HM Medical Clinic

 

Can-14-3362 227.238

Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Microenvironment and Immunology Control of PD-L1 Expression by OncogenicActivation of the AKT–mTOR Pathway inNon–Small Cell Lung CancerKristin J. Lastwika1,2, Willie Wilson III3, Qing Kay Li4, Jeffrey Norris1, Haiying Xu5,Sharon R. Ghazarian6, Hiroshi Kitagawa1, Shigeru Kawabata1, Janis M. Taube5,Sheng Yao7, Linda N. Liu7, Joell J. Gills1, and Phillip A. Dennis1 Alterations in EGFR, KRAS, and ALK are oncogenic drivers in squamous cell carcinomas, membranous expression of PD-L1 lung cancer, but how oncogenic signaling influences immunity in was significantly associated with mTOR activation. These data the tumor microenvironment is just beginning to be understood.
suggest that oncogenic activation of the AKT–mTOR pathway Immunosuppression likely contributes to lung cancer, because promotes immune escape by driving expression of PD-L1, which drugs that inhibit immune checkpoints like PD-1 and PD-L1 have was confirmed in syngeneic and genetically engineered mouse clinical benefit. Here, we show that activation of the AKT–mTOR models of lung cancer where an mTOR inhibitor combined with a pathway tightly regulates PD-L1 expression in vitro and in vivo.
PD-1 antibody decreased tumor growth, increased tumor-infil- Both oncogenic and IFNg-mediated induction of PD-L1 was trating T cells, and decreased regulatory T cells. Cancer Res; 76(2); dependent on mTOR. In human lung adenocarcinomas and 227–38. 2015 AACR.
cells. Mice that lacked FoxP3 cells developed fewer lung tumorsthan mice with mutant KRAS alone.
Despite the development of targeted therapies, lung cancer Although multiple mechanisms can contribute to immune remains the leading cause of cancer-related death worldwide suppression in the tumor microenvironment, programmed death (1). Most of the oncogenic drivers in non–small cell lung cancer ligand 1 (PD-L1 and B7-H1), an inhibitory member of the B7 (NSCLC), such as EGFR or KRAS, activate the PI3K–AKT–mTOR family, plays a central role in many cancer types (6). This cell pathway, which increases cell proliferation, metabolism, and surface protein is normally found on immune cells and in survival. Activation of this pathway is a critical event during lung immune privileged tissues, but its expression is upregulated in tumorigenesis. Previously, we showed that genetic deletion of many epithelial tumors, including lung cancer (7). PD-L1 binds to AKT1 or inhibitors of mTOR such as rapamycin or metformin either PD-1 or CD80 receptors on activated immune cells to prevent KRAS-driven lung tumorigenesis (2–4). We also demon- inhibit their activation and effector responses (8). The interaction strated a relationship between AKT/mTOR signaling and immune of PD-L1 and PD-1 induces differentiation of na€ve CD4þ T cells suppression, because inhibition of tumorigenesis by rapamycin into Tregs and maintains Treg-suppressive functions. PD-L1 can was associated with reduced influx of lung associated FoxP3þ also act as a receptor by sending reverse signals to limit tumor cell regulatory T cells (Tregs) into the tumors (5). This was confirmed apoptosis. The importance of PD-L1 and PD-1 in lung cancer is by creating mice that harbored mutant KRAS but lacked FoxP3 reflected by the antitumor activity observed using PD-1– or PD-L1–blocking antibodies as single agents in heavily pretreatedNSCLC patients (9, 10). Clinical responses were sometimes sustained over many months, suggesting recovered ability of Department of Oncology, Johns Hopkins University, Baltimore, Mary- land. 2The George Washington University, Institute for Biomedical immune effectors to control tumor growth (11). This clinical Sciences, Washington, DC. 3Cancer Biology and Genetics Branch, benefit supports efforts to study the mechanisms that regulate Center for Cancer Research, National Institutes of Health, Bethesda, tumor PD-L1 expression and therapeutic interventions to decrease Maryland. 4Department of Pathology, Johns Hopkins University, Bal- PD-L1 levels.
timore, Maryland. 5Department of Dermatology, Johns Hopkins Uni-versity, Baltimore, Maryland. 6Biostatistics, Epidemiology and Data Tumors can express PD-L1 either constitutively or through Management Core, Johns Hopkins University, Baltimore, Maryland.
flammatory cytokines, especially members of the Amplimmune, Inc., Gaithersburg, Maryland.
interferon family. Cytokine-driven PD-L1 expression is indicative Note: Supplementary data for this article are available at Cancer Research of an ongoing immune response in the tumor microenvironment, whereas intrinsic PD-L1 expression does not depend on the Corresponding Author: Phillip A. Dennis, Johns Hopkins University School of presence of tumor-infiltrating lymphocytes.
Medicine, 4940 Eastern Avenue, 301 Building/Suite 4500 Baltimore, MD 21224.
Multiple mechanisms can contribute to intrinsic tumor PD-L1 Phone: 410-550-9250; Fax: 410-550-5445; E-mail: [email protected] expression. Expressions of PD-L1 and PD-L2 (another ligand for PD-1) are increased in Hodgkin's disease and mediastinal large B- 2015 American Association for Cancer Research.
cell lymphoma through chromosomal amplification (12). T-cell Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Lastwika et al.
lymphomas carrying NPM–ALK fusions induce PD-L1 expression In vivo treatments through STAT3 activation (13). PTEN loss or PIK3CA mutations For the transgenic KRASLA2 mouse model treatment began in glioma, breast, and prostate cancers have been shown to at weaning and lasted 4 weeks. One hundred and fifty micrograms activate the AKT–mTOR pathway and subsequently increase of anti–PD-1 blocking antibody (Amplimmune) was given on PD-L1 expression (14, 15). A correlation between activating the first treatment day in combination with rapamycin. The mutations in EGFR and increased immunosuppression markers, control vehicle was given on treatment day 1 in combination including PD-L1 and PD-1, was established (16). Recently, a with 150 mg IgG (Rockland) . A previously optimized rapamycin mouse model of lung squamous carcinoma demonstrated high dosing schedule was used to obtain trough levels that are readily PD-L1 expression in tumor-promoting cells with loss of LKB1 and tolerated in humans (25). The control and anti–PD-1 antibody PTEN (17). In NSCLC patients, the relationship of oncogenic were given by i.p. injection once a week for 3 wks and tumors were drivers with PD-L1 expression is still unclear with one study harvested 1 hour after the last injection. Mice were weighed QOD associating PD-L1 expression with mutant EGFR but not KRAS to monitor for toxicity. Tumor burden was calculated as the sum or ALK (18), and another demonstrating no clear difference in PD- of individual lung tumor volumes per mouse.
