Human Pathology (2014) 45, 127–136
www.elsevier.com/locate/humpath
Original contribution
Significance of Akt activation and AKT gene increases in soft tissue tumors☆,☆☆ Yoh Dobashi MD, PhD a,⁎, Eiichi Sato MD, PhD b , Yoshinao Oda MD, PhD c , Johji Inazawa MD, PhD d , Akishi Ooi MD, PhD e a
Department of Pathology, Jichi Medical University, Saitama, Japan Department of Orthopedic Surgery, Interdisciplinary Graduate School of Medicine and Engineering, University of Yamanashi, Chuo, Japan c Department of Anatomic Pathology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan d Department of Molecular Cytogenetics, Medical Research Institute, Tokyo Medical and Dental University, Tokyo, Japan e Department of Molecular and Cellular Pathology, Kanazawa University Graduate School of Medical Science, Kanazawa, Japan b
Received 27 May 2013; revised 20 June 2013; accepted 26 June 2013
Keywords: Bone and soft tissue; AKT; Polysomy; Activation; Isoforms; Metastasis; Molecular targeting therapy
Summary To clarify the aberrations of AKT genes, their protein products and clinicopathologic significance in bone and soft tissue tumors, expression profiles of total Akt, its isoforms and activated Akt, and increases in copy number of AKT1/AKT2 genes were examined. Immunohistochemical analysis in 77 cases revealed overexpression of total Akt, Akt1, Akt2, and phosphorylated Akt in 84.4%, 67.5%, 72.7%, and 71.4%, respectively. Positive results were also observed in benign lesions but at a lower frequency. Overexpression of Akt1 was more frequent than that of Akt2 in well-differentiated liposarcoma (6/7 versus 3/7 cases) and schwannoma (4/4 versus 1/4 cases), whereas Akt2 overexpression and Akt activation were more frequent than Akt1 overexpression in malignant nerve sheath (3/4 and 4/4, respectively, versus 2/4 cases) and muscular tumors (8/9 and 8/9 versus 4/9 cases). By fluorescence in situ hybridization analysis, increase of gene copy number was observed in 13.3% for AKT1 and in 25.0% for AKT2 due to polysomy of chromosome 14 or 19, respectively, but not gene amplification. One case of schwannoma exhibited polysomy of both chromosomes 14 and 19. Akt activation was correlated with total Akt cytoplasmic localization (P = .0031) and subsequent metastasis (P = .0454). Moreover, AKT2 gene increase correlated with tumor size (P = .0352) and metastasis (P = .0344). In conclusion, in a defined subset of bone and soft tissue tumors, including benign tumors, Akt was frequently overexpressed and activated, and AKT1/2 copy number was increased. Because abnormality of Akt/AKT correlated with clinicopathologic profiles, novel therapies targeting isoformspecific Akts may be useful for these particular types of tumors. © 2014 Elsevier Inc. All rights reserved.
☆
No financial disclosure or conflicts of interests are declared. Funding support: This work was partially supported by a grant from the Japan Society for the Promotion of Science C23590409 (Y. D.) and C22590310 (A. O.) and Smoking Research Foundation (Y. D). ⁎ Corresponding author. Department of Pathology Saitama Medical Center, Jichi Medical University, 1-847, Amanuma, Omiya, Saitama, 3308503, Japan. E-mail address: [email protected] (Y. Dobashi). ☆☆
0046-8177/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.humpath.2013.06.024
1. Introduction Malignant bone and soft tissue tumors currently represent approximately 1% of adult malignancies and 15% to 20% of pediatric malignancies [1]. These are heterogeneous
128 neoplasms consisting of more than 50 subtypes [2]. Although the overall survival rates average 60% to 70%, they are far lower in some subtypes, due primarily to limited therapy options and the development of distant metastases [1,3]. Therefore, there is much interest in supplementing conventional chemotherapy with novel “tailored therapies” that target signaling molecules critical for each tumor. Although an increasing number of such promising novel therapies have been developed, progress against soft tissue tumors has been slow due to their rarity compared with carcinomas. Nevertheless, there is some promise for tailored therapies, especially given the progress in targeting growth factor receptors and critical effector molecules downstream of them [3,4]. Among the myriad of growth-promoting signaling cascades, the phospho-inositol-3 kinase (PI3K)/Akt pathway is especially critical for both normal and pathologic processes [5,6]. In particular, Akt family proteins function at the crossroads of multiple intracellular pathways and can potentially influence thousands of downstream targets [4-6]. Akt activities are frequently dysregulated in human cancers, and there are numerous reports describing the correlations between Akt activation and clinicopathologic profiles [4,5,7]. Recently, Akt pathway activation was shown to be correlated with deep location, high histologic grade, a higher probability of metastasis, and worse prognosis in soft tissue tumors [1,8,9]. This would indicate that Akt may be a candidate for targeted therapies against soft tissue tumors. Akt family proteins are encoded by 3 related genes (AKT1-3), located on chromosomes 14q32, 19q13, and 1q44, respectively [6,10]. Their gene products, Akt1-3, are 56-kd serine/threonine kinases that display distinct tissue and subcellular distributions [6]. In response to upstream signals, they translocate to specific subcellular compartments where they regulate cellular functions such as cell metabolism, growth, and survival [5,6,9]. Although Akt1 is expressed ubiquitously at a high level, Akt2 is elevated in insulinresponsive tissues such as fat, skeletal muscle, and the liver [1,5,6]. Although Akt1 and Akt2 typically require membrane localization for their activity, they can also localize to the nucleus where they function in cell differentiation and/or suppression of apoptosis [6,11]. In cancers, distinct changes in the expression of Akts have been observed. Akt1 is up-regulated and/or activated in gastric, prostatic, ovarian, and breast carcinomas, and this up-regulation is often associated with poor prognosis [4-6]. Akt2 is activated in hepatocellular, colorectal, ovarian, pancreatic, and breast carcinomas [4,5]. Moreover, both Akt1 and Akt2 are implicated in invasion and in metastatic processes [5]. One mechanism of Akt overexpression is gene amplification, and AKT1 amplification has been reported in sporadic cases of the lung, gastric, breast, and prostatic carcinomas [5,7,10]. AKT2 gene amplification has been frequently detected in breast, ovarian, and pancreatic carcinomas and is associated with a poor prognosis in ovarian carcinoma [5,10].
Y. Dobashi et al. In our previous study in lung carcinoma, fluorescence in situ hybridization (FISH) analysis revealed amplification of AKT1 and/or AKT2 in 6% and high-level polysomy in 7% of the cases [7]. Although the Akt-driven pathway may constitute a critical mechanism in the pathology of bone and soft tissue tumors, there are few studies to date that comprehensively describe the expression and activation status of Akt or alterations in the AKT genes. In addition, the relative contributions of the different Akt isoforms to bone and soft tissue tumors are not well understood. Herein, we conducted a comprehensive analysis of Akts and specifically asked if (1) total Akt (T-Akt) is constitutively overexpressed/ activated globally or in particular subtypes, (2) there is any preponderance in the expression of particular Akt isoforms and/or gene increase specific to histologic subtypes, and (3) Akt activation or gene increases correlate with clinicopathologic profiles.
2. Materials and methods 2.1. Cases and classification Under protocols approved by the Institutional Tissue Committees, 77 cases (65 malignant and 12 benign) were obtained from surgeries performed at the Department of Orthopedic Surgery, University of Yamanashi Hospital, and University of Kanazawa Hospital. Pathologic diagnoses were made according to World Health Organization classification [2]. Staging was determined in accordance with the American Joint Committee on Cancer (AJCC) staging system [12]. These details are presented in Tables 1 and 2. Cases were restricted to patients who had not received preoperative chemotherapy.
