• nonsmall cell lung carcinoma;
  • mammalian target of rapamycin pathway;
  • activated kinases;
  • prognosis;
  • protein expression profiles


  1. Top of page
  2. Abstract


Dysregulation of the mammalian target of rapamycin (mTOR) pathway has been shown to contribute to tumorigenesis. This study explored protein expression profiles of mTOR pathway and the relationship with prognosis in patients with nonsmall cell lung carcinoma (NSCLC).


The protein expression profiles of mTOR/phosphorylated (p-)mTOR, phosphoinositide-dependent kinase 1 (PDK1)/p-PDK1, p-Akt1, and P70 ribosomal protein S6 kinase (P70S6K)/p-P70S6K were determined via immunohistochemical staining assay. The clinical prognostic values of both single and combined protein expression were investigated with univariate and multivariate survival analysis.


Compared with normal lung tissues, the protein levels of mTOR/p-mTOR, p-Akt1 Ser473/Thr308, and P70S6K/p-P70S6K were higher (all P < .05), whereas p-PDK1 was lower (P < .05) in tumor tissues. p-mTOR expression was associated with histological differentiation, histological type, lymph node invasion, and stage (all P < .05). Overall survival in NSCLC patients was significantly shorter in cases with positive phenotype for p-mTOR, p-PDK1, and p-P70S6K (all P < .05). Subjects with coexpression of any 2 of p-mTOR, p-PDK1, p-Akt1 Ser473, and p-P70S6K demonstrated worse prognosis than those expressing no biomarker or any 1 biomarker alone (all P < .05). Multivariate analysis showed that the combination of p-mTOR/p-P70S6K is an independent prognostic factor in addition to tumor stage.


This study provides clinical evidence that activated components of mTOR pathway, not total protein, are predictors of poor prognosis in NSCLC. Moreover, evaluating protein-expression profiles of these molecules might be a new strategy for individual therapy in subjects with NSCLC. Cancer 2011;. © 2011 American Cancer Society.

Lung cancer is the leading cause of cancer-related deaths worldwide; nonsmall cell lung carcinoma (NSCLC) is the most common form. Despite recent advances in understanding the biology of lung cancer and the introduction of new chemotherapeutic agents for its treatment, the 5-year survival rate for patients with lung cancer is <15%.1 Recently, progression in the understanding of oncogenic kinase signaling pathways has provided better targets for developing effective therapeutic strategies,2 which may help to improve clinical outcome.

A potential candidate is the mammalian target of rapamycin (mTOR) pathway, which plays a central role in regulating cellular functions, including proliferation, growth, survival, mobility, and angiogenesis.3, 4 Dysregulation of mTOR pathway has been reported in many types of cancer, such as breast,5 prostate,6 and liver cancer.7 mTOR, a Ser/Thr protein kinase, is activated by an upstream molecule Akt1. On stimulation, Akt1 is phosphorylated at the Thr308 site by phosphoinositide-dependent kinase 1 (PDK1) at the cell membrane, then is phosphorylated at Ser473 for its activation.8 Once fully activated, Akt1 phosphorylates mTOR at Ser2448, which initiates series of cellular physiologic processes, including the regulation of cell growth through activating P70 ribosomal protein S6 kinase (P70S6K).9 Up-regulation of P70S6K also contributes to tumorigenesis.10 The full activation of P70S6K is controlled by multiple phosphorylation events localized at the Thr421, Ser424, Thr229, and Thr389 sites.11, 12 Phosphorylation at Thr421/Ser424 is the initial event and is necessary for P70S6K sufficient activation.13