L1 staining between samples with mutations in EGFR, KRAS, orALK (19). Because the AKT–mTOR pathway serves as a conver- gence point for activation of many of the oncogenes involved in Cell lysates were prepared in 2xLSB. Antibodies were from Cell NSCLC, we hypothesized that this pathway was likely responsible Signaling Technology unless otherwise noted and included for the control of PD-L1 expression. We used NSCLC cell lines, anti–PD-L1 antibody (AbCam; ab58810), anti–phospho- mouse models, and primary human lung cancers to show that PD- AKTS473(9271), anti-AKT(9272), anti–phospho-S6(4858), anti- L1 protein expression is dependent on active AKT–mTOR signal- ing, regardless of specific oncogenic or cytokine stimuli. These data identify a common mechanism of PD-L1 regulation in lung cancer, and provided rationale for clinical trials of oncogenic (2532), anti–phospho-ACCS79(3661), anti-ACC(3662), anti- pathway inhibitors combined with inhibitors of immune Materials and Methods STAT3(8768), anti-p53(2524), anti-p21(2947), and anti-atubulin CL30, IO33, CL13, and CL25 cell lines were derived from 4- Quantitative RT-PCR for PD-L1 lung adenocarcinomas developed in A/J mice, and were a gener- RNA was isolated from CL13 cells using the Qiagen RNeasy ous gift from Dr. Steven Belinsky (Lovelace Respiratory Research Mini Kit (Qiagen). cDNA was made using the SuperScript Institute, Albuquerque, NM) in 1999 (20). Immortalized Beas2B II RT Reaction Kit (Invitrogen) from 2 mg of isolated RNA.
and isogeneic Beas2B transformed with NNK have been described previously (21). HCT-116 parent, PTEN/, PIK3CA mutant or (Mm03928990_g1) primers were purchased from Applied KrasD13/ isogeneic cells were obtained from the JHU Genetic Biosystems. Samples were analyzed on a StepOnePlus RT-PCR Resources Core Facility. Human lung cancer cell lines were early System Instrument using TaqMan Universal PCR Master Mix, No passages (<20) of the original cell lines from the National Cancer AmpErase UNG (Applied Biosystems) according to the manufac- Institute obtained in the years spanning 2000 to 2012. H1975 and H157 cell lines had higher passages (>20) and were authenticatedby JHU Genetic Resources Core Facility. H1299 was authenticated by DDC Medical in 2012. Immortalized Beas2B and isogeneic A total of 1  106 human and mouse lung cancer cells were Beas2B transformed with NNK were a gift in 1996. HCT-116 harvested and stained for 30 minutes at 4C with primary anti- parent, PTEN/, PIK3CA mutant or KrasD13/ isogeneic cells body to PE-anti–mouse-PD-L1 (BioLegend; #10F.9G2), PE-anti– were obtained from the JHU Genetic Resources Core Facility in mouse-B7-H4 (eBioscience; Clone 188), PE-anti–human-PD-L1 2012. All cell lines were passaged for fewer than 6 months after (eBioscience; Clone M1H1), APC-anti–human-B7-H4 (BD Phar- mingen; Clone M1H43) or isotype-matched controls. Sampleswere run on a FACS Caliber (BD Biosciences) and analyzed using FlowJo software (TreeStar).
CL13 cells were transfected with DharmaFECT (Thermo Scien- tific) and a pool of 4 mouse PTEN siRNA or scrambled siRNA (L-040700-00-0005; Thermo Scientific). The pLKO.1 plasmids con- taining shRNA targeted to human RAPTOR or RICTOR have been Formalin-fixed lung tissues were incubated in PD-L1 (CST# described previously (22).
13684), pS6S235/235 (CST#4858), FoxP3 (eBio #14-5773-82),CD3 (A0452 Dako), Ki67 (Ab16667 AbCam), Cleaved caspase- 3 (CST#9664), pHP-1 gamma (ab45270 AbCam), and detection All animal studies were conducted using a protocol approved was completed using the VECTASTAIN Elite ABC Kit (Vector by the Animal Care and Use Committee at the National Cancer Laboratories) per the manufacturer's instructions. Tissues were Institute. The genetically engineered KRASLA2 and CC10þ also incubated in the presence of an isotype-matched control EGFRL585R/T790M mice, as well as, the NNK-induced A/J mouse antibody (sc-2027; Santa Cruz Biotechnology). All stains were lung tumor model have been described previously (3, 23, 24).
quantified in 10 tumor-containing 40 magnification fields. For Cancer Res; 76(2) January 15, 2016 Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Control of PD-L1 by Oncogenic Activation of AKT/mTOR in NSCLC murine PD-L1, the percentage of positive tumor cell surface immunoblotting was analyzed by unpaired the Student t test.
staining was scored as (<5%), 1þ(5–20%), 2þ(20%–50%) Tumor volume and tumor-infiltrating lymphocytes were analyzed or 3þ(50%). pS6 staining was quantified by assigning a score of by two-way ANOVA followed by Tukey's post hoc test. Statistical absent (0), minimal (1), moderate (2), or strong (3) to each significance was reached with a P value less than or equal to 0.05.
tumor. The staining index was calculated for each tumor mymultiplying the staining intensity by its distribution. FoxP3, CD3, Ki67, Cl. Caspase-3 and pHP-1g stains were quantified by count- Expression of PD-L1 in mutant EGFR and mutant KRAS murine ing the number of positive cells. The investigator was blinded to sample identities during scoring.
PD-L1 expression was examined in mouse models of lung TMA slides were stained with the 5H1 antibody for PD-L1 cancer driven by activating mutations in KRAS or EGFR that are expression and a mouse IgG isotype antibody using a previously used to model lung cancer in smokers and never smokers, respec- described protocol by a board certified pathologist (Q.K. Li; tively. In the KRASLA2 mouse model, lung adenocarcinomas ref. 26). Approximately 10% of randomly chosen cores were develop after spontaneous recombination events induce onco- scored to confirm PD-L1 by a second board certified pathologist genic KRASG12D expression. Immunoblotting of lung lysates from (J.M. Taube). Both TMAs were also analyzed for pS6S235/236 KRASLA2 mice demonstrated increased activation of AKT/mTOR (CST#4858) expression. Tumor with >10% phospho-S6 expres- and PD-L1 expression compared with age-matched, wild-type sion were considered positive.
littermates (Fig. 1A). EGFRL858R/T790M mice have doxycycline-inducible expression of human mutant EGFR. Lung lysates har- Statistical analysis vested from mice exposed to doxycycline for 3 weeks show Data in bar graphs are presented as mean  SE. c2 analyses increased expression of EGFR, active AKT/mTOR signaling and tested for differences between the distributions of clinical vari- PD-L1 (Fig. 1A, middle). The tobacco-specific carcinogen NNK ables across histologic samples. The Fisher exact test examined induces KRAS mutations and causes primarily lung adenomas in potential statistical associations of the association between PD-L1 susceptible mouse strains. We previously showed that activation and phospho-S6 expression in both TMAs. Quantification of of the AKT–mTOR pathway is critical for NNK-induced lung C57Bl/6 lung lysates cells + 55
Figure 1.
Expression of PD-L1 and oncogenic activation of the Akt–mTOR pathway. A, lung lysates were harvested from C57BL/6 KRAS LA2 or wt littermates (left),from FVB mEGFRþ/CC10þ littermates treated with or without doxycycline (middle), or from A/J mice exposed to i.p. saline or the tobacco carcinogen NNK (right),and processed for immunoblotting. Each lane represents one mouse. B, A/J mice treated as in A showing PD-L1 expression in lung lesions but not in normallung epithelium. Scale bar, 10 mm. C and D, human NSCLC cell lines (C) and NNK-derived murine lung adenocarcinoma cell lines (D) have activationof AKT–mTOR, as well as expression of PD-L1 as shown by flow cytometry and immunoblotting.