2.2. Immunohistochemistry Serial sections were stained with 4 primary antibodies followed by visualization with a CSAII kit (Catalyzed Signal Amplification System 2; Dako, Glostrup, Denmark), T-Akt (1:600; Cell Signaling Technology [CST], Beverly, MA), Akt1 (1:40; monoclonal, C73H10; CST), Akt2 (1:150, polyclonal; Abcam, Cambridge, UK), phosphorylated Akt (p-Akt) (p-AktSer473; 1:300, monoclonal; CST) [7]. The sensitivity and the specificity of the antibodies had been previously validated on cell lines and tissue specimens by immunohistochemistry (IHC) and immunoblotting [1,7,8,13]. IHC expression was evaluated by 2 observers (Y. D. and A. O.) and scored based on staining intensity and the fraction of positive cells. Significant staining was defined as staining that was more intense compared with endothelial cells [8,9]. IHC expression level was quantitatively evaluated by the fraction of stained tumor cells: negative, less than 10%; 1+, 10% or more and less than 50%; 2+, 50% or more cells with significant staining [7,8]. When discordance was noted in evaluation of the results between
AKTs in bone and soft tissue tumors Table 1
129
2.4. Statistical analysis
Clinicopathologic profiles of 77 cases
Clinicopathologic factors Histology Malignant Benign Sex Male Female Age (y), mean (range) b60 ≥60 Sarcoma only Diameter (pT) b5 cm (pT1) ≥5 cm (pT2) Recurrence + − Metastasis + − AJCC stage I II III IV
n 65 12 43 34 57 (33-81) 46 31
11 54 5 60 10 55
Agreement among observers in the interpretation of IHC results was evaluated by κ statistics [15]. Other statistical analysis was performed with the StatView package (version 5.0; SAS Inc., Cary, NC, USA). Differences in frequencies of positive immunostaining between 2 groups were analyzed by Fisher exact test. For correlations among more than 3 categories, Mann-Whitney U test or Fisher protected least significant difference (PLSD) analysis was applied. Log-rank analysis with the Kaplan-Meier method was used to determine the correlation between the variables and survival period via univariate analysis. In the multivariate analysis, a Cox proportional hazards regression analysis was performed. Two-sided P b .05 was placed to determine statistical significance.
3. Results 3.1. IHC
17 25 23 0
2 observers, it was resolved by discussion. When scores were classified into 2 groups for statistical analysis, 1+ and 2+ were combined as positive [7,14].
2.3. FISH analysis AKT1 and AKT2 gene copy aberrations were examined by FISH analysis, as described previously [7,14]. Samples consisted of 65 cases exhibiting T-Akt overexpression by IHC and 9 cases of sarcoma without T-Akt overexpression. For FISH probes, we used the bacterial artificial chromosome clones CTD-2507D9 (AKT1, 14q32.32) and CTB166E20 (AKT2, 19q13.1-13.2). For reference probes adjacent to the centromere, we used RP11-203 M5 (14q11.2) for AKT1 and RP11-124 K10 (19p13.11) for AKT2 [7]. The specificity and localization of both probes were previously validated [7]. FISH signals were examined in at least 30 nuclei for each tumor independently by 2 observers (Y. D. and A. O.). Results were classified into 4 strata based on AKT copy number using the previously established scoring system: 1, disomy (≤2 copies in N90% of cancer cells); 2, low-level polysomy (≥3 copies with an equivalent number of reference genes in N10% and b40%, without amplification); 3, highlevel polysomy (polysomy in ≥40%, without amplification); and 4, amplification (tight clusters, average AKT/chromosome ≥2, or ≥15 copies of AKT per cell in ≥10%) [7,14]. When discordance was noted in evaluation of the results between 2 observers, it was resolved by discussion.
Overall results of IHC are presented in Figs. 1 and 2 as well as Table 2. In normal tissues, T-Akt, Akt1, Akt2, and pAkt staining was observed in the cytoplasm of skeletal muscle, vascular smooth muscle, and weakly in occasional endothelial cells and lymphocytes. Interobserver agreement was “almost perfect” (T-Akt, κ = 0.86; Akt1, κ = 0.89; Akt2, κ = 0.88; p-Akt, κ = 0.91). 3.1.1. T-Akt Positive staining was observed in the nuclei and/or the cytoplasm in 65 (84.4%) of 77 cases, including 56 (86.2%) of 65 malignant and 9 (75.0%) of 12 benign lesions. More than 50% of the cases exhibited positive staining across all histologic types. Positive staining was more prominent in the cytoplasm than in the nucleus (59 versus 45 cases), and nuclear staining was more frequently found in liposarcoma (LS), malignant peripheral nerve sheath tumor (MPNST), chordoma, and others. 3.1.2. Akt1 Akt1 staining was observed in wide varieties of tumors in 52 (67.5%) of 77 cases, including 45 (69.2%) of 65 malignant and 7 (58.3%) of 12 benign lesions. Compared with Akt2, Akt1 expression was found more frequently in welldifferentiated LS (WD-LS, 6/7 versus 3/7 cases) and schwannoma (4/4 versus 1/4 cases). Akt1 expression was observed in 80.0% (52/65) of the cases positive for T-Akt and was more prominent in the nucleus compared with the cytoplasm (38 versus 30 cases). 3.1.3. Akt2 Akt2 staining was observed in 56 (72.7%) of 77 cases, including 50 (76.9%) of 65 malignant and 6 (50.0%) of 12 benign lesions. Compared with Akt1, Akt2 was more
130
Table 2
Overall results of immunohistochemical and FISH analyses
Histology
Malignant tumors Undifferentiated pleomorphic sarcoma Myxofibrosarcoma Liposarcoma Well differentiated Myxoid/round cell Dedifferentiated Synovial sarcoma MPNST Rhabdomyosarcoma Pleomorphic Embryonal Leiomyosarcoma DFSP Osteosarcoma Chondrosarcoma Grade I Grade II Chordoma Ewing sarcoma/PNET Total Benign tumors Schwannoma Dermatofibroma Leiomyoma Giant cell tumor Lipoma Total Total
Total cases
T-Akt
6
5 (2)
[3, 5]
4 (2)
[2, 4]
5 (1)
[4, 2]
5 (1)
6
4 (1)
[3, 3]
4 (1)
[3, 4]
4 (0)
[3, 1]
7 7 2 4 4
6 6 2 3 4
(1) (5) (0) (2) (3)
[6, 3] [6, 6] [2, 1] [2, 3] [3, 4]
3 5 2 3 3
(0) (4) (1) (0) (2)
[3, 3] [4, 5] [1, 1] [1, 2] [1, 3]
6 5 2 2 2
(0) (0) (0) (1) (1)
2 2 5 3 5
2 2 5 2 4
(1) (1) (3) (1) (2)
[1, 2] [0, 2] [2, 5] [2, 2] [2, 4]
2 2 4 2 4
(0) (0) (3) (0) (1)
[1, 1] [2, 0] [1, 4] [2, 2] [2, 4]
1 1 3 2 3
2 2 4 4 65
1 2 4 4 56
(0) (0) (0) (2) (24)
[0, 1] [1, 2] [4, 4] [2, 4] [39, 51]
1 2 4 3 48
(0) (0) (0) (1) (15)
[0, 1] [1, 2] [4, 4] [1, 3] [29, 43]
4 2 2 2 2 12 77
4 2 1 1 1 9 65
(2) (0) (0) (0) (0) (2) (26)
[3, 4] [2, 2] [0, 1] [1, 1] [0, 1] [6, 9] [45, 60]
3 (1) 2 (0) 1 (0) 1 (0) 0 (0) 7 (1) 55 (16)
[2, 3] [0, 2] [0, 1] [1, 1] [0, 0] [3, 7] [32, 50]
Positive (2+)
p-Akt [N, C]
Positive (2+)
Akt1 [N, C]
Positive (2+)
Akt2 [N, C]
Positive (2+)
AKT1 [N, C]
AKT2
D
P
D
P
[0, 5]
5
−
3
2 (L, H)
3 (1)
[0, 3]
4
−
3
1 (L)
[5, 1] [3, 4] [2, 1] [1, 1] [1, 2]
3 6 2 3 4
(0) (5) (1) (2) (3)
[3, 3] [4, 6] [2, 1] [0, 3] [1, 4]
5 4 2 4 2
− 1 (L) − − 2 (L)
4 3 1 4 2
1 (L) 2 (L) 1 (H) − 2 (H)
(0) (0) (1) (0) (2)
[1, 0] [0, 1] [2, 2] [0, 2] [2, 3]
2 2 5 2 2
(1) (1) (2) (0) (0)
[0, 2] [0, 2] [0, 5] [0, 2] [0, 2]
2 2 4 2 2
− − 1 (H) − 3 (H, H, L)
2 1 2 2 4
− 1 (L) 3 (L) − 1 (L)
1 2 4 2 45
(0) (0) (0) (2) (8)
[0, 1] [1, 2] [4, 4] [2, 0] [31, 27]
1 2 4 4 50
(0) (1) (1) (1) (20)
[0, 1] [0, 3] [0, 4] [0, 4] [10, 50]
2 2 4 4 50
− − − − 7
2 2 4 4 43
− − − − 14
4 2 0 1 1 8 53
(1) (0) (0) (0) (0) (1) (9)
[3, 4] [2, 1] [0, 0] [0, 1] [0, 1] [5, 7] [36, 34]
1 (1) 2 (0) 1 (0) 0 (0) 1 (0) 5 (1) 55 (21)
[0, 1] [0, 2] [0, 1] [0, 0] [0, 1] [0, 5] [10, 55]