mTOR pathway is constitutively active in NSCLC and promotes resistance to chemotherapy and radiation therapy in vitro.14 Their potential roles in NSCLC progression provide attractive targets for anticancer therapy. Several inhibitors of the pathway, such as CCI-779 (temsirolimus) or RAD001 (everolimus), are being developed and tested in clinical trials.15, 16 However, results of early clinical trials have shown that mTOR inhibitors have a limited efficacy as single agents in NSCLC17, 18 as well as other tumors, such as soft tissue sarcoma, uterine carcinoma, and cervical carcinoma.19 The insufficiency of therapeutic response may be caused by nonspecifically selected patients for mTOR inhibitor treatment. Therefore, it is very important to explore more clinical data on mTOR pathway to devise better treatment strategies for NSCLC. Here, via immunohistochemical staining assay, we analyzed the clinical prognostic value of mTOR pathway in NSCLC, including PDK1, Akt1, mTOR, and P70S6K. Our data provide clinical evidence that activated components of mTOR pathway, not total protein, are predictors of poor prognosis in NSCLC and suggest that evaluating protein-expression profiles of these molecules might be a new strategy for individual therapy in subjects with NSCLCs.


  1. Top of page
  2. Abstract

Tissue Collection

One hundred seventy-two surgically resected primary NSCLC cases, during the period from January 2002 to December 2004, were obtained from the archives of the Pathology Department of West China Hospital, Sichuan University. The tissue specimens consisted of 75 squamous cell carcinomas (SCC), 77 adenocarcinomas, and 20 other types, including large-cell carcinomas, adenosquamous carcinomas, and carcinoids, along with 56 adjacent normal pulmonary tissue specimens in the same blocks. Twenty normal lung specimens underwent surgery because of bronchiectasis (6 cases), tuberculosis (9 cases), and benign tumors (5 cases) during the same period as normal controls. All the tissues were fixed in 10% formalin immediately and embedded with paraffin within 12 to 24 hours postresection. Data on stage were according to the International Union Against Cancer's tumor-node-metastasis system, and differentiation and histological type were according to the World Health Organization classification for NSCLC.20, 21 All patients were adjuvant therapy-free before surgical resection and underwent standard therapeutic procedure after surgical resection according to the Clinical Oncology Information Network guidelines for nonsurgical management of lung cancer22; the median follow-up time was 37.25 months (range, 0-86.5 months). Because of inadequate paraffin-embedded fixed tissue blocks, only 136 cases were included for evaluating all the biomarkers. Institutional review board approval for this study was obtained from West China Hospital.

Antibody Preparation

Primary antibodies were selected to represent key elements of the mTOR signaling pathway. Antibodies used were as follows: mTOR (7C10, 1:50 dilution, #2983, Cell Signaling Technology, Beverly, Mass), phosphorylated (p-)mTOR (Ser2448, 49F9, 1:100 dilution, #2976, Cell Signaling Technology), PDK1 (1:300 dilution, #21,005, Signalway Antibody, Pearland, Tex), p-PDK1 (Ser241, 1:300 dilution, #11,005, Signalway Antibody), p-Akt1 (Ser473, 736E11, 1:50 dilution, #3787, Cell Signaling Technology; p-Akt1 Thr308, 1:50 dilution, clone 244F9H2, #9266, Cell Signaling Technology), P70S6K (49D7, 1:60 dilution, #2708, Cell Signaling Technology), and p-P70S6K (Ser424, 1:300 dilution, #11,284, Signalway Antibody).

Immunohistochemical Staining

Before immunostaining, deparaffinization and hydration for 4 μm paraffin sections were done in xylene and graded ethanol in distilled water. After hydration, a 15-minute blocking for endogenous peroxides was done in 3% H2O2 in 100% methanol. Antigen retrieval was done by heating Tris/ethylenediaminetetraacetic acid retrieval solution (pH 9.0, Dako Cytomation, Carpinter, Calif) at 95°C for 30 minutes in water bath. Slides were blocked with 3% bovine serum albumin for 30 minutes at 37°C and then incubated overnight at 4°C with the specific primary antibodies. The next day, sections were incubated with peroxidase-labeled polymer conjugated to goat antirabbit antibody (EnVsion+ Symstem, K4007, Dako) for 30 minutes, and developed in 3,3-diaminobenzidine for 3 to 5 minutes. Finally, the sections were counterstained with Harris hematoxylin, dehydrated and cleared in xylene, and mounted.