Cancer Res; 76(2) January 15, 2016 Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Lastwika et al.
tumorigenesis (27). In lung lysates from 1-year-old mice previ- PD-L1 by immunoblotting and flow cytometry. The H1770 ously treated with NNK, PD-L1 expression was observed in NNK- (NOTCH1) cell line did not have active AKT/mTOR signaling or but not saline-exposed lungs. Lungs from NNK-treated mice also express PD-L1. Murine cell lines established from NNK-induced had higher activation of AKT and mTOR (Fig. 1A, right). IHC lung adenocarcinomas also had AKT/mTOR activation and PD-L1 staining of lung tissues from mice demonstrates PD-L1 expression expression (Fig. 1D). Expression of PD-L1 in these cell lines in resident immune cells but not in normal lung epithelium (Fig.
appeared selective, because expression of another immunosup- 1B, top and middle). In contrast, PD-L1 was detected in early lung pressive ligand, B7-H4, was only observed in 10% of cells for all lesions after a single NNK exposure. Collectively, these data but one cell line tested (H520; Supplementary Fig. S1A). These demonstrate that PD-L1 is expressed in mouse models of NSCLC studies show that activation of AKT and mTOR is associated with driven by mutations in KRAS or EGFR.
PD-L1 expression in NSCLC lines that harbor a wide spectrum ofdriver mutations.
Expression of PD-L1 and AKT/mTOR activation in NSCLC celllines Inhibition of PI3K, AKT, or mTOR decreases PD-L1 expression NSCLC cell lines were examined for total PD-L1 expression by in NSCLC cell lines immunoblotting, and membranous PD-L1 expression by flow To test whether PD-L1 expression was dependent on active cytometry. The panel of human cell lines was chosen to include a PI3K–AKT–mTOR signaling, murine and human NSCLC cell lines variety of oncogenic drivers, in an effort to reflect the mutational with mutations in KRAS or EGFR and high PD-L1 expression were spectrum seen in patients. AKT/mTOR activation was detected in treated with pharmacologic inhibitors of components in the adenocarcinoma and squamous cell carcinoma cell lines with pathway. Inhibitors of PI3K (LY294002), AKT (TCN-P), or mTOR mutations in NRAS (H1299), KRAS (H157/A549), EGFR (rapamycin) decreased PD-L1 expression in a time-dependent (H1975/H1650), BRAF (H2087), PIK3CA (H1650), EML4-ALK manner (Fig. 2A–C). Although some cell line specificity was (H3122), RET (H1563), autocrine production of FGF2 (H226), or observed, inhibition of PI3K, AKT, and mTOR activity appeared FGFR1 amplification (H520; Fig. 1C). These cell lines expressed to coincide with or precede decreased PD-L1 expression. After Figure 2.
Inhibition of the PI3K–AKT–mTOR pathway decreases PD-L1 expression. A–C, NSCLC cell lines were treated with 10 mmol/L of a PI3K inhibitor (LY294002; A)1 mmol/L of an AKT inhibitor (TCN-P; B), or 100 nmol/L of an mTOR inhibitor (rapamycin; C). D, cells were treated with 2 mmol/L AICAR or vehicle.
E, cell lines were treated with 100 nmol/L of a dual mTORC1/2 inhibitor (AZD8055). F, stable shRNA knockdown of RAPTOR (mTORC1) but not RICTOR(mTORC2) decreases PD-L1 in H157 cells (A–F).
Cancer Res; 76(2) January 15, 2016 Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Control of PD-L1 by Oncogenic Activation of AKT/mTOR in NSCLC 48 hours of PI3K or AKT inhibition, recovery of pathway activa- (Supplementary Fig. S1A). Although the majority of these cell tion and expression of PD-L1 occurred with similar kinetics in cell lines did not express B7-H4, the highest B7-H4–expressing cell lines. In contrast, recovery of mTOR activity or PD-L1 expression line, H520, was treated with rapamycin. Rapamycin did not alter was not observed with rapamycin at the time points examined, B7-H4 protein in H520 cells, suggesting that mTOR specifically possibly due to its long half-life. To investigate whether PI3K and regulates PD-L1 (Supplementary Fig. S1B). To determine whether AKT inhibition were required for mTOR inhibition and decreased other signaling pathways downstream of oncogenic drivers such expression of PD-L1, we used AICAR, an activator of AMPK that as the MEK–ERK pathway might play a role in regulating PD-L1, can inhibit mTOR independently of PI3K and AKT. AICAR acti- cells were treated with an MEK inhibitor, U0126. U0126 did not vated AMPK, increased phosphorylation of the AMPK substrate alter PD-L1 expression despite inhibiting ERK phosphorylation ACC, inhibited mTORC1 activation, and decreased PD-L1 expres- and proliferation (Supplementary Fig. S2), indicating that control sion at 16 hours (Fig. 2D). Taken together, these results demon- of PD-L1 expression was specific to the PI3K–AKT–mTOR path- strate that inhibition of PI3K, Akt, or mTOR (through allosteric way and was not due to stimulation of the MEK–ERK pathway or inhibition with rapamycin or AMPK activation), decreases PD-L1 to indirect effects on cellular proliferation.
To confirm the results obtained with rapamycin and AICAR, we Rapamycin decreases PD-L1 expression in murine lung tumors also tested a dual mTORC1/2 inhibitor, AZD8055. AZD8055 To validate these in vitro studies, we examined the effects of decreased PD-L1 expression coincident with decreased activation rapamycin on PD-L1 expression in vivo. One year after exposure of AKT and mTOR (Fig. 2E). Because inhibition of PD-L1 by to NNK, 1 week of rapamycin treatment significantly reduced rapamycin correlated more closely with inhibition of pS6 but not mTOR signaling and decreased PD-L1 expression in A/J mouse pAKT at early time points, this suggested that mTORC1 exerts lung tumors compared with vehicle-treated lung tumors (Fig.
more control over PD-L1 expression than mTORC2. To discern 3A). Similarly, lung tumors from KRASLA2 mice treated for 10 whether PD-L1 expression is dependent on mTORC1 or weeks with rapamycin also had lower PD-L1 expression and mTORC2, shRNA-mediated knockdown of a key component in mTOR activation compared with vehicle-treated littermates mTORC1 (RAPTOR) or mTORC2 (RICTOR) was performed in (Fig. 3B). Six weeks after doxycycline administration, mutant H157 cells. Knockdown of RAPTOR but not RICTOR decreased EGFRL858R/T790M mice were treated for 1 week with vehicle or PD-L1 expression, even though RICTOR knockdown decreased rapamycin. Lung tumors from mice treated with rapamycin had phosphorylation of AKT at serine 473 (Fig. 2F).
reduced mTOR activation and PD-L1 expression compared to To determine whether mTOR could regulate other immuno- vehicle-treated mice (Fig. 3C). These studies indicate that suppressive ligands expressed on tumors, we performed flow mTOR activation is correlated with PD-L1 expression in murine cytometry for B7-H4 using the same panel of NSCLC cells lung tumors.
Figure 3.