3 2 0 0 1 6 56
1 (L) − − − − 1 8
3 2 0 0 1 6 49
1 (H) − − − − 1 15
Y. Dobashi et al.
NOTE. Numbers represent cases for each category. Numbers in parentheses indicate the cases exhibiting 2+ staining intensity. Abbreviations: DFSP, dermatofibrosarcoma protuberans; PNET, primitive neuroectodermal tumor; N, nuclear positivity; C, cytoplasmic positivity; D, disomy; P, polysomy; L, low-level polysomy; H, high-level polysomy.
AKTs in bone and soft tissue tumors frequently expressed in MPNST (4/4 versus 2/4 cases) and muscular tumors (rhabdomyosarcoma [RMS] and leiomyosarcoma [LMS], 8/9 versus 4/9 in total). Akt2 was found to be localized predominantly in the cytoplasm compared with the nucleus (55 versus 12 cases), and nuclear staining was commonly observed in LS and MPNST. 3.1.4. p-Akt p-Akt staining was observed in 55 (71.4%) of 77 cases, including 48 (73.8%) of 65 malignant and 7 (58.3%) of 12 benign lesions, and was more prominent in the cytoplasm compared with the nucleus (44 versus 27 cases). Positive staining was frequently found in chordoma, RMS, dedifferentiated LS, LMS and MPNST, and others. In benign lesions, schwannoma and dermatofibroma showed positive staining at a higher frequency. The frequency of positive staining for all of these proteins was not significantly different between malignant and benign
131 tumors (T-Akt, P = .3277; Akt1, P = .4598; Akt2, P = .0543; p-Akt1, P = .2744).
3.2. AKT1 and AKT2 genetic changes FISH was performed on 74 tumors, including 65 cases exhibiting T-Akt overexpression and 9 cases of “T-Akt– negative” sarcomas. The FISH signals were successfully visualized in 64 cases, of which 58 (51 malignant and 7 benign) were cases overexpressing T-Akt and 6 cases were “T-Akt–negative” sarcoma (Table 2). Although no amplification of AKT1 was observed, polysomy was detected in 8 cases; 3 cases as high-level (all sarcoma cases) and 5 cases (4 sarcomas and 1 schwannoma) as low-level polysomy. Altogether, 56 cases (50 malignant and 6 benign cases) exhibited disomy, including 6 cases of “T-Akt–negative” sarcoma.
Fig. 1 Immunohistochemical staining for T-Akt and p-Akt, Akt1 and Akt2, and FISH analysis. A case of malignant peripheral nerve sheath tumor that exhibited diffuse nuclear/cytoplasmic positive for T-Akt (A), cytoplasmic staining for p-Akt (B), Akt2 (D), and cytoplasmic (Nnuclear) staining for Akt1 (C). FISH revealed an increase in gene-specific signals (orange fluorescence) and reference probe signals (green fluorescence), indicating low-level polysomy of chromosome 14, which houses AKT1 (E), and high-level polysomy of chromosome 19, AKT2 (F).
132
Y. Dobashi et al.
Fig. 2 Immunohistochemical staining for T-Akt and p-Akt, Akt1 and Akt2, and FISH analysis. A case of schwannoma that exhibited positive staining for T-Akt (A), p-Akt (B), Akt1 (C), and Akt2 (D), predominantly in the cytoplasm. FISH revealed low-level polysomy of chromosome 14 (E) and high-level polysomy of chromosome 19 (F).
For AKT2, no amplification was observed, but polysomy was detected in 15 cases: 5 cases as high-level (4 sarcomas and 1 schwannoma) and 10 cases (all sarcoma) as low-level polysomy. Altogether, 49 cases (43 malignant and 6 benign cases), including all 6 cases of “T-Akt–negative” sarcoma, exhibited disomy. Among 18 cases that exhibited polysomy, 5 had polysomy of both chromosomes. Of these, none exhibited high-level polysomy of both, 4 cases exhibited high polysomy for 1 chromosome and low for the other (LMS, MPNST, osteosarcoma, and schwannoma), and 1 case (myxoid/round-LS) exhibited low-level polysomy of both chromosomes. Thirteen cases exhibited gene gain of either AKT1 or AKT2. Because all T-Akt–negative cases exhibited disomy, the frequency of AKT gain was estimated as 28.3% of sarcoma (17/60 cases, ie, 51 cases successfully analyzed plus 9 T-Akt–negative cases) and 10.0% of benign lesion (1/10 case, ie, 7 cases successfully analyzed plus 3 T-Akt–negative cases). It would be 11.7% (7/60 cases) of sarcoma for AKT1 gain and 23.3% (14/60) for AKT2.
3.3. Comparison of IHC and FISH The results of IHC and FISH were statistically evaluated (Tables 3 and 4). First, overexpression of T-Akt correlated with expression of p-Akt (P = .0094), Akt1 (P b .0001), and Akt2 (P b .0001). Overexpression of both isoforms was significantly correlated (P = .0005), and p-Akt was correlated with Akt1 (P = .008) and Akt2 (P = .0015) (Table 3). Furthermore, cytoplasmic overexpression of T-Akt correlated with Akt activation (P = .0031) and Akt2 overexpression (P = .0007). Conversely, Akt1 overexpression was correlated with nuclear T-Akt (P = .0007) and nuclear p-Akt (P = .0035) (Table 4). We found no statistical difference between the groups showing 2+ or 1+ staining in any of proteins. Second, there was strong correlation between AKT1 and AKT2 gene increases (P b .0001) (Table 3). Lastly, all the tumors exhibiting AKT1 and/or AKT2 increase were accompanied by overexpression and activation of Akt. Furthermore, tumors that did not exhibit T-Akt overexpression did not exhibit polysomy. Overall, a significant
AKTs in bone and soft tissue tumors Table 3
133
Immunohistochemical results and statistical analyses of sarcomas
Factor
n
T-Akt
p-Akt
Akt1
Akt2
AKT1
AKT2
DSS, P Univariate a
T-Akt + − p-Akt + − Akt1 + − Akt2 + − AKT1 + − NA AKT2 + − NA Benign/malignant Diameter (pT) Histologic type c Metastasis AJCC stage d
– 56 9 48 17 45 20
Multivariate b
–
–
–
–
–
.5642
.3455
–
–
–
–
–
.8732
.1029
b.0001 ⁎
.0008 ⁎
–
–
–
–
.9746
.6289
b.0001 ⁎
.0015 ⁎
.0005 ⁎
–
–
–
.8168
.1652
.637
.1134
.0973
NA
–
b.0001 ⁎
.6640
.5505
.946
.0071 ⁎
.7371
.0756
b.0001 ⁎
– –
.1418
.5332
.3277 .6164 .8918 .1452 .1451
.2744 .5092 .8404 .0454 ⁎ .0054 ⁎
.4589 .7828 .8639 .4838 .7054
.0543 .6725 .2146 .0793 .0594
NA .6594 .0314 ⁎ .386 .1753
NA .0352 ⁎ .6078 .0344 ⁎ .074
.0094 ⁎
50 15 8 49 8 15 42 8
NA .2370 .0056 ⁎ b.0001 ⁎ .0041 ⁎
.2249 .1388 .0126 ⁎ .2293
NOTE. Fisher exact test was used if not otherwise specified. Abbreviations: DSS, disease-specific survival; NA, not analyzed. a Log-rank analysis. b Cox proportional hazards regression analysis. c Fisher PLSD test. d Mann-Whitney U test. ⁎ Statistically significant.
correlation was found only between increases in AKT2 gene number and Akt activation (P = .0071) (Table 3). Among 40 cases of sarcoma exhibiting disomy of both AKT1 and AKT2, Akt overexpression was found in 35 cases (87.5%), and activation, in 30 cases (75.0%). Therefore, a substantial fraction of tumors that did not exhibit AKT gene increase nevertheless exhibited Akt overexpression/activation.
of metastasis (P b .0001) and higher AJCC grade (P = .0041), but not with abnormalities in Akts or AKTs. In multivariate analysis, only a history of metastasis (P = .0126) had a statistical impact (Table 3). We found no statistical correlation between subcellular localization and DSS.