Immunohistochemical Scoring

Evaluation of sections was carried out by 2 pathologists (Drs. D. N. Liang and J. B. Yu) without the knowledge of clinical information as previous described.23, 24 Expression of each marker was assessed semiquantitatively according to criteria that accounted for both the fraction and the intensity of immunostaining of tumor cells involved. The fraction score was defined as the average of 5 randomly selected fields by light microscope: 0, no tumor cell stained; 1, <20% of cells stained; 2, 20% to 50% of cells stained; and 3, >50% of cells stained. The intensity score was defined as follows: 0, no appreciable staining in the tumor cells; 1, barely detectable staining in the cytoplasm and/or nucleus compared with the stromal elements; 2, readily appreciable brown staining; and 3, dark brown staining in tumor cells obscuring the cytoplasm and/or nucleus. The total score was calculated by multiplying the fraction score and the intensity score, producing a total range from 0 to 9. For statistical analysis, total scores of 0 to 2 were considered negative/low expression, and 2 to 9 positive/high expression.

Statistical Analysis

Pearson chi-square test was performed to determine the association of clinical characteristics with status of protein expression. Spearman rank test was used to assess the correlation between protein phenotypes. The Kaplan-Meier method was used to estimate univariate survival. The log-rank test and univariate Cox regression analysis were used to compare survival distributions between positive and negative staining groups. Independent prognostic factors of survival were identified with a multivariate Cox regression analysis. Only markers that were significant predictors in the univariate analysis were included in the multivariate analysis. For factors with only 2 levels, the first was compared with the second. P < .05 (2-sided) was considered to be statistically significant. Data analysis and summarization were conducted using SPSS 13.0 for Windows (SPSS Inc., Chicago, Ill).


  1. Top of page
  2. Abstract

Association of Clinical Characteristics With Survival

A cohort in our study consisted of 172 patients, including 138 men and 34 women with an average age of 60.4 years. The 5-year survival rate for all these patients was 37.5%. Characteristics of the patients are summarized in Table 1.

Table 1. Relation Between Clinical Variables and Survival
VariableCategoryAvailable No. for Survival (N=172)Median Survival, moLog-rank P
  • ADC indicates adenocarcinoma; SCC, squamous cell carcinoma.

  • a

    Statistically significant.

Age, y<608137.708
Histological typeADC7731.5.291
Moderate or well10037.5
Tumor size≤3 cm1866.4.000a
>3 cm15437
Lymph node invasionN0 + N110752.000a
N2 + N36516
Distant metastasisM016638.000a
StageI + II8561.2.000a
III + IV8717

Expression of mTOR Pathway-Related Proteins in NSCLC and Normal Control Groups

Representative expression patterns of mTOR pathway were shown in Figure 1. PDK1/p-PDK1, P70S6K/p-P70S6K, p-Akt1 Ser473, and p-Akt Thr308 expressions were detected in both cytoplasm and nuclei. mTOR and p-mTOR expressions were predominately in cytoplasm, although several membrane-staining specimens were also observed. The results are consistent with previous observations.6, 25

thumbnail image

Figure 1. Representative examples of protein expression profiles for mTOR, p-mTOR, PDK1, p-PDK1, p-Akt1 Ser473/Thr308, P70S6K, and p-P70S6K (from left column) of normal lung (A) and nonsmall cell lung carcinoma (B) specimens are shown. (B) From the first to the third line, immunohistochemical staining of scores of 0 to 2, 3 to 5, and 6 to 9, respectively, are shown for each kinase. Original magnification, ×400.

Download figure to PowerPoint

Comparisons of mTOR pathway-related protein expressions in NSCLC with normal controls are shown in Table 2. Protein levels of mTOR/p-mTOR, P70S6K/p-P70S6K, and p-Akt1 Ser473/Thr308 were significantly increased (all P < .05), whereas p-PDK1 was decreased (P < .05) in NSCLC compared with normal control.