Rapamycin decreases PD-L1 expression in lung tumors in vivo. Immunohistochemical staining for pS6 or PD-L1 in lung tumors treated with vehicle or rapamycinfrom A/J mice exposed to the tobacco-carcinogen NNK (A), C57BL/6 KRAS LA2 mice (B), or dox-exposed FVB mEGFRþ/CC10þ littermates (C). Bar graphsquantify decreased staining of pS6 and PD-L1. Scale bar, 10 mm. , P  0.05 by unpaired Student t test.
Cancer Res; 76(2) January 15, 2016 Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Lastwika et al.
Activation of the AKT–mTOR pathway increases PD-L1 genetic results linking active AKT/mTOR signaling to PD-L1 expression in NSCLC, pairs of isogenic HCT116 cells were used On the basis of the observation that inhibition of the PI3K– to determine whether single genetic alterations of the pathway AKT–mTOR pathway decreases PD-L1 expression, we tested could increase PD-L1 expression. Increased activation of Akt and whether cell lines with low basal levels of PD-L1 could increase mTOR, as well as increased expression of PD-L1 was observed in PD-L1 expression after stimulation of the AKT–mTOR pathway.
HCT116 PTEN/ cells, suggesting that regulation of PD-L1 by Administration of EGF to NSCLC cell lines activated the pathway PTEN may occur in several tumor types. HCT116 cells that express and increased PD-L1 expression (Fig. 4A). Likewise, mouse and mutant KRAS or mutant PIK3CA alleles also had higher AKT human cell lines rapidly activated AKT and mTOR and increased activation and PD-L1 expression compared with isogeneic cells PD-L1 expression (Fig. 4B). Comparison of BEAS-2B cells with with corresponding wild-type alleles (Fig. 4E).
BEAS-2B cells fully transformed by NNK showed increased PD-L1expression and activation of AKT/mTOR in cells fully transformed EGF and IFNg increase PD-L1 protein expression through by NNK (Fig. 4C). Knockdown of PTEN, a negative regulator of activation of mTOR PI3K, increased the activation of AKT and PD-L1 expression in PD-L1 expression can also be induced in tumors in response to CL13 cells (Fig. 4D). To complement the pharmacologic and proinflammatory cytokines like IFNg via JAK/STAT signaling and Figure 4.
Activation of the AKT–mTOR pathway increases PD-L1 expression. A, NSCLC cell lines were treated with 5 ng/mL EGF or vehicle in 1% serum. B, NSCLC cell lines weretreated with 100 nmol/L NNK or vehicle in 1% serum. C, lysates from B2B and B2B-NNK isogeneic cell lines were evaluated by immunoblotting. D, cells weredirectly transfected with scrambled siRNA or PTEN-targeted siRNA for 24 hours in serum-free media. Cells with wt PTEN or that had lost PTEN are shown on theright. E, comparison of cells that have lost either the mutant KRAS or PIK3CA alleles or the corresponding wt alleles through homologous recombination (HR).
Cancer Res; 76(2) January 15, 2016 Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Control of PD-L1 by Oncogenic Activation of AKT/mTOR in NSCLC interferon-stimulated response elements in the PD-L1 promoter ylation events are rapidly controlled (Fig. 4C), we also included a (7, 28). STAT1 and to a lesser extent STAT3 typically mediates 30-minute time point to observe EGF-induced AKT and mTOR IFNg signaling. Although both STAT1 and STAT3 can bind to the activation. At 16 hours, EGF and IFNg increased PD-L1 expression PD-L1 promoter, STAT3 binds with higher affinity and stimulates and activated mTOR signaling. Upregulation of PD-L1 was depen- more PD-L1 transcript in dendritic cells (DC; ref. 29). Phosphor- dent on mTOR activation, because rapamycin pretreatment pre- ylation of JAK/STAT occurs minutes after exposure to IFNg, but in vented EGF- and IFNg-mediated increases in PD-L1 expression, multiple cancer cell lines maximum induction of PD-L1 occurs but not IFNg-induced p-STAT3 (Fig. 5B).
much later (9–24 hours; ref. 28). For these reasons, we chose to Although EGF and IFNg induce PD-L1 protein expression in an use phospho-STAT3 as a readout for IFNg signaling and evaluated mTOR-dependent manner, it is unclear whether mTOR exerts PD-L1 expression at later time points. Cells were treated with IFNg transcriptional control of PD-L1. Therefore, we measured PD-L1 for 16 and 24 hours (Fig. 5A). IFNg activated JAK2 and STAT3 transcription. IFNg increased transcription of PD-L1 but EGF did signaling, as well as PD-L1 expression. Parallel cultures of cells not. Rapamycin did not inhibit IFNg-induced transcription, were also treated with EGF to compare PD-L1 regulation and suggesting that mTOR provides translational control of PD-L1 signaling pathway activation. Because EGF-stimulated phosphor- (Fig. 5C). To confirm translational regulation of PD-L1, NSCLC Figure 5.
The AKT–mTOR pathway controls PD-L1 protein expression. A, CL13 and H1299 cell lines were treated with 10 ng/mL IFNg or 5 ng/mL EGF in 1% serum forthe indicated times. An early time point (30 m) was included after EGF addition to confirm pathway activation. B, cells were treated for 24 hours with100 nmol/L rapamycin alone, for 23 hours with 5 ng/mL EGF or 10 ng/mL IFNg alone, or the combination by treating with rapamycin for 1 hour, then addingEGF or IFNg to culture media and harvesting 23 hours later. C, CL13 cells treated as in B and RNA was harvested for RT-PCR. D, cells were treated with100 mg/mL cycloheximide for the indicated time points. E, cells were treated with 5 mg/mL actinomycin D for the indicated time points. Immunoblotsshown are representative of three independent experiments. F, cells were treated for 6 hours with 300 mmol/L chloroquine (CLQ) alone, for 4 hours with100 nmol/L rapamycin alone, or the combination by treating with CLQ for 2 hours, then adding rapamycin to culture media and harvesting 4 hours later.
G, cells were treated for 6 hours with 100 nmol/L PS-341 alone, for 5 hours with 100 nmol/L rapamycin alone, or the combination by treating with PS-341for 2 hours, then adding rapamycin to culture media and harvesting 5 hours later.
Cancer Res; 76(2) January 15, 2016 Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Lastwika et al.
Table 1. mTOR activation is required, but may not be sufficient to induce PD-L1 combined for further analyses. Approximately 90% of tumors expression in primary lung adenocarcinoma and squamous cell carcinoma with PD-L1 expression had activation of mTOR and 54% of tumors with mTOR activation also expressed PD-L1, suggesting that mTOR activation was necessary, but not sufficient, for PD-L1 expression (Table 1). Distribution of mTOR activation tended to be similar to staining patterns for PD-L1, suggesting that the same NOTE: Correlation between p-S6S235/236 and PD-L1 markers in the TMAs.
cells co-express both markers (Supplementary Fig. S4). The major- Statistical analyses were performed using the Fisher exact test.
ity (83%) of tumors negative for pS6 were also negative for PD-L1.