4. Discussion 3.4. Clinicopathologic analysis Our overall results were analyzed statistically for their correlation with clinicopathologic profiles (Fig. 3 and Table 3). In all sarcoma cases, AKT1 gain was correlated with specific histologic types (P = .0314), clustering significantly in MPNST and osteosarcoma. AKT2 gene gain was correlated with tumor size (b5 cm or ≥5 cm, P = .0352). Moreover, Akt activation and AKT2 gain were predictors of subsequent metastasis (P = .0454 and P = .0344, respectively). When analyzed with respect to disease-specific survival (DSS), a significant correlation was observed with a history
The development of molecularly targeted therapies for bone and soft tissue tumors lags behind at present. However, Akt is a promising candidate for treating sarcomas because the involvement of Akt in myogenic, adipogenic, and neurogenic tumors is well known [4,9], and the association of Akt deregulation with aggressive behavior in RMS and LMS has recently been noted [1,8]. However, the expression profiles of the Akt isoforms and the significance of AKT gene increases have not been well characterized. We herein undertook the characterization of the expression profiles of Akt isoforms and AKT gains in bone and soft tissue tumors focusing on Akt1 and Akt2 because Akt3 displays the most
134 Table 4 Marker T-Akt N C N/C −/− p-Akt N C N/C −/− Akt1 N C N/C −/− Akt2 N C N/C −/−
Y. Dobashi et al. Correlation between subcellular localization and factors analyzed (P) n
T-Akt
p-Akt
Akt1
Akt2
AKT1
AKT2
Metastasis
DSS
–
.0031 ⁎ (C + N)/C∝ a
.0007 ⁎ N∝ b
.0007 ⁎ (C + N)/C∝ a
.3208
.1245
.1682
.2899
NE
–
.0035 ⁎ N∝ b
.6479
.4246
.4479
.0701
.7535
.5277
.5291
–
.5954
.0549
NA
.1325
.3587
NE
.136
.8381
–
NA
.2054
.6777
.0536
5 17 34 9 3 20 24 18 20 12 13 20 1 40 9 15
Abbreviations: DSS, disease-specific survival; NE, not evaluable because only negative groups were correlated. a Cytoplasmic and both nuclear/cytoplasmic positivity of T-Akt correlated with p-Akt and Akt2 expression. b Nuclear positivity of T-Akt and p-Akt correlated with Akt1 expression. ⁎ Statistically significant by the Mann-Whitney U test and Fisher PLSD test.
restricted tissue distribution, and only minor gene gains have been reported in a few tumor types [5,16]. Our results offer a number of insights: (1) although Akt2 protein has been recognized to be a predominant isoform abundantly expressed in mesenchymal tissues, frequent expression/activation of Akt1 and AKT1 gains was also observed. Indeed, physiologically, both Akt1 and Akt2 have been described to play distinct and essential roles in mesenchymal tissue: Akt1 regulates skeletal muscle growth and adipogenesis [1,11], whereas Akt2 regulates cellular metabolism of insulin-dependent cells and suppresses apoptosis [6]. (2) Benign lesions also showed frequent Akt activation, as previously described for leiomyoma, schwannoma, and giant cell tumors [8,17]. In particular, 1 case of
schwannoma revealed gene gains of both AKT1 and AKT2. (3) Overexpression of each isoform may alter the subcellular localization of T-Akt or p-Akt. (4) AKT gains and overexpression of their isoform products were found to be correlated with various clinicopathologic parameters and thus could be potential determinants of clinical behavior. In a previous study, we found that 6% of lung carcinomas exhibited amplification of AKT1 or AKT2, and 7% exhibited high-level polysomy of chromosome 14 or 19 [7]. In the current study, AKT1 and/or AKT2 gains were found in 28.3% of sarcoma cases, but no amplification was detected. AKT2 gains were found in a larger fraction of cases (23.3%) compared with AKT1 (11.7%). Furthermore, a correlation between AKT2 gain and tumor size suggests that AKT2 gain is
Fig. 3 Kaplan-Meier survival curves for DSS. Higher AJCC staging grade (A) and a history of metastasis (B) were significantly correlated with overall survival by the log-rank test.
AKTs in bone and soft tissue tumors a later event after tumorigenesis. A previous study reported that tumors overexpressing Akt2 were more metastatic in vivo [5,16]. Consistent with this, we observed that tumors with AKT2 gain or Akt activation exhibited a higher frequency of metastasis. Tumors overexpressing Akt2 also showed this trend, but this was not statistically significant (Table 3). Although past studies have suggested that amplification of AKT and Akt activation in tumors indicate a poor prognosis [5,6], we did not find such correlations in lung carcinoma [7], and others found favorable prognosis in colorectal [18] and lung carcinomas [13] exhibiting Akt activation. Thus, the mode of Akt involvement may differ depending on the organ involved and/or may depend on the site of Akt phosphorylation: p-Akt-Thr308 is correlated with poor prognosis, whereas p-Akt-Ser473 is not in lung carcinomas [1,19]. In the current study, we did not find any correlation between AKT gain or Akt activation and prognosis. Akt isoforms have been shown to function differentially, either in complementary or opposing manners [16,20]. The complex interplay between these isoforms is reflected in the varied descriptions of their roles and correlation with clinicopathologic parameters. For example, several studies reported that Akt2 and p-Akt expressions are associated with significantly shorter survival in soft tissue tumors [1,9]. However, this may depend on the histologic type because Akt isoforms play diverse roles in mesenchymal cells, depending on the tissue [21]. In adipose tissue, Akt1 and Akt2 are both important for the regulation of insulin-mediated metabolic signaling, whereas Akt2 is particularly indispensable for the regulation of adipocyte number [22]. Consistently, zebrafish expressing constitutively active Akt2 developed WD-LS, and Akt activation was noted in 32% of human WD-LS [23]. Because, in the current study, Akt activation and gene gains were also frequently found in myxoid/round-LS, the Akts are likely to be involved not only in WD-LS but also in LS in general. In skeletal muscle, Akt1 and Akt2 differentially regulate muscle-specific gene transcription, and Akt2 predominantly promotes differentiation of myoblasts [11,16,24]. Consistently, we observed frequent overexpression of both Akt1 and Akt2 in RMS, and AKT2 gain was found in 1 case. In neurogenic tumors, T-Akt overexpression was observed in all 4 cases of schwannoma and MPNST, and p-Akt, in 3 cases each, with a high frequency of polysomy. Therefore, Akt may play a more functional role in nerve sheath tumors. Furthermore, although Akt1 was overexpressed in all cases of schwannoma, Akt2 was more predominantly overexpressed in MPNST, suggesting that Akt2 plays a critical role in the malignant phenotype. The specific functions of Akt1 and Akt2 [25] are regulated by proteins that modulate their activities [26], which are determined by differential subcellular localization. With respect to soft tissue tumors, one group reported that nuclear p-Akt expression was correlated with a favorable prognosis than cytoplasmic p-Akt [1], but another group reported that detection of p-Akt in both the nucleus and the
135 cytoplasm was correlated with a worse prognosis [9]. In the current study, cytoplasmic localization of T-Akt was correlated with Akt activation. Furthermore, the linkage between Akt2 overexpression and cytoplasmic T-Akt or between Akt1 overexpression and nuclear T-Akt/p-Akt may signify that Akt1 plays a critical role in the nucleus and Akt2 in the cytoplasm. This may explain the isoform-dependent subcellular localization of T-Akt and p-Akt. In tumors exhibiting Akt activation, inhibition of Akt induces apoptosis and thereby impedes cell proliferation, and this observation has led to the development of several Aktspecific inhibitors as cancer therapeutics [4,5,26]. Potent Akt1/Akt2 dual inhibitors have been developed at the preclinical stage [16,27]. However, given the different functions of the 2 isoforms, there may be potential for treatments targeting isoform-specific events [16,28]. In conclusion, we describe that a very substantial fraction of bone sarcoma and soft tissue tumors exhibited not only Akt activation but also AKT gene gains. The presence of the latter suggests that involvement of Akt in these particular kinds of tumors may not be a passive consequence of oncogenesis or merely nonpathogenic “passengers” in this process. Although it may be a physiological phenomenon, these results support the idea that activation and molecular alterations of Akt could serve as therapeutic biomarkers, which would allow the implementation of more coordinated molecular approaches. Using appropriate isoform-specific Akt inhibitors may further enhance the efficacy of chemotherapy for bone and soft tissue tumors in combination with conventional regimens.