Table 2. Expression of mTOR Pathway-Related Proteins in NSCLC and Normal Control Groups
ProteinP/NNSCLC, No. (%)Lung Tissue, No. (%)P
  • mTOR indicates mammalian target of rapamycin; NSCLC, nonsmall cell lung carcinoma; P/N, positive/negative; PDK1, phosphoinositide-dependent kinase 1; p-, phosphylated.

  • a

    Statistically significant.

PDK1P78 (54.9)49 (64.5).173a
N64 (45.1)27 (35.5)
p-PDK1P70 (50)53 (69.7).005a
N70 (50)23 (30.3)
p-Akt1 Ser473P118 (74.7)17 (22.4).000a
N40 (25.3)59 (77.6)
p-Akt1 Thr308P103 (73.6)31 (40.8).000a
N37 (26.4)45 (59.2)
mTORP106 (79.1)43 (56.6).001a
N28 (20.9)33 (43.4)
p-mTORP89 (51.7)28 (37.3).037a
N83 (48.3)47 (62.7)
P70S6KP92 (66.7)30 (39.5).000a
N46 (33.3)46 (60.5)
p-P70S6KP100 (70.4)26 (34.2).000a
N42 (29.6)50 (65.8)

Relationship Between Clinical Characteristics and Protein Expression Profiles

The univariate analysis was performed to estimate the relationship between clinical variables and protein phenotypes (Table 3). The expressions of p-mTOR, PDK1, and p-Akt1 Ser473/Thr308 were associated with histological differentiation grade (P = .025, P = .025, P = .017 and P = .018, respectively). Overexpression of p-mTOR was also related to lymph node invasion, distant metastasis, and tumor stage (P = .008, P = .029 and P = .015, respectively). p-mTOR expression was significantly increased in adenocarcinoma compared with SCC and the other cancer types (P < .001).

Table 3. Relation Between Clinical Characteristics and Protein Expression Profile
AntibodyP/NHistological TypeHistological DifferentiationTumor SizeLymph Node InvasionDistant MetastasisStage
ADCSCCOtherPPoorModerate/WellP≤3 cm>3 cmPN0+N1N2+N3PM0M1PI+IIIII+IVP
  • P/N indicates positive/negative; ADC, adenocarcinoma; SCC, squamous cell carcinoma.

  • a

    Statistically significant.

(n = 134)N10126 1216 424 1810 262 1216 
(n = 172)N244415 4241 677 6023 830 4934 
(n = 142)N33283 2539 559 4123 622 3232 
(n = 140)N28357 2644 1060 5317 700 4228 
p-Akt Ser473P584911.8053781.017a16102.1417642.1261144.6436256.587
(n = 158)N19165 2119 238 319 382 2317 
p-Akt Thr308P395014.1344558.018a1390.1456835.8641003.6085548.945
(n = 140)N21124 829 532 2512 352 2017 
(n = 138)N24166 1036 541 3016 442 2521 
(n = 142)N20148 1230 537 3111 411 2121 

Comparison of Survival Among Different Status of Single and Combined Proteins Expression

As shown in Figure 2 and Table 4, subjects with positive expression of p-PDK1, p-mTOR, and p-P70S6K had shorter survival time (log-rank, P = .001, P = .000, and P = .004, respectively), compared with subjects with negative expression, whereas subjects with PDK1, mTOR, and P70S6K positive expression did not (P > .05). P70S6K expression had a trend toward shorter survival. In contrast, mTOR expression conferred a favorable survival trend, although the difference did not reach statistical significance (log-rank, P = .063 and P = .104, respectively).

thumbnail image

Figure 2. Relations between protein expression and survival are shown. Kaplan-Meier curves of proteins compare positive (red lines) with negative phenotype (blue lines).

Download figure to PowerPoint

Table 4. Relation Between Protein Expression and Survival (Univariate Analysis): Uniprotein Profiles
ProteinP/NHR95% CIP
  • P/N indicates positive/negative; HR, hazard ratio; CI, confidence interval.

  • a

    Statistically significance.