A small subset of tumors expressed PD-L1 without mTOR acti-vation, indicating that there may be additional mechanisms cell lines were exposed to the protein translation inhibitor cyclo- inducing PD-L1 expression. Sixty-three of 158 (40%) of lung heximide (Fig. 5D). Cycloheximide rapidly decreased PD-L1 tumors had both active mTOR signaling and PD-L1 expression, protein expression, indicating that PD-L1 likely has rapid turnover and a Fisher exact test revealed a statistically significant correlation in lung cancer cells. In contrast with cycloheximide, inhibiting between the two markers (P ¼ 0.0001; Table 1). These results transcription with actinomycin D did not change PD-L1 expres- underscore the clinical relevance of our preclinical associations.
sion, even at later time points (Fig. 5E). The accumulation of p53in IO33 cells and the accumulation of p21 in the mutant p53 cellline H1975 demonstrated that transcription was successfully The combination of rapamycin and a PD-1 blocking antibody inhibited. These results suggest that PD-L1 expression is predom- decreases lung tumor growth inantly controlled at the protein level and that mTOR exerts its Monoclonal antibodies that block PD-L1 or PD-1 have shown regulation at this level.
clinical benefit in NSCLC (9, 10). However, it is possible that To examine how rapamycin was decreasing PD-L1 protein simultaneous inhibition of both PD-L1 and PD-1 may increase expression, we studied two main pathways of protein degradation therapeutic benefit because each has additional immunosuppres- via the lysosome or the proteasome. Pretreatment with a lysosome sive binding partners. To test the efficacy of systemically blocking acidification inhibitor (chloroquine) but not a proteasome inhib- PD-1 while reducing the expression of PD-L1 in tumor tissue, a itor (PS-341) prevented rapamycin-mediated decreases in PD-L1 murine anti–PD-1 antibody and rapamycin were administered in protein (Fig. 5F and G). This suggests that rapamycin inhibits PD- the KRASLA2 mouse model (Fig. 6A). Rapamycin alone decreased L1 expression through a combination of decreased protein syn- the tumor burden of KRAS-driven lung tumors by approximately thesis and increased lysosomal protein degradation.
50% whereas PD-1 blockade had no effect as a single agent (Fig.
6B). The combination of rapamycin and anti–PD-1 significantlyreduced lung tumor burden by comparison with any other treat- Expression of PD-L1 and activation of mTOR in human lung ment group. The combination therapy increased CD3þ T cells and adenocarcinomas and squamous cell carcinomas reduced FoxP3þ Tregs (Fig. 6C). This led to a higher ratio of CD3þ To determine whether these findings are clinically relevant, two T cells to Tregs, indicating a shift towards an immune activated human lung tissue microarrays (TMA) were stained and scored for rather than immunosuppressive microenvironment. Lung tumors membranous and/or cytoplasmic PD-L1 expression (Supplemen- from mice treated with rapamycin had a reduction in PD-L1 tary Fig. S3). One TMA included 63 lung adenocarcinomas with expression and mTOR activation (Fig. 6D and E). A marker of matched normal lung and assorted normal tissues. The other TMA apoptosis, cleaved caspase-3, was increased in tumors treated with contained 96 lung squamous cell carcinomas with assorted nor- the combination. Although rapamycin alone inhibited tumor mal tissues. These normal tissues served as internal positive proliferation, tumors treated with the combination also had more (placenta) or negative (soft tissue) controls for PD-L1 expression.
pHP1gþ cells, suggesting that these tumors had undergone senes- Each TMA was simultaneously stained with an IgG antibody to cence. These findings demonstrate enhanced antitumor efficacy control for background. Clinical and pathologic characteristics of with the combination of rapamycin and a PD-1 blocking antibody the patient population are summarized in Supplementary Table through increased apoptosis and cellular senescence. We con- S1. Sixty-two of 63 adenocarcinoma and 96 of 96 squamous cell firmed the efficacy of rapamycin and PD-1 blockade in a second carcinoma tumors were evaluable. Twenty of 62 (32.2%) lung mouse model of KRAS-driven lung cancer (Supplementary Figs.
adenocarcinomas and 50 of 96 (52.1%) lung squamous cell S5 and S6). These findings demonstrate enhanced antitumor carcinomas expressed membranous PD-L1, which is consistent efficacy against two mutant KRAS mouse models of lung cancer with previous observations (Supplementary Fig. S3B and S3D; when rapamycin and a PD-1 blocking antibody are combined.
refs. 30–34). No clinical or pathologic characteristics were asso-ciated with PD-L1 expression. PD-L1 membranous expression was observed on lung tumor tissue and on resident alveolarmacrophages, but not on non-neoplastic lung tissue. These data PD-L1 plays a prominent role in the balance of the immune support a potential common role of this protein in mediating system between the stimulatory signals needed for effective immunosuppression in NSCLC.
immune responses and maintenance of self-tolerance or tissue To explore the potential regulation of PD-L1 in human primary integrity. PD-L1 can be expressed on hematopoietic and non- lung tumors by mTOR activation, both TMAs were also stained hematopoietic cells, as well as in lymphoid and peripheral tissues.
with an antibody specific for phosphorylation of S6 at S235/236 Consequently, the regulation of PD-L1 is complex and most likely (Supplementary Fig. S3C and S3D). Because there were no sig- depends on the status of underlying transcriptional and signaling nificant differences between the TMA characteristics (Supplemen- networks. Here, our studies reveal a strong association between tary Table S1; stage significance is likely due to a sample size bias), PD-L1 protein and activation of the AKT–mTOR pathway in lung the adenocarcinomas and squamous cell carcinomas were cancer. The dependence of PD-L1 expression on mTOR is Cancer Res; 76(2) January 15, 2016 Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Control of PD-L1 by Oncogenic Activation of AKT/mTOR in NSCLC Sacrifice 7 weeks
4.5 mg/kg LD; 1.5 mg/kg veh QOD and
N = 12
μg IgG weekly i.p.
N = 12
4.5 mg/kg LD; 1.5 mg/kg rapa QOD
N = 12
μg αPD-1 weekly i.p.
4.5 mg/kg LD; 1.5 mg/kg rapa QOD and
N = 12
150 μg αPD-1 weekly i.p.
pS6S235/236
P < 0.0001 P < 0.0001 cells/HPF 0
Cl. Caspase
Tumor burden
g+ cells/HPF 20
cells/HPF 30
/FoxP3 10
Figure 6.
The combination of rapamycin and aPD-1 blockade significantly reduces lung tumor burden in the KRAS LA2 mouse model. A, KRAS LA2 mice were treatedwith either IP vehicle and IgG, rapamycin, aPD-1 antibody, or rapamycin and aPD-1 for 4 weeks beginning at the time of weaning. B, tumor burden after4 weeks of treatment; , P  0.05 by Mann–Whitney. C, quantification of IHC staining for CD3þ or FoxP3þ cells. The ratio of CD3þ over FoxP3þ cells is also shown; , P  0.05 by two-way ANOVA. D, images represent IHC staining for PD-L1, pS6, cleaved caspase-3, Ki67, and pHP1g. Scale bar, 10 mm. E, quantification of IHCstains in D. , P  0.05 by two-way ANOVA.
consistent with studies in glioma, breast, prostate, ovarian, and regulatory steps on PD-L1 expression will probably depend on cell pancreatic cancer. Interestingly, this relationship does not extend type, context, and may vary over the course of response to a to melanoma, emphasizing multiple mechanisms for PD-L1 regulation in solid tumors (35).