Acknowledgments The authors thank the staff in the Department of Pathology, Saitama Medical Center, for their helpful support.
References [1] Valkov A, Kilvaer TK, Sorbye SW, et al. The prognostic impact of Akt isoforms, PI3K and PTEN related to female steroid hormone receptors in soft tissue sarcomas. J Transl Med 2011;9:200. [2] Fletcher CDM, Bridge JA, Hogendoorn P, Mertens F, editors. WHO Classification of Tumours of Soft Tissue and Bone. Lyon: IARC Press; 2013. [3] Vemulapalli S, Mita A, Alvarado Y, Sankhala K, Mita M. The emerging role of mammalian target of rapamycin inhibitors in the treatment of sarcomas. Target Oncol 2011;6:29-39. [4] Bellacosa A, Kumar CC, Di Cristofano A, Testa JR. Activation of AKT kinases in cancer: implications for therapeutic targeting. Adv Cancer Res 2005;94:29-86. [5] Altomare DA, Testa JR. Perturbations of the AKT signaling pathway in human cancer. Oncogene 2005;24:7455-64. [6] Osaki M, Oshimura M, Ito H. PI3K-Akt pathway: its functions and alterations in human cancer. Apoptosis 2004;9:667-76. [7] Dobashi Y, Kimura M, Matsubara H, et al. Molecular alterations in AKT and its protein activation in human lung carcinomas. HUM PATHOL 2012;43:2229-40.
136 [8] Setsu N, Yamamoto H, Kohashi K, et al. The Akt/mammalian target of rapamycin pathway is activated and associated with adverse prognosis in soft tissue leiomyosarcomas. Cancer 2011;118:1637-48. [9] Tomita Y, Morooka T, Hoshida Y, et al. Prognostic significance of activated AKT expression in soft-tissue sarcoma. Clin Cancer Res 2006;12:3070-7. [10] Kirkegaard T, Witton CJ, Edwards J, et al. Molecular alterations in AKT1, AKT2 and AKT3 detected in breast and prostatic cancer by FISH. Histopathology 2010;56:203-11. [11] Vandromme M, Rochat A, Meier R, et al. Protein kinase B beta/Akt2 plays a specific role in muscle differentiation. J Biol Chem 2001;276: 8173-9. [12] Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A. AJCC Cancer Staging Manual. 7th ed. New York: Springer; 2010. [13] Shah A, Swain WA, Richardson D, et al. Phospho-akt expression is associated with a favorable outcome in non-small cell lung cancer. Clin Cancer Res 2005;11:2930-6. [14] Cappuzzo F, Hirsch FR, Rossi E, et al. Epidermal growth factor receptor gene and protein and gefitinib sensitivity in non-small-cell lung cancer. J Natl Cancer Inst 2005;97:643-55. [15] Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-74. [16] Heron-Milhavet L, Khouya N, Fernandez A, Lamb NJ. Akt1 and Akt2: differentiating the aktion. Histol Histopathol 2011;26:651-62. [17] Dobashi Y, Suzuki S, Sato E, et al. EGFR-dependent and independent activation of Akt/mTOR cascade in bone and soft tissue tumors. Mod Pathol 2009;22:1328-40. [18] Baba Y, Nosho K, Shima K, et al. Phosphorylated AKT expression is associated with PIK3CA mutation, low stage, and favorable outcome in 717 colorectal cancers. Cancer 2011;117:1399-408.
Y. Dobashi et al. [19] Al-Saad S, Donnem T, Al-Shibli K, et al. Diverse prognostic roles of Akt isoforms, PTEN and PI3K in tumor epithelial cells and stromal compartment in non-small cell lung cancer. Anticancer Res 2009;29: 4175-83. [20] Zhou GL, Tucker DF, Bae SS, et al. Opposing roles for Akt1 and Akt2 in Rac/Pak signaling and cell migration. J Biol Chem 2006;281:36443-53. [21] Mukherjee A, Wilson EM, Rotwein P. Selective signaling by Akt2 promotes bone morphogenetic protein 2–mediated osteoblast differentiation. Mol Cell Biol 2010;30:1018-27. [22] Fischer-Posovszky P, Tews D, Horenburg S, Debatin KM, Wabitsch M. Differential function of Akt1 and Akt2 in human adipocytes. Mol Cell Endocrinol 2012;358:135-43. [23] Gutierrez A, Snyder EL, Marino-Enriquez A, et al. Aberrant AKT activation drives well-differentiated liposarcoma. Proc Natl Acad Sci U S A 2011;108:16386-91. [24] Cenni V, Bavelloni A, Beretti F, et al. Ankrd2/ARPP is a novel Akt2 specific substrate and regulates myogenic differentiation upon cellular exposure to H(2)O(2). Mol Biol Cell 2011;22:2946-56. [25] Gonzalez E, McGraw TE. Insulin-modulated Akt subcellular localization determines Akt isoform-specific signaling. Proc Natl Acad Sci U S A 2009;106:7004-9. [26] de Frias M, Iglesias-Serret D, Cosialls AM, et al. Akt inhibitors induce apoptosis in chronic lymphocytic leukemia cells. Haematologica 2009; 94:1698-707. [27] Zhao Z, Leister WH, Robinson RG, et al. Discovery of 2,3,5trisubstituted pyridine derivatives as potent Akt1 and Akt2 dual inhibitors. Bioorg Med Chem Lett 2005;15:905-9. [28] Saji M, Narahara K, McCarty SK, et al. Akt1 deficiency delays tumor progression, vascular invasion, and distant metastasis in a murine model of thyroid cancer. Oncogene 2011;30:4307-15.