PDK1P vs N1.1480.774-1.704.490
p-PDK1P vs N2.6621.431-4.952.001a
mTORP vs N0.6450.377-1.103.104
p-mTORP vs N1.9171.349-2.724.000a
p-Akt Ser473P vs N0.9710.656-1.435.881
p-Akt Thr308P vs N0.8370.521-1.345.459
P70S6KP vs N1.6540.964-2.838.063
p-P70S6KP vs N1.9171.217-3.019.004a

p-Akt1 Ser473 or p-Akt1 Thr308 expression alone did not significantly influence survival (log-rank, P = .881 and P = .459). Subjects with coexpression of p-Akt1 Ser473 and p-Akt1 Thr308 also had no difference of survival time compared with the other 2 combinations, including p-Akt Ser473(−) and/or p-Akt Thr308(−) (log-rank, P = .699) (data not shown).

In addition, we evaluated the prognostic value, combining different activated proteins using Kaplan-Meier analysis (Fig. 3 and Table 5). Subjects with coexpression of p-mTOR and p-PDK1 had significantly worse prognosis than the other 2 combinations, p-mTOR (−) and/or p-PDK (−) (log-rank, P < .001, Fig. 3A). The survival time was also substantially shortened for subjects with expression of both p-mTOR and p-P70S6K, with a median survival of 17 months versus 61.67 months (log-rank, P < .001, Fig. 3B). Likewise, coexpression of p-PDK1 and p-P70S6K significantly reduced the survival time, with a median survival of 31.2 months (log-rank, P < .001, Fig. 3C). It is noteworthy that combining p-Akt1 Ser473 expression with p-PDK1, p-mTOR, or p-P70S6K showed poor prognostic significance for survival, although its expression alone did not (log-rank, all P < .05, Fig. 3D-F). Subjects with p-PDK (+)/p-mTOR (+)/p-AKT1 Ser473 (+)/p-P70S6K (+) had the worst prognosis in all, with a median survival of 15.20 months (log-rank, P < .001 Fig. 3G).

thumbnail image

Figure 3. Relations between different combined protein profiles and survival are shown. Kaplan-Meier curves compare both positive patterns (red lines) with either or neither negative patterns of proteins (blue lines).

Download figure to PowerPoint

Table 5. Relation Between Protein Expression and Survival (Univariate Analysis): Combined Protein Profiles
Combined Protein PatternsP/OHR95% CIP
  1. P/N indicates positive/negative; P/O, all positive expression combinations vs other expression combinations including either or neither negatives; HR, hazard ratio; CI, confidence interval.

p-mTOR/p-PDK1P vs O2.4991.580-3.953.000
p-mTOR/p-P70S6KP vs O3.9102.453-6.234.000
p-PDK1/p-P70S6KP vs O2.5751.622-4.089.000
p-mTOR/p-Akt1 Ser473P vs O2.3161.457-3.680.000
p-PDK1/p-Akt1 Ser473P vs O1.6260.966-2.656.048
p-P70S6K/p-Akt1 Ser473P vs O1.7301.077-2.780.021
p-Akt1 Ser473/p-PDK1/ p-mTOR/p-P70S6KP vs O3.6212.230-5.877.000

Multivariate Cox Regression Analysis

By using multivariate Cox regression models, we evaluated whether expression of p-PDK1, p-mTOR, and p-P70S6K could have prognostic value in the assessment of NSCLC. Variables included in the model were those that have been previously shown to be influential, which were sex, tumor size, lymph node invasion, distant metastasis, tumor stage, p-PDK1, p-mTOR, and p-P70S6K (Table 6). The analysis revealed that tumor stage, p-PDK1, p-mTOR, and p-P70S6K were independent prognostic factors for NSCLC development (log-rank, all P < .05).

Table 6. Multivariate Survival Analysis (Cox Regression Model): Univariate Model (Method: Forward LR)
VariantsHR95% CIP
  • LR indicates likelihood ratio; HR, hazard ratio; CI, confidence interval.

  • a

    Statistically significant.