Transcription of PD-L1 can be induced by many cytokines, of Our TMA study suggested that mTOR activation is necessary but which IFNg is the most potent (7). Activation of the AKT–mTOR not sufficient for PD-L1 expression. It is possible that tumors with pathway plays a central role in the initiation of IFN-stimulated mTOR activation but no PD-L1 protein lack PD-L1 transcripts, gene translation, in a mechanism parallel to but independent of which would preclude mTOR-dependent translation. Approxi- activation of the JAK–STAT pathway (39). Therefore, although mately 53% (810/1,537) of lung cancer specimens in The Cancer PD-L1 transcription does not depend on mTOR activation, trans- Genome Atlas set do not have detectable PD-L1 mRNA (36).
lation of IFNg-induced transcripts, including PD-L1, may be Other studies have identified that PD-L1 mRNA levels were only dependent on activation of PI3K, AKT, and mTOR kinase activity.
higher than normal lung tissue in stage IV lung tumors (37). Thus, The dependence of PD-L1 translation on PI3K–AKT–mTOR activ- there may be additional levels of PD-L1 regulation between ity is also observed during viral infections. In HIV-1–infected transcription and translation. A recent study directly compared macrophages and dendritic cells, the viral protein Nef induces PD- samples for PD-L1 mRNA and protein expression and observed L1 transcription by binding to the promoter but PD-L1 protein that PD-L1 mRNA had a complex, nonlinear positive association expression depends on active PI3K/AKT signaling (40). Our data with PD-L1 protein expression. This finding was consistent in two indicate that multiple types of stimuli, including growth factors separate TMA cohorts and suggests that PD-L1 is regulated at both cytokines and oncogenes, converge at mTOR to increase PD-L1 transcription and translational levels (34). In DCs, LPS and IFNg- mediated induction of PD-L1 protein depends on both active Inhibiting ligation of tumor-derived PD-L1 with PD-1 on T cells transcription and translation (38). The relative contribution of is proposed as a major therapeutic target to revert tumor-mediated Cancer Res; 76(2) January 15, 2016 Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Lastwika et al.
immunosuppression. However, because PD-L1 and PD-1 have PD-L1 expression, it is possible that response to PD-1 or PD-L1 additional binding partners perhaps blocking multiple interac- blockade depends on a critical threshold of TILs at the start of tions is needed to fully rescue antitumor immunity. Combining therapy (50). We have observed that the number of tumor- PD-1 blockade with rapamycin, which inhibited oncogenic KRAS infiltrating CD3þ T cells per high powered field is doubled in signaling and PD-L1 expression, resulted in a significant reduction the mutant EGFRL858R/T790M compared with the KRASLA2 mouse in tumor burden compared with either drug alone. Only the model (88.9 vs. 37.5 CD3þ T cells; unpublished data). Further- combination therapy significantly increased the ratio of CD3þ to more, a recent study demonstrated higher nonsynonymous muta- FoxP3þ cells, supporting this change in T-cell populations as a tional burden is associated with response to PD-1 blockade as a readout for antitumor activity. In addition, only the combination single agent, in part by enhancing neoantigen-specific CD8þ T-cell was associated with decreased proliferation and increased apo- responses (51). Identifying the mechanisms responsible for the ptotic and senescent markers. Drug-induced senescence with differences in lung tumors and TILs between responders and non- DNA-damaging agents is well established, but a role for adaptive responders of PD-1 blockade would have important insight into immunity in driving cancer cell senescence was recently identified.
In multiple murine models and in human cancers, T helper 1 cell Activation of PI3K–AKT–mTOR signaling is driven by mul- production of IFNg and TNFa induce immune-dependent tumor tiple mechanisms in NSCLC and is vital to tumor develop- cell senescence (41). Although this is the first demonstration of ment, progression, and prognosis. We show that activation of immune-induced senescence in tumors, immune cells promote AKT–mTOR, regardless of the driving oncogene or exogenous senescence to regulate other leukocytes. For example, Tregs can stimulus, increases PD-L1 protein expression in NSCLC. Our induce senescence in na€ve and memory T cells through a mech- data extend a growing body of evidence that oncogenes have anism dependent on toll like receptor 8, p38, and ERK1/2 (42). A tumor cell autonomous effects by altering the immune system remarkable aspect of checkpoint blockade with PD-1 or PD-L1 is in the tumor microenvironment. Clinical trials combining the generation of long-term stable disease in the absence of anti–PD-1 antibodies and current standard-of-care treatments complete tumor regression, raising the possibility that these are already underway and include combining targeted thera- tumors have undergone senescence.
pies with immunotherapy (73–75). Our studies provide ratio- Implementing rapamycin as a cancer therapy raises issues about nale to combine and optimize PI3K–AKT–mTOR inhibitors its own role in immunosuppression. Rapamycin has a black box with anti–PD-1 antibodies.
warning from the FDA stemming from a study of renal transplantpatients who were also taking cyclosporine and corticosteroids Disclosure of Potential Conflicts of Interest (43), but multiple trials of single-agent rapamycin or rapamycin J.M. Taube reports receiving a commercial research grant and is a consultant/ analogues in cancer patients have shown no evidence of increased advisory board member for Bristol Myers Squibb. No potential conflicts of incidence of immunosuppression (25, 44). In fact, many basic interest were disclosed by the other authors.
and clinical studies have associated rapamycin with activeimmune responses (45, 46). Our studies in the NNK-induced Authors' Contributions lung cancer model have shown only modest decreases in CD4þ Conception and design: K.J. Lastwika, W. Wilson III, S. Yao, L.N. Liu, P.A.
levels with short-term or continuous rapamycin treatment. Fac- tors that are likely to play important roles in the cumulative effects Development of methodology: W. Wilson III, J. Norris, H. Xu, P.A. Dennis of rapamycin on the immune system include the timing and Acquisition of data (provided animals, acquired and managed patients, degree of mTOR inhibition, as well as cell type and modulation of provided facilities, etc.): K.J. Lastwika, J. Norris, H. Kitagawa, S. Kawabata, mTORC2 signaling. Although precise mechanisms remain unclear, we demonstrate the potential to use rapamycin in com- Analysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): K.J. Lastwika, S.R. Ghazarian, S. Kawabata, J.M.
bination with a PD-1 blocking antibody to increase antitumor Taube, P.A. Dennis immunity. Rapamycin administration has been shown to sensi- Writing, review, and/or revision of the manuscript: K.J. Lastwika, W. Wilson tize tumors to immunotherapy in other mouse model systems.
III, Q.K. Li, S.R. Ghazarian, H. Kitagawa, J.M. Taube, S. Yao, J.J. Gills, P.A. Dennis For example, treatment of fibrosarcoma or colorectal cancers with Administrative, technical, or material support (i.e., reporting or organizing rapamycin increased tumor sensitivity to adoptive cellular immu- data, constructing databases): Q.K. Li, S. Kawabata, S. Yao, P.A. Dennis notherapy (47). Although PD-L1 expression was not examined in Study supervision: P.A. Dennis this study, it is tempting to speculate PD-L1 as a contributingfactor in immunosuppression.
