; 30: 35–42. 8 Segrelles C, Ruiz S, Perez P, Murga C, Santos M, Budunova IV et al. Functional roles 34 Berg CE, Lavan BE, Rondinone CM. Rapamycin partially prevents insulin resistance of Akt signaling in mouse skin tumorigenesis. Oncogene 2002; 21: 53–64. induced by chronic insulin treatment. Biochem Biophys Res Commun 2002; 293: 9 Mao JH, To MD, Perez-Losada J, Wu D, Del Rosario R, Balmain A et al. Mutually 1021–1027. exclusive mutations of the Pten and ras pathways in skin tumor progression. 35 Zhande R, Mitchell JJ, Wu J, Sun XJ. Molecular mechanism of insulin-induced Genes Dev 2004; 18: 1800–1805. degradation of insulin receptor substrate 1. Mol Cell Biol 2002; 22: 1016–1026. 10 Calautti E, Li J, Saoncella S, Brissette JL, Goetinck PF. Phosphoinositide 3-kinase 36 Pirola L, Bonnafous S, Johnston AM, Chaussade C, Portis F, Van Obberghen E. signaling to Akt promotes keratinocyte differentiation versus death. J Biol Chem Phosphoinositide 3-kinase-mediated reduction of insulin receptor substrate-1/2 2005; 280: 32856–32865. protein expression via different mechanisms contributes to the insulin-induced

& 2013 Macmillan Publishers Limited Oncogene (2013) 3254 – 3262 mTOR inhibition and epidermal Akt/PKB signaling K Sully et al 3262 desensitization of its signaling pathways in L6 muscle cells. J Biol Chem 2003; 278: 50 Wertheimer E, Spravchikov N, Trebicz M, Gartsbein M, Accili D, Avinoah I et al. 15641–15651. The regulation of skin proliferation and differentiation in the IR null mouse: 37 Sun XJ, Rothenberg P, Kahn CR. Backer JM, Araki E, Wilden PA, et al. Structure of implications for skin complications of diabetes. Endocrinology 2001; 142: the insulin receptor substrate IRS-1 defines a unique protein. 1234–1241. Nature 1991; 352: 73–77. 51 Sadagurski M, Yakar S, Weingarten G, Holzenberger M, Rhodes CJ, Breitkreutz D 38 Haruta T, Uno T, Kawahara J, Takano A, Egawa K, Sharma PM et al. A rapamycin- et al. Insulin-like growth factor 1 receptor signaling regulates skin development sensitive pathway down-regulates insulin signaling via phosphorylation and and inhibits skin keratinocyte differentiation. Mol Cell Biol 2006; 26: 2675–2687. proteasomal degradation of insulin receptor substrate-1. Mol Endocrinol 2000; 14: 52 Wang Z, Pedersen E, Basse A, Lefever T, Peyrollier K, Kapoor S et al. Rac1 is crucial 783–794. for Ras-dependent skin tumor formation by controlling Pak1-Mek-Erk hyper- 39 Gual P, Le Marchand-Brustel Y, Tanti JF. Positive and negative regulation of insulin activation and hyperproliferation in vivo. Oncogene 2010; 29: 3362–3373. signaling through IRS-1 phosphorylation. Biochimie 2005; 87: 99–109. 53 Chen WS, Xu PZ, Gottlob K, Chen ML, Sokol K, Shiyanova T et al. Growth retar- 40 Esposito DL, Li Y, Cama A, Quon MJ. Tyr(612) and Tyr(632) in human insulin dation and increased apoptosis in mice with homozygous disruption of the Akt1 receptor substrate-1 are important for full activation of insulin-stimulated phos- gene. Genes Dev 2001; 15: 2203–2208. phatidylinositol 3-kinase activity and translocation of GLUT4 in adipose cells. 54 Haratake A, Uchida Y, Schmuth M, Tanno O, Yasuda R, Epstein JH et al. UVB- Endocrinology 2001; 142: 2833–2840. induced alterations in permeability barrier function: roles for epidermal hyperpro- 41 Boura-Halfon S, Zick Y. Phosphorylation of IRS proteins, insulin action, and insulin liferation and thymocyte-mediated response. J Invest Dermatol 1997; 108: 769–775. 