Stage (III + IV vs I + II)4.9172.885-8.378.000a
p-PDK1 (P vs N)2.0591.035-4.096.040a
p-mTOR (P vs N)3.2991.928-5.645.000a
p-P70S6K (P vs N)2.9841.722-5.170.000a

Further Cox regression analysis to assess the potential of combined protein patterns previously described is shown in Table 7. In addition to stage, the combination of p-mTOR/p-p70S6k was a powerful and independent predictor in NSCLC (P < .001).

Table 7. Multivariate Survival Analysis (Cox Regression Model): Composite Covariate Model (Method: Forward LR)
VariantsHR95% CIP
  • LR indicates likelihood ratio; HR, hazard ratio; CI, confidence interval; P, all positive expression combinations; O, other combinations, including either or neither negatives; PDK1, phosphoinositide-dependent kinase 1.

  • a

    Statistically significant.

Stage (III + IV vs I + II)5.4993.193-9.470.000a
p-mTOR/p-P70S6K (P vs O)6.3003.682-10.780.000a
p-PDK1/p-P70S6K (P vs O)2.5441.118-5.789.026a

Correlations Between Phosphorylated Kinases

The correlations between phosphorylated kinases were explored using Spearman rank test as summarized in Table 8. Expression of p-mTOR significantly correlated with expression of p-PDK1 and p-Akt1 Ser473 (r = 0.157, P = .033 and r = 0.252, P < .001, respectively). p-PDK1 expression was also associated with p-Akt1 Ser473 (r = 0.220, P = .003). p-Akt Thr308 expression was correlated with p-Akt Ser473 and p-PDK1 (r = 0.261, P = .000; r = 0.248, P = .001, respectively). The results suggest that these proteins are involved in the same pathway. Conversely, p-P70S6k expression was not associated with the other kinases investigated except P70S6K. We further explored the correlation between total P70S6K and all of the phosphorylated proteins. The data showed that total P70S6K expression correlated with p-P70S6K, p-mTOR, p-PDK1, and p-Akt1 Ser473 (r = 0.187, P = .012; r = 0.239, P = .001; r = 0.189, P = .011; r = 0.157, P = .032, respectively), indicating that protein level, not activation of P70S6K, is related to PDK1-Akt1-mTOR pathway in NSCLC.

Table 8. Correlations Between Phosphorylated Proteins
AntibodyAntibodySpearman Coefficients (r)P
  • a

    Statistically significant.

 p-Akt1 Ser4730.252.000a
 p-Akt1 Thr3080.032.679
p-PDK1p-Akt1 Ser4730.220.003a
 p-Akt1 Thr3080.248.001a
p-Akt1 Ser473p-P70S6K0.057.446
 p-Akt Thr3080.261.000a
p-Akt Thr308p-P70S6K0.094.227
 p-Akt1 Ser4730.157.032a
 p-Akt1 Thr3080.099.201


  1. Top of page
  2. Abstract

The activation of mTOR pathway occurs through a complex signaling cascade during tumorigenesis. This study investigated associations between the components of mTOR pathway and their significance in the clinical outcome of NSCLC patients. We found that among all mTOR pathway components, except for PDK1/p-PDK1, all total proteins and their phosphorylated forms expressed more frequently in NSCLC compared with normal lung tissues. p-PDK1, p-mTOR, and p-P70S6K were significantly correlated with survival. In addition, combinations of kinases were powerful prognostic factors.

We also evaluated prognostic values of high-risk factors, including past history of cancer, coexisting illness, and smoking history, and their associations with protein expressions. No correlation was found between survival and the factors aforementioned by univariate and multivariate survival analysis. However, past history of cancer and smoking history were associated with mTOR pathway expression (data not shown). These suggested that mTOR pathway might be involved in chronic inflammation and smoking-related tumorigenesis.