www.elsevier.com/locate/humpath
Original contribution
Significance of Akt activation and AKT gene increases in soft tissue tumors☆,☆☆ Yoh Dobashi MD, PhD a,⁎, Eiichi Sato MD, PhD b , Yoshinao Oda MD, PhD c , Johji Inazawa MD, PhD d , Akishi Ooi MD, PhD e a
Department of Pathology, Jichi Medical University, Saitama, Japan Department of Orthopedic Surgery, Interdisciplinary Graduate School of Medicine and Engineering, University of Yamanashi, Chuo, Japan c Department of Anatomic Pathology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan d Department of Molecular Cytogenetics, Medical Research Institute, Tokyo Medical and Dental University, Tokyo, Japan e Department of Molecular and Cellular Pathology, Kanazawa University Graduate School of Medical Science, Kanazawa, Japan b
Received 27 May 2013; revised 20 June 2013; accepted 26 June 2013
Keywords: Bone and soft tissue; AKT; Polysomy; Activation; Isoforms; Metastasis; Molecular targeting therapy
Summary To clarify the aberrations of AKT genes, their protein products and clinicopathologic significance in bone and soft tissue tumors, expression profiles of total Akt, its isoforms and activated Akt, and increases in copy number of AKT1/AKT2 genes were examined. Immunohistochemical analysis in 77 cases revealed overexpression of total Akt, Akt1, Akt2, and phosphorylated Akt in 84.4%, 67.5%, 72.7%, and 71.4%, respectively. Positive results were also observed in benign lesions but at a lower frequency. Overexpression of Akt1 was more frequent than that of Akt2 in well-differentiated liposarcoma (6/7 versus 3/7 cases) and schwannoma (4/4 versus 1/4 cases), whereas Akt2 overexpression and Akt activation were more frequent than Akt1 overexpression in malignant nerve sheath (3/4 and 4/4, respectively, versus 2/4 cases) and muscular tumors (8/9 and 8/9 versus 4/9 cases). By fluorescence in situ hybridization analysis, increase of gene copy number was observed in 13.3% for AKT1 and in 25.0% for AKT2 due to polysomy of chromosome 14 or 19, respectively, but not gene amplification. One case of schwannoma exhibited polysomy of both chromosomes 14 and 19. Akt activation was correlated with total Akt cytoplasmic localization (P = .0031) and subsequent metastasis (P = .0454). Moreover, AKT2 gene increase correlated with tumor size (P = .0352) and metastasis (P = .0344). In conclusion, in a defined subset of bone and soft tissue tumors, including benign tumors, Akt was frequently overexpressed and activated, and AKT1/2 copy number was increased. Because abnormality of Akt/AKT correlated with clinicopathologic profiles, novel therapies targeting isoformspecific Akts may be useful for these particular types of tumors. © 2014 Elsevier Inc. All rights reserved.
☆
No financial disclosure or conflicts of interests are declared. Funding support: This work was partially supported by a grant from the Japan Society for the Promotion of Science C23590409 (Y. D.) and C22590310 (A. O.) and Smoking Research Foundation (Y. D). ⁎ Corresponding author. Department of Pathology Saitama Medical Center, Jichi Medical University, 1-847, Amanuma, Omiya, Saitama, 3308503, Japan. E-mail address: [email protected] (Y. Dobashi). ☆☆
0046-8177/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.humpath.2013.06.024
1. Introduction Malignant bone and soft tissue tumors currently represent approximately 1% of adult malignancies and 15% to 20% of pediatric malignancies [1]. These are heterogeneous
128 neoplasms consisting of more than 50 subtypes [2]. Although the overall survival rates average 60% to 70%, they are far lower in some subtypes, due primarily to limited therapy options and the development of distant metastases [1,3]. Therefore, there is much interest in supplementing conventional chemotherapy with novel “tailored therapies” that target signaling molecules critical for each tumor. Although an increasing number of such promising novel therapies have been developed, progress against soft tissue tumors has been slow due to their rarity compared with carcinomas. Nevertheless, there is some promise for tailored therapies, especially given the progress in targeting growth factor receptors and critical effector molecules downstream of them [3,4]. Among the myriad of growth-promoting signaling cascades, the phospho-inositol-3 kinase (PI3K)/Akt pathway is especially critical for both normal and pathologic processes [5,6]. In particular, Akt family proteins function at the crossroads of multiple intracellular pathways and can potentially influence thousands of downstream targets [4-6]. Akt activities are frequently dysregulated in human cancers, and there are numerous reports describing the correlations between Akt activation and clinicopathologic profiles [4,5,7]. Recently, Akt pathway activation was shown to be correlated with deep location, high histologic grade, a higher probability of metastasis, and worse prognosis in soft tissue tumors [1,8,9]. This would indicate that Akt may be a candidate for targeted therapies against soft tissue tumors. Akt family proteins are encoded by 3 related genes (AKT1-3), located on chromosomes 14q32, 19q13, and 1q44, respectively [6,10]. Their gene products, Akt1-3, are 56-kd serine/threonine kinases that display distinct tissue and subcellular distributions [6]. In response to upstream signals, they translocate to specific subcellular compartments where they regulate cellular functions such as cell metabolism, growth, and survival [5,6,9]. Although Akt1 is expressed ubiquitously at a high level, Akt2 is elevated in insulinresponsive tissues such as fat, skeletal muscle, and the liver [1,5,6]. Although Akt1 and Akt2 typically require membrane localization for their activity, they can also localize to the nucleus where they function in cell differentiation and/or suppression of apoptosis [6,11]. In cancers, distinct changes in the expression of Akts have been observed. Akt1 is up-regulated and/or activated in gastric, prostatic, ovarian, and breast carcinomas, and this up-regulation is often associated with poor prognosis [4-6]. Akt2 is activated in hepatocellular, colorectal, ovarian, pancreatic, and breast carcinomas [4,5]. Moreover, both Akt1 and Akt2 are implicated in invasion and in metastatic processes [5]. One mechanism of Akt overexpression is gene amplification, and AKT1 amplification has been reported in sporadic cases of the lung, gastric, breast, and prostatic carcinomas [5,7,10]. AKT2 gene amplification has been frequently detected in breast, ovarian, and pancreatic carcinomas and is associated with a poor prognosis in ovarian carcinoma [5,10].
Y. Dobashi et al. In our previous study in lung carcinoma, fluorescence in situ hybridization (FISH) analysis revealed amplification of AKT1 and/or AKT2 in 6% and high-level polysomy in 7% of the cases [7]. Although the Akt-driven pathway may constitute a critical mechanism in the pathology of bone and soft tissue tumors, there are few studies to date that comprehensively describe the expression and activation status of Akt or alterations in the AKT genes. In addition, the relative contributions of the different Akt isoforms to bone and soft tissue tumors are not well understood. Herein, we conducted a comprehensive analysis of Akts and specifically asked if (1) total Akt (T-Akt) is constitutively overexpressed/ activated globally or in particular subtypes, (2) there is any preponderance in the expression of particular Akt isoforms and/or gene increase specific to histologic subtypes, and (3) Akt activation or gene increases correlate with clinicopathologic profiles.
2. Materials and methods 2.1. Cases and classification Under protocols approved by the Institutional Tissue Committees, 77 cases (65 malignant and 12 benign) were obtained from surgeries performed at the Department of Orthopedic Surgery, University of Yamanashi Hospital, and University of Kanazawa Hospital. Pathologic diagnoses were made according to World Health Organization classification [2]. Staging was determined in accordance with the American Joint Committee on Cancer (AJCC) staging system [12]. These details are presented in Tables 1 and 2. Cases were restricted to patients who had not received preoperative chemotherapy.
2.2. Immunohistochemistry Serial sections were stained with 4 primary antibodies followed by visualization with a CSAII kit (Catalyzed Signal Amplification System 2; Dako, Glostrup, Denmark), T-Akt (1:600; Cell Signaling Technology [CST], Beverly, MA), Akt1 (1:40; monoclonal, C73H10; CST), Akt2 (1:150, polyclonal; Abcam, Cambridge, UK), phosphorylated Akt (p-Akt) (p-AktSer473; 1:300, monoclonal; CST) [7]. The sensitivity and the specificity of the antibodies had been previously validated on cell lines and tissue specimens by immunohistochemistry (IHC) and immunoblotting [1,7,8,13]. IHC expression was evaluated by 2 observers (Y. D. and A. O.) and scored based on staining intensity and the fraction of positive cells. Significant staining was defined as staining that was more intense compared with endothelial cells [8,9]. IHC expression level was quantitatively evaluated by the fraction of stained tumor cells: negative, less than 10%; 1+, 10% or more and less than 50%; 2+, 50% or more cells with significant staining [7,8]. When discordance was noted in evaluation of the results between
AKTs in bone and soft tissue tumors Table 1
129
2.4. Statistical analysis
Clinicopathologic profiles of 77 cases
Clinicopathologic factors Histology Malignant Benign Sex Male Female Age (y), mean (range) b60 ≥60 Sarcoma only Diameter (pT) b5 cm (pT1) ≥5 cm (pT2) Recurrence + − Metastasis + − AJCC stage I II III IV
n 65 12 43 34 57 (33-81) 46 31
11 54 5 60 10 55
Agreement among observers in the interpretation of IHC results was evaluated by κ statistics [15]. Other statistical analysis was performed with the StatView package (version 5.0; SAS Inc., Cary, NC, USA). Differences in frequencies of positive immunostaining between 2 groups were analyzed by Fisher exact test. For correlations among more than 3 categories, Mann-Whitney U test or Fisher protected least significant difference (PLSD) analysis was applied. Log-rank analysis with the Kaplan-Meier method was used to determine the correlation between the variables and survival period via univariate analysis. In the multivariate analysis, a Cox proportional hazards regression analysis was performed. Two-sided P b .05 was placed to determine statistical significance.