Responses to PD-1 and PD-L1 blockade have been proposed to The authors thank Dr. Leiping Chen for providing the anti–PD-L1 mono- be associated with the presence of PD-L1 and many ongoing clonal antibody 5H1.
clinical trials require PD-L1þ pretreatment biopsies. Despitestrong expression of PD-L1 in lung tumors, PD-1 blockade had no effect on tumorigenesis in the KRASLA2 mouse model. This is in This work was supported by intramural funding from the National agreement with a report demonstrating PD-1 blockade reduced Cancer Institute, the George Washington University, and NIH grant P30 tumor burden in mouse models of mutant EGFR- but not KRAS- driven lung cancer (16). Because both mutant EGFR and KRAS The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked tumor models express PD-L1, this may indicate specific genomic advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate subsets of lung tumors predict response to single-agent anti–PD-1 outside of PD-L1 expression. However, multiple clinical studieshave not identified the presence of mutant KRAS or EGFR as Received November 19, 2014; revised August 30, 2015; accepted September predictors for successful PD-1 blockade (48, 49). In addition to 20, 2015; published OnlineFirst December 4, 2015.
Cancer Res; 76(2) January 15, 2016 Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Control of PD-L1 by Oncogenic Activation of AKT/mTOR in NSCLC 1. Siegel R, Ma J, Zou Z, Jemal A. Cancer statistics, 2014. CA Cancer J Clin 22. Sarbassov DD, Guertin DA, Ali SM, Sabatini DM. Phosphorylation and regulation of Akt/PKB by the rictor–mTOR complex. Science 2005;307: 2. Memmott RM, Mercado JR, Maier CR, Kawabata S, Fox SD, Dennis PA.
Metformin prevents tobacco carcinogen–induced lung tumorigenesis.
23. Politi K, Zakowski MF, Fan PD, Schonfeld EA, Pao W, Varmus HE. Lung Cancer Prev Res 2010;3:1066–76.
adenocarcinomas induced in mice by mutant EGF receptors found in 3. Granville CA, Warfel N, Tsurutani J, Hollander MC, Robertson M, Fox SD, human lung cancers respond to a tyrosine kinase inhibitor or to down- et al. Identification of a highly effective rapamycin schedule that markedly regulation of the receptors. Genes Dev 2006;20:1496–510.
reduces the size, multiplicity, and phenotypic progression of tobacco 24. Johnson L, Mercer K, Greenbaum D, Bronson RT, Crowley D, Tuveson DA, carcinogen-induced murine lung tumors. Clin Cancer Res 2007;13: et al. Somatic activation of the K-ras oncogene causes early onset lung cancer in mice. Nature 2001;410:1111–6.
4. Hollander MC, Maier CR, Hobbs EA, Ashmore AR, Linnoila RI, Dennis PA.
25. O'Donnell A, Faivre S, Burris HA III, Rea D, Papadimitrakopoulou V, Shand Akt1 deletion prevents lung tumorigenesis by mutant K-ras. Oncogene N, et al. Phase I pharmacokinetic and pharmacodynamic study of the oral mammalian target of rapamycin inhibitor everolimus in patients with 5. Granville CA, Memmott RM, Balogh A, Mariotti J, Kawabata S, Han W, et al.
advanced solid tumors. J Clin Oncol 2008;26:1588–95.
A central role for Foxp3þ regulatory T cells in K-Ras–driven lung tumor- 26. Taube JM, Anders RA, Young GD, Xu H, Sharma R, McMiller TL, et al.
igenesis. PLoS ONE 2009;4:e5061.
Colocalization of inflammatory response with B7-h1 expression in human 6. Dong H, Zhu G, Tamada K, Chen L. B7-H1, a third member of the B7 melanocytic lesions supports an adaptive resistance mechanism of family, co-stimulates T-cell proliferation and interleukin-10 secretion. Nat immune escape. Sci Transl Med 2012;4:127ra37.
27. West KA, Linnoila IR, Belinsky SA, Harris CC, Dennis PA. Tobacco carcin- 7. Dong H, Strome SE, Salomao DR, Tamura H, Hirano F, Flies DB, et al.
ogen-induced cellular transformation increases activation of the phospha- Tumor-associated B7-H1 promotes T-cell apoptosis: a potential mecha- tidylinositol 30-kinase/Akt pathway in vitro and in vivo. Cancer Res 2004;64: nism of immune evasion. Nat Med 2002;8:793–800.
8. Butte MJ, Keir ME, Phamduy TB, Sharpe AH, Freeman GJ. Programmed 28. Lee SJ, Jang BC, Lee SW, Yang YI, Suh SI, Park YM, et al. Interferon regulatory death-1 ligand 1 interacts specifically with the B7-1 costimulatory molecule factor-1 is prerequisite to the constitutive expression and IFN-gamma- to inhibit T-cell responses. Immunity 2007;27:111–22.
induced upregulation of B7-H1 (CD274). FEBS Lett 2006;580:755–62.
9. Topalian SL, Hodi FS, Brahmer JR, Gettinger SN, Smith DC, McDermott DF, 29. Wolfle SJ, Strebovsky J, Bartz H, Sahr A, Arnold C, Kaiser C, et al. PD-L1 et al. Safety, activity, and immune correlates of anti–PD-1 antibody in expression on tolerogenic APCs is controlled by STAT-3. Eur J Immunol cancer. N Engl J Med 2012;366:2443–54.
10. Brahmer JR, Tykodi SS, Chow LQM, Hwu WJ, Topalian SL, Hwu P, et al.
30. Mu CY, Huang JA, Chen Y, Chen C, Zhang XG. High expression of PD-L1 in Safety and activity of anti–PD-L1 antibody in patients with advanced lung cancer may contribute to poor prognosis and tumor cells immune cancer. N Engl J Med 2012;366:2455–65.
escape through suppressing tumor infiltrating dendritic cells maturation.
11. Lipson EJ, Sharfman WH, Drake CG, Wollner I, Taube JM, Anders RA, Med Oncol 2011;28:682–8.
et al. Durable cancer regression off-treatment and effective reinduc- 31. Chen YB, Mu CY, Huang JA. Clinical significance of programmed death-1 tion therapy with an anti–PD-1 antibody. Clin Cancer Res 2013;19: ligand-1 expression in patients with non–small cell lung cancer: a 5-year- follow-up study. Tumori 2012;98:751–5.
12. Green MR, Monti S, Rodig SJ, Juszczynski P, Currie T, O'Donnell E, et al.
32. Boland JM, Kwon ED, Harrington SM, Wampfler JA, Tang H, Yang P, et al.
Integrative analysis reveals selective 9p24.1 amplification, increased PD-1 Tumor B7-H1 and B7-H3 expression in squamous cell carcinoma of the ligand expression, and further induction via JAK2 in nodular sclerosing lung. Clin Lung Cancer 2013;14:157–63.
Hodgkin lymphoma and primary mediastinal large B-cell lymphoma.
33. Konishi J, Yamazaki K, Azuma M, Kinoshita I, Dosaka-Akita H, Nishimura M. B7-H1 expression on non–small cell lung cancer cells and its relation- 13. Marzec M, Zhang Q, Goradia A, Raghunath PN, Liu X, Paessler M, et al.
ship with tumor-infiltrating lymphocytes and their PD-1 expression. Clin Oncogenic kinase NPM/ALK induces through STAT3 expression of immu- Cancer Res 2004;10:5094–100.
nosuppressive protein CD274 (PD-L1, B7-H1). Proc Natl Acad Sci U S A 34. Velcheti V, Schalper KA, Carvajal DE, Anagnostou VK, Syrigos KN, Sznol M, et al. Programmed death ligand-1 expression in non–small cell lung cancer.