In our findings, p-mTOR expression had more clinical significance in NSCLC, but not total mTOR. The rate of p-mTOR expression was more prominent in advanced stage and lymph node invasion subjects with NSCLC, suggesting that p-mTOR particularly contributed to NSCLC progression. This result was also found in breast,2 prostate,26 and colorectal cancer.27 We also found that phosphorylation of mTOR was more frequently activated in adenocarcinoma than in other tumor types, which could have implications for tumorigenesis and the use of targeted therapies. The higher prevalence of active mTOR in adenocarcinoma may underlie the selective response of patients with adenocarcinoma to mTOR inhibitors. More importantly, p-mTOR status was significantly correlated with worse prognosis and was an independent, strong predictor of survival in NSCLC, which is in keeping with the findings of Yoshizawa et al and Dhillon et al.24, 28 All these indicated that p-mTOR, as the active form of mTOR, plays a crucial role in NSCLC progression. However, as opposed to p-mTOR, the expression of mTOR showed a trend toward better outcome, although the difference did not reach statistical significance. This result was also found by Anagnostou et al in early stage lung adenocarcinoma.29 This indicated that the functions of total mTOR are more complex than just translational control, and crosstalk between pathways could alter its oncogenic potential in NSCLC. To date, very limited studies were reported to assess the prognostic value of total mTOR expression in NSCLC. This result remained to be confirmed in an expanded and intensive study. High p-mTOR expression could be used not only as a reliable indicator of worse prognosis but also an inclusion criterion in clinical trials of mTOR inhibitors. In future investigations of cancer treatment, phosphorylation of mTOR, not total mTOR expression, should be examined as a potential target in anticancer therapy.

In the present study, p-PDK1 level was lower in NSCLC than those of normal lung tissues. However, PDK1 only showed a trend toward lower expression (P = .173). This was similar to Koukourakis's30 report that both PDK1 and its downstream pyruvate dehydrogenase (PDH) levels were reduced in lung carcinoma cells. We also found that the upstream molecules p-Akt1 Ser473 and p-PDK1 were indeed correlated with p-mTOR, confirming the activation cascade of the PDK1/Akt1/mTOR pathway in NSCLC.31 PDK1, as the gate of mTOR pathway activation, is frequently phosphorylated and activated in many types of cancer.32 Ser241 lies in the activation loop of PDK1 kinase, and could phosphorylate itself, leading to PDK1 activation. p-PDK1, but not total PDK1, confers worse prognosis in subjects with NSCLC. The findings suggest that activation of PDK1 contribute to tumorigenesis in NSCLC, and phosphorylation at Ser241 is crucial for the activity of PDK1, which has also been confirmed in vitro by Sato et al.33 Our results provide clinical evidence for targeting p-PDK1 as a novel approach to cancer intervention in NSCLC. PDK1 has been known to phosphorylate Akt1 at the Thr308 site; however, the identity of kinase responsible for phosphorylation of Ser473 is controversial. In this study, we find that p-PDK1 expression is associated with p-Akt1 at Ser473, suggesting that p-PDK1 is involved in Ser473 phosphorylation and inconsistent with the findings by Balendran et al34 that in the presence of PRK2 (PKC-related kinase 2) or PIF (PDK1-interacting fragment), PDK1 phosphorylated Akt1 not only at Thr308 but also at Ser473. Chan and Tsichlis35 also report that the phosphorylation of Akt1 at Thr308 by PDK1 could be caused by the removal of an inhibitory effect exerted by PDK1 on Ser473 phosphorylation. All these suggest that PDK1 contributes to the phosphorylation of Akt1 at the Ser473 site. However, the results derived from PDK1 knockout embryonic stem cells indicated that PDK1 was not necessary for phosphorylation of Akt1 at Ser473.36 Whether PDK1 is involved in the process needs further elucidation.