3. Results 3.1. IHC
17 25 23 0
2 observers, it was resolved by discussion. When scores were classified into 2 groups for statistical analysis, 1+ and 2+ were combined as positive [7,14].
2.3. FISH analysis AKT1 and AKT2 gene copy aberrations were examined by FISH analysis, as described previously [7,14]. Samples consisted of 65 cases exhibiting T-Akt overexpression by IHC and 9 cases of sarcoma without T-Akt overexpression. For FISH probes, we used the bacterial artificial chromosome clones CTD-2507D9 (AKT1, 14q32.32) and CTB166E20 (AKT2, 19q13.1-13.2). For reference probes adjacent to the centromere, we used RP11-203 M5 (14q11.2) for AKT1 and RP11-124 K10 (19p13.11) for AKT2 [7]. The specificity and localization of both probes were previously validated [7]. FISH signals were examined in at least 30 nuclei for each tumor independently by 2 observers (Y. D. and A. O.). Results were classified into 4 strata based on AKT copy number using the previously established scoring system: 1, disomy (≤2 copies in N90% of cancer cells); 2, low-level polysomy (≥3 copies with an equivalent number of reference genes in N10% and b40%, without amplification); 3, highlevel polysomy (polysomy in ≥40%, without amplification); and 4, amplification (tight clusters, average AKT/chromosome ≥2, or ≥15 copies of AKT per cell in ≥10%) [7,14]. When discordance was noted in evaluation of the results between 2 observers, it was resolved by discussion.
Overall results of IHC are presented in Figs. 1 and 2 as well as Table 2. In normal tissues, T-Akt, Akt1, Akt2, and pAkt staining was observed in the cytoplasm of skeletal muscle, vascular smooth muscle, and weakly in occasional endothelial cells and lymphocytes. Interobserver agreement was “almost perfect” (T-Akt, κ = 0.86; Akt1, κ = 0.89; Akt2, κ = 0.88; p-Akt, κ = 0.91). 3.1.1. T-Akt Positive staining was observed in the nuclei and/or the cytoplasm in 65 (84.4%) of 77 cases, including 56 (86.2%) of 65 malignant and 9 (75.0%) of 12 benign lesions. More than 50% of the cases exhibited positive staining across all histologic types. Positive staining was more prominent in the cytoplasm than in the nucleus (59 versus 45 cases), and nuclear staining was more frequently found in liposarcoma (LS), malignant peripheral nerve sheath tumor (MPNST), chordoma, and others. 3.1.2. Akt1 Akt1 staining was observed in wide varieties of tumors in 52 (67.5%) of 77 cases, including 45 (69.2%) of 65 malignant and 7 (58.3%) of 12 benign lesions. Compared with Akt2, Akt1 expression was found more frequently in welldifferentiated LS (WD-LS, 6/7 versus 3/7 cases) and schwannoma (4/4 versus 1/4 cases). Akt1 expression was observed in 80.0% (52/65) of the cases positive for T-Akt and was more prominent in the nucleus compared with the cytoplasm (38 versus 30 cases). 3.1.3. Akt2 Akt2 staining was observed in 56 (72.7%) of 77 cases, including 50 (76.9%) of 65 malignant and 6 (50.0%) of 12 benign lesions. Compared with Akt1, Akt2 was more
130
Table 2
Overall results of immunohistochemical and FISH analyses
Histology
Malignant tumors Undifferentiated pleomorphic sarcoma Myxofibrosarcoma Liposarcoma Well differentiated Myxoid/round cell Dedifferentiated Synovial sarcoma MPNST Rhabdomyosarcoma Pleomorphic Embryonal Leiomyosarcoma DFSP Osteosarcoma Chondrosarcoma Grade I Grade II Chordoma Ewing sarcoma/PNET Total Benign tumors Schwannoma Dermatofibroma Leiomyoma Giant cell tumor Lipoma Total Total
Total cases
T-Akt
6
5 (2)
[3, 5]
4 (2)
[2, 4]
5 (1)
[4, 2]
5 (1)
6
4 (1)
[3, 3]
4 (1)
[3, 4]
4 (0)
[3, 1]
7 7 2 4 4
6 6 2 3 4
(1) (5) (0) (2) (3)
[6, 3] [6, 6] [2, 1] [2, 3] [3, 4]
3 5 2 3 3
(0) (4) (1) (0) (2)
[3, 3] [4, 5] [1, 1] [1, 2] [1, 3]
6 5 2 2 2
(0) (0) (0) (1) (1)
2 2 5 3 5
2 2 5 2 4
(1) (1) (3) (1) (2)
[1, 2] [0, 2] [2, 5] [2, 2] [2, 4]
2 2 4 2 4
(0) (0) (3) (0) (1)
[1, 1] [2, 0] [1, 4] [2, 2] [2, 4]
1 1 3 2 3
2 2 4 4 65
1 2 4 4 56
(0) (0) (0) (2) (24)
[0, 1] [1, 2] [4, 4] [2, 4] [39, 51]
1 2 4 3 48
(0) (0) (0) (1) (15)
[0, 1] [1, 2] [4, 4] [1, 3] [29, 43]
4 2 2 2 2 12 77
4 2 1 1 1 9 65
(2) (0) (0) (0) (0) (2) (26)
[3, 4] [2, 2] [0, 1] [1, 1] [0, 1] [6, 9] [45, 60]
3 (1) 2 (0) 1 (0) 1 (0) 0 (0) 7 (1) 55 (16)
[2, 3] [0, 2] [0, 1] [1, 1] [0, 0] [3, 7] [32, 50]
Positive (2+)
p-Akt [N, C]
Positive (2+)
Akt1 [N, C]
Positive (2+)
Akt2 [N, C]
Positive (2+)
AKT1 [N, C]
AKT2
D
P
D
P
[0, 5]
5
−
3
2 (L, H)
3 (1)
[0, 3]
4
−
3
1 (L)
[5, 1] [3, 4] [2, 1] [1, 1] [1, 2]
3 6 2 3 4
(0) (5) (1) (2) (3)
[3, 3] [4, 6] [2, 1] [0, 3] [1, 4]
5 4 2 4 2
− 1 (L) − − 2 (L)
4 3 1 4 2
1 (L) 2 (L) 1 (H) − 2 (H)
(0) (0) (1) (0) (2)
[1, 0] [0, 1] [2, 2] [0, 2] [2, 3]
2 2 5 2 2
(1) (1) (2) (0) (0)
[0, 2] [0, 2] [0, 5] [0, 2] [0, 2]
2 2 4 2 2
− − 1 (H) − 3 (H, H, L)
2 1 2 2 4
− 1 (L) 3 (L) − 1 (L)
1 2 4 2 45
(0) (0) (0) (2) (8)
[0, 1] [1, 2] [4, 4] [2, 0] [31, 27]
1 2 4 4 50
(0) (1) (1) (1) (20)
[0, 1] [0, 3] [0, 4] [0, 4] [10, 50]
2 2 4 4 50
− − − − 7
2 2 4 4 43
− − − − 14
4 2 0 1 1 8 53
(1) (0) (0) (0) (0) (1) (9)
[3, 4] [2, 1] [0, 0] [0, 1] [0, 1] [5, 7] [36, 34]
1 (1) 2 (0) 1 (0) 0 (0) 1 (0) 5 (1) 55 (21)
[0, 1] [0, 2] [0, 1] [0, 0] [0, 1] [0, 5] [10, 55]
3 2 0 0 1 6 56
1 (L) − − − − 1 8
3 2 0 0 1 6 49
1 (H) − − − − 1 15
Y. Dobashi et al.
NOTE. Numbers represent cases for each category. Numbers in parentheses indicate the cases exhibiting 2+ staining intensity. Abbreviations: DFSP, dermatofibrosarcoma protuberans; PNET, primitive neuroectodermal tumor; N, nuclear positivity; C, cytoplasmic positivity; D, disomy; P, polysomy; L, low-level polysomy; H, high-level polysomy.