14. Parsa AT, Waldron JS, Panner A, Crane CA, Parney IF, Barry JJ, et al. Loss of Lab Invest 2014;94:107–16.
tumor suppressor PTEN function increases B7-H1 expression and immu- 35. Jiang X, Zhou J, Giobbie-Hurder A, Wargo J, Hodi FS. The activation of noresistance in glioma. Nat Med 2007;13:84–8.
MAPK in melanoma cells resistant to BRAF inhibition promotes PD-L1 15. Crane CA, Panner A, Murray JC, Wilson SP, Xu H, Chen L, et al. PI(3) kinase expression that is reversible by MEK and PI3K inhibition. Clin Cancer Res is associated with a mechanism of immunoresistance in breast and prostate cancer. Oncogene 2009;28:306–12.
36. Rhodes DR, Kalyana-Sundaram S, Mahavisno V, Varambally R, Yu J, 16. Akbay EA, Koyama S, Carretero J, Altabef A, Tchaicha JH, Christensen CL, Briggs BB, et al. Oncomine 3.0: genes, pathways, and networks in a et al. Activation of the PD-1 pathway contributes to immune escape in collection of 18,000 cancer gene expression profiles. Neoplasia 2007;9: EGFR-driven lung tumors. Cancer Discov 2013;3:1355–63.
17. Xu C, Fillmore CM, Koyama S, Wu H, Zhao Y, Chen Z, et al. Loss of Lkb1 37. Sasaki H, Suzuki A, Shitara M, Hikosaka Y, Okuda K, Moriyama S, et al.
and Pten leads to lung squamous cell carcinoma with elevated PD-L1 gene expression in Japanese lung cancer patients. Biomed Rep 2013;1: expression. Cancer Cell 2014;25:590–604.
18. D'Incecco A, Andreozzi M, Ludovini V, Rossi E, Capodanno A, Landi L, et al.
38. Loke P, Allison JP. PD-L1 and PD-L2 are differentially regulated by Th1 and PD-1 and PD-L1 expression in molecularly selected non–small cell lung Th2 cells. Proc Natl Acad Sci U S A 2003;100:5336–41.
cancer patients. Br J Cancer 2015;112:95–102.
39. Kaur S, Sassano A, Dolniak B, Joshi S, Majchrzak-Kita B, Baker DP, et al.
19. Garon EB, Rizvi NA, Hui R, Leighl N, Balmanoukian AS, Eder JP, et al.
Role of the Akt pathway in mRNA translation of interferon-stimulated Pembrolizumab for the treatment of non–small cell lung cancer. N Engl J genes. Proc Natl Acad Sci U S A 2008;105:4808–13.
40. Muthumani K, Shedlock DJ, Choo DK, Fagone P, Kawalekar OU, Good- 20. Jones-Bolin SE, Johansson E, Palmisano WA, Anderson MW, Wiest JS, man J, et al. HIV-mediated phosphatidylinositol 3-kinase/serine-threonine Belinsky SA. Effect of promoter and intron 2 polymorphisms on murine kinase activation in APCs leads to programmed death-1 ligand upregula- lung K-ras gene expression. Carcinogenesis 1998;19:1503–8.
tion and suppression of HIV-specific CD8 T cells. J Immunol 2011;187: 21. Klein-Szanto AJ, Iizasa T, Momiki S, Garcia-Palazzo I, Caamano J, Metcalf R, et al. A tobacco-specific N-nitrosamine or cigarette smoke condensate 41. Braumuller H, Wieder T, Brenner E, Assmann S, Hahn M, Alkhaled M, et al.
causes neoplastic transformation of xenotransplanted human bronchial T-helper-1-cell cytokines drive cancer into senescence. Nature 2013;494: epithelial cells. Proc Natl Acad Sci U S A 1992;89:6693–7.
Cancer Res; 76(2) January 15, 2016 Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Lastwika et al.
42. Ye J, Huang X, Hsueh EC, Zhang Q, Ma C, Zhang Y, et al. Human regulatory 47. Hahnel PS, Thaler S, Antunes E, Huber C, Theobald M, Schuler M. Targeting T cells induce T-lymphocyte senescence. Blood 2012;120:2021–31.
AKT signaling sensitizes cancer to cellular immunotherapy. Cancer Res 43. Hidalgo M, Buckner JC, Erlichman C, Pollack MS, Boni JP, Dukart G, et al. A phase I and pharmacokinetic study of temsirolimus (CCI-779) adminis- 48. Garon EB, Balmanoukian A, Hamid O, Hui R, Gandhi L, Leighl N, et al.
tered intravenously daily for 5 days every 2 weeks to patients with advanced Abstract A20: MK-3475 monotherapy for previously treated non–small cell cancer. Clin Cancer Res 2006;12:5755–63.
lung cancer (NSCLC): preliminary safety and clinical activity. Clin Cancer 44. Bissler JJ, McCormack FX, Young LR, Elwing JM, Chuck G, Leonard JM, et al.
Res 2014;20(2 Suppl):A20-A.
Sirolimus for angiomyolipoma in tuberous sclerosis complex or lymphan- 49. Creelan BC. Update on immune checkpoint inhibitors in lung cancer.
gioleiomyomatosis. N Engl J Med 2008;358:140–51.
Cancer Control 2014;21:80–9.
45. Haydar AA, Denton M, West A, Rees J, Goldsmith DJ. Sirolimus-induced 50. Herbst RS, Soria JC, Kowanetz M, Fine GD, Hamid O, Gordon MS, et al.
pneumonitis: three cases and a review of the literature. Am J Transplant Predictive correlates of response to the anti–PD-L1 antibody MPDL3280A in cancer patients. Nature 2014;515:563–7.
46. Rao RR, Li Q, Odunsi K, Shrikant PA. The mTOR kinase determines 51. Rizvi NA, Hellmann MD, Snyder A, Kvistborg P, Makarov V, Havel JJ, effector versus memory CD8þ T-cell fate by regulating the expression of et al. Cancer immunology. Mutational landscape determines sensitivity transcription factors T-bet and Eomesodermin. Immunity 2010;32: to PD-1 blockade in non–small cell lung cancer. Science 2015;348: Cancer Res; 76(2) January 15, 2016 Published OnlineFirst December 4, 2015; DOI: 10.1158/0008-5472.CAN-14-3362 Control of PD-L1 Expression by Oncogenic Activation of the AKT
mTOR Pathway in Non Small Cell Lung Cancer
Kristin J. Lastwika, Willie Wilson III, Qing Kay Li, et al. 2016;76:227-238. Published OnlineFirst December 4, 2015.
Cancer Res Access the most recent version of this article at: Access the most recent supplemental material at: This article cites 51 articles, 22 of which you can access for free at: Reprints and
To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at To request permission to re-use all or part of this article, contact the AACR Publications Department at

Source: https://www.roswellpark.edu/sites/default/files/cancer_res-2016-lastwika-227-38.pdf