Akt1, as an important component of mTOR pathway, has been reported to be correlated with poor outcome in several cancers.37 Evaluation of both Thr308 and Ser473 phosphorylation is a better surrogate for Akt activity, and is more likely to yield important clinical correlations.38 Thr308 phosphorylation is necessary and sufficient for Akt activation; however, maximal activation requires additional phosphorylation at Ser473.39 In our study, expression of p-Akt1Thr308 or p-Akt1 Ser473 alone or their coexpression did not correlate with survival. In contrast, Tsurutani et al demonstrated that being positive for both p-Akt Thr308 and p-Akt Ser473 was associated with worse survival in patients with NSCLC (P = .041), but not Thr308 or Ser473 alone.38 Interestingly, they also reported that the sites of phosphorylation are not prognostic for bronchioloalveolar carcinoma patients, either combined or individually.40 Whether evaluation of 2 phosphorylation sites (Thr308 and Ser473) can improve the prognostic significance of Akt activation needs to be further studied. Furthermore, in our study, the combinations of p-Akt1 Ser473 with p-PDK1, p-mTOR, or p-P70S6K were all predictors of shortened survival in NSCLC; however, p-Akt Thr308 or p-Akt Ser473/Thr308 combined with others were not (data not shown). These suggested that p-Akt1 Ser473 might be more valuable for evaluating prognosis than p-Akt Thr308.

P70S6K, as the main indicator of mTOR pathway activity,41 has been found to promote carcinogenesis in various types of cancer.42 Phosphorylation of Ser424 is an initial and necessary step for its activation. Our results showed that p-P70S6K at Ser424 was an independent, strong prognostic factor, whereas total P70S6K only showed a trend toward worse survival of NSCLC. The data indicated that activated P70S6K plays a critical role in NSCLC development, and phosphorylation of P70S6K at Ser424 is essential for P70S6K effect. According to previous studies, mTOR activated P70S6K through a series of phosphorylating processes to regulate cellular physiological function.43 However, the correlations between p-P70S6K at Ser424 and p-mTOR, p-PDK1, or p-Akt1 Ser473 were not observed, although total protein P70S6K expression is indeed correlated with them in our data. These findings indicated that there could be additional signaling pathways to regulate P70S6K activation independent of mTOR pathway. The results are consistent with the reports by Thomas et al and Lam et al that several mTOR-independent pathways also seem to directly phosphorylate P70S6K, such as PDK1 and ERK pathway.44, 45 The precise mechanism of Ser424 phosphorylation in vivo is still under investigation. It has been shown by Lehman et al that Ser424 phosphorylation is mainly regulated by ERK in non-neuronal cells.46 In contrast, in another study, coactivation of mTOR and ERK signaling pathways may be required to regulate P70S6K phosphorylation at Ser424 in BDNF (brain-derived neurotrophic factor)-stimulated neurons.11 Therefore, p-P70S6K at Ser424 could be less reliable as a surrogate marker for mTOR activation, and mTOR inhibitor could fail to block p-P70S6K tumorigenic function sufficiently. Although the exact mechanism of phosphorylation of P70S6K at Ser424 still needs intensive investigation, the prognostic significance of P70S6K is not negligible. Importantly, multivariate Cox regression analysis showed that in addition to tumor stage, the combination of p-mTOR and p-P70S6K was the strongest independent prognostic factor. The results also suggest that multitargeted interventions may be more effective for tumor therapy, compared with single mTOR inhibitor treatment.

It is noteworthy that combinations of these kinases are of additional prognostic value compared with a single kinase. Despite the limited number of valid cases, the simultaneous expressions of 2 or 3 proteins strongly correlated with poor survival. In addition, our study highlighted the value of the combination of p-PDK (+)/p-mTOR (+)/p-AKT1 Ser473 (+)/p-P70S6K (+), pivotal molecules in the mTOR signaling axis, which showed the worst prognosis in all, as comarkers for evaluation of NSCLC.

In conclusion, the dysregulation of mTOR pathway plays a crucial role in the development in NSCLC. mTOR pathway is highly associated with clinical parameters, especially with survival. These also provide novel targets for cancer intervention, such as p-PDK1 and p-P70S6K. Combined p-P70S6K and p-mTOR represents a more valuable predictor for prognosis. Moreover, by evaluating the status of molecular predictors, we could predict tumor response to mTOR inhibitor treatments, thereby selecting appropriate therapy strategy for individual patients.


  1. Top of page
  2. Abstract

Supported by grants from the National Basic Research Program of China (2009CB941200) and Nature Science Foundation of China (30771228 to H.X.; 30771227 to X.M.).


  1. Top of page
  2. Abstract