55 Yamamoto A, Serizawa S, Ito M, Sato Y. Stratum corneum lipid abnormalities in resistance. Am J Physiol Endocrinol Metab 2009; 296: 26. atopic dermatitis. Arch Dermatol Res 1991; 283: 219–223. 42 Pfeifer GP, Besaratinia A. UV wavelength-dependent DNA damage and 56 Kim MA, Kim HJ, Jee HJ, Kim AJ, Bae YS, Bae SS et al. Akt2, but not Akt1, is required human non-melanoma and melanoma skin cancer. Photochem Photobiol Sci 2012; for cell survival by inhibiting activation of JNK and p38 after UV irradiation. 11: 90–97. Oncogene 2009; 19:19. 43 Tabernero J, Rojo F, Calvo E, Burris H, Judson I, Hazell K et al. Dose- and schedule- 57 Yuan ZQ, Feldman RI, Sun M, Olashaw NE, Coppola D, Sussman GE et al. Inhibition dependent inhibition of the mammalian target of rapamycin pathway with of JNK by cellular stress- and tumor necrosis factor alpha-induced AKT2 through everolimus: a phase I tumor pharmacodynamic study in patients with advanced activation of the NF kappa B pathway in human epithelial Cells. J Biol Chem 2002; solid tumors. J Clin Oncol 2008; 26: 1603–1610. 277: 29973–29982. 44 Koehl GE, Spitzner M, Ousingsawat J, Schreiber R, Geissler EK, Kunzelmann K. 58 Cao C, Wan Y. Parameters of protection against ultraviolet radiation-induced skin Rapamycin inhibits oncogenic intestinal ion channels and neoplasia in cell damage. J Cell Physiol 2009; 220: 277–284. APC(Min/ þ ) mice. Oncogene 2010; 29: 1553–1560. 59 Wang HQ, Quan T, He T, Franke TF, Voorhees JJ, Fisher GJ. Epidermal growth 45 Rao RR, Li Q, Odunsi K, Shrikant PA. The mTOR kinase determines effector versus factor receptor-dependent, NF-kappaB-independent activation of the phospha- memory CD8 þ T cell fate by regulating the expression of transcription factors tidylinositol 3-kinase/Akt pathway inhibits ultraviolet irradiation-induced cas- T-bet and Eomesodermin. Immunity 2010; 32: 67–78. pases-3, -8, and -9 in human keratinocytes. J Biol Chem 2003; 278: 45737–45745. 46 Gan X, Wang J, Su B, Wu D. Evidence for direct activation of mTORC2 kinase 60 Meeran SM, Katiyar N, Singh T, Katiyar SK. Loss of endogenous interleukin-12 activity by phosphatidylinositol 3,4,5-trisphosphate. J Biol Chem 2011; 286: activates survival signals in ultraviolet-exposed mouse skin and skin tumors. 10998–11002. Neoplasia 2009; 11: 846–855. 47 Tato I, Bartrons R, Ventura F, Rosa JL. Amino acids activate mammalian target of 61 Demehri S, Turkoz A, Kopan R. Epidermal Notch1 loss promotes skin tumor- rapamycin complex 2 (mTORC2) via PI3K/Akt signaling. J Biol Chem 2011; 286: igenesis by impacting the stromal microenvironment. Cancer Cell 2009; 16: 55–66. 6128–6142. 62 Hsieh AC, Bo R, Manola J, Vazquez F, Bare O, Khvorova A et al. A library of siRNA 48 Gonzalez E, McGraw TE. The Akt kinases: isoform specificity in metabolism and duplexes targeting the phosphoinositide 3-kinase pathway: determinants of gene cancer. Cell Cycle 2009; 8: 2502–2508. silencing for use in cell-based screens. Nucl Acids Res 2004; 32: 893–901. 49 Sadagurski M, Nofech-Mozes S, Weingarten G, White MF, Kadowaki T, 63 Reeve VE, Domanski D, Slater M. Radiation sources providing increased UVA/UVB Wertheimer E. Insulin receptor substrate 1 (IRS-1) plays a unique role in normal ratios induce photoprotection dependent on the UVA dose in hairless mice. epidermal physiology. J Cell Physiol 2007; 213: 519–527. Photochem Photobiol 2006; 82: 406–411.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene (2013) 3254 – 3262 & 2013 Macmillan Publishers Limited