AKTs in bone and soft tissue tumors frequently expressed in MPNST (4/4 versus 2/4 cases) and muscular tumors (rhabdomyosarcoma [RMS] and leiomyosarcoma [LMS], 8/9 versus 4/9 in total). Akt2 was found to be localized predominantly in the cytoplasm compared with the nucleus (55 versus 12 cases), and nuclear staining was commonly observed in LS and MPNST. 3.1.4. p-Akt p-Akt staining was observed in 55 (71.4%) of 77 cases, including 48 (73.8%) of 65 malignant and 7 (58.3%) of 12 benign lesions, and was more prominent in the cytoplasm compared with the nucleus (44 versus 27 cases). Positive staining was frequently found in chordoma, RMS, dedifferentiated LS, LMS and MPNST, and others. In benign lesions, schwannoma and dermatofibroma showed positive staining at a higher frequency. The frequency of positive staining for all of these proteins was not significantly different between malignant and benign
131 tumors (T-Akt, P = .3277; Akt1, P = .4598; Akt2, P = .0543; p-Akt1, P = .2744).
3.2. AKT1 and AKT2 genetic changes FISH was performed on 74 tumors, including 65 cases exhibiting T-Akt overexpression and 9 cases of “T-Akt– negative” sarcomas. The FISH signals were successfully visualized in 64 cases, of which 58 (51 malignant and 7 benign) were cases overexpressing T-Akt and 6 cases were “T-Akt–negative” sarcoma (Table 2). Although no amplification of AKT1 was observed, polysomy was detected in 8 cases; 3 cases as high-level (all sarcoma cases) and 5 cases (4 sarcomas and 1 schwannoma) as low-level polysomy. Altogether, 56 cases (50 malignant and 6 benign cases) exhibited disomy, including 6 cases of “T-Akt–negative” sarcoma.
Fig. 1 Immunohistochemical staining for T-Akt and p-Akt, Akt1 and Akt2, and FISH analysis. A case of malignant peripheral nerve sheath tumor that exhibited diffuse nuclear/cytoplasmic positive for T-Akt (A), cytoplasmic staining for p-Akt (B), Akt2 (D), and cytoplasmic (Nnuclear) staining for Akt1 (C). FISH revealed an increase in gene-specific signals (orange fluorescence) and reference probe signals (green fluorescence), indicating low-level polysomy of chromosome 14, which houses AKT1 (E), and high-level polysomy of chromosome 19, AKT2 (F).
132
Y. Dobashi et al.
Fig. 2 Immunohistochemical staining for T-Akt and p-Akt, Akt1 and Akt2, and FISH analysis. A case of schwannoma that exhibited positive staining for T-Akt (A), p-Akt (B), Akt1 (C), and Akt2 (D), predominantly in the cytoplasm. FISH revealed low-level polysomy of chromosome 14 (E) and high-level polysomy of chromosome 19 (F).
For AKT2, no amplification was observed, but polysomy was detected in 15 cases: 5 cases as high-level (4 sarcomas and 1 schwannoma) and 10 cases (all sarcoma) as low-level polysomy. Altogether, 49 cases (43 malignant and 6 benign cases), including all 6 cases of “T-Akt–negative” sarcoma, exhibited disomy. Among 18 cases that exhibited polysomy, 5 had polysomy of both chromosomes. Of these, none exhibited high-level polysomy of both, 4 cases exhibited high polysomy for 1 chromosome and low for the other (LMS, MPNST, osteosarcoma, and schwannoma), and 1 case (myxoid/round-LS) exhibited low-level polysomy of both chromosomes. Thirteen cases exhibited gene gain of either AKT1 or AKT2. Because all T-Akt–negative cases exhibited disomy, the frequency of AKT gain was estimated as 28.3% of sarcoma (17/60 cases, ie, 51 cases successfully analyzed plus 9 T-Akt–negative cases) and 10.0% of benign lesion (1/10 case, ie, 7 cases successfully analyzed plus 3 T-Akt–negative cases). It would be 11.7% (7/60 cases) of sarcoma for AKT1 gain and 23.3% (14/60) for AKT2.
3.3. Comparison of IHC and FISH The results of IHC and FISH were statistically evaluated (Tables 3 and 4). First, overexpression of T-Akt correlated with expression of p-Akt (P = .0094), Akt1 (P b .0001), and Akt2 (P b .0001). Overexpression of both isoforms was significantly correlated (P = .0005), and p-Akt was correlated with Akt1 (P = .008) and Akt2 (P = .0015) (Table 3). Furthermore, cytoplasmic overexpression of T-Akt correlated with Akt activation (P = .0031) and Akt2 overexpression (P = .0007). Conversely, Akt1 overexpression was correlated with nuclear T-Akt (P = .0007) and nuclear p-Akt (P = .0035) (Table 4). We found no statistical difference between the groups showing 2+ or 1+ staining in any of proteins. Second, there was strong correlation between AKT1 and AKT2 gene increases (P b .0001) (Table 3). Lastly, all the tumors exhibiting AKT1 and/or AKT2 increase were accompanied by overexpression and activation of Akt. Furthermore, tumors that did not exhibit T-Akt overexpression did not exhibit polysomy. Overall, a significant
AKTs in bone and soft tissue tumors Table 3
133
Immunohistochemical results and statistical analyses of sarcomas
Factor
n
T-Akt
p-Akt
Akt1
Akt2
AKT1
AKT2
DSS, P Univariate a
T-Akt + − p-Akt + − Akt1 + − Akt2 + − AKT1 + − NA AKT2 + − NA Benign/malignant Diameter (pT) Histologic type c Metastasis AJCC stage d
– 56 9 48 17 45 20
Multivariate b
–
–
–
–
–
.5642
.3455
–
–
–
–
–
.8732
.1029
b.0001 ⁎
.0008 ⁎
–
–
–
–
.9746
.6289
b.0001 ⁎
.0015 ⁎
.0005 ⁎
–
–
–
.8168
.1652
.637
.1134
.0973
NA
–
b.0001 ⁎
.6640
.5505
.946
.0071 ⁎
.7371
.0756
b.0001 ⁎
– –
.1418
.5332
.3277 .6164 .8918 .1452 .1451
.2744 .5092 .8404 .0454 ⁎ .0054 ⁎
.4589 .7828 .8639 .4838 .7054
.0543 .6725 .2146 .0793 .0594
NA .6594 .0314 ⁎ .386 .1753
NA .0352 ⁎ .6078 .0344 ⁎ .074
.0094 ⁎
50 15 8 49 8 15 42 8
NA .2370 .0056 ⁎ b.0001 ⁎ .0041 ⁎
.2249 .1388 .0126 ⁎ .2293
NOTE. Fisher exact test was used if not otherwise specified. Abbreviations: DSS, disease-specific survival; NA, not analyzed. a Log-rank analysis. b Cox proportional hazards regression analysis. c Fisher PLSD test. d Mann-Whitney U test. ⁎ Statistically significant.
correlation was found only between increases in AKT2 gene number and Akt activation (P = .0071) (Table 3). Among 40 cases of sarcoma exhibiting disomy of both AKT1 and AKT2, Akt overexpression was found in 35 cases (87.5%), and activation, in 30 cases (75.0%). Therefore, a substantial fraction of tumors that did not exhibit AKT gene increase nevertheless exhibited Akt overexpression/activation.
of metastasis (P b .0001) and higher AJCC grade (P = .0041), but not with abnormalities in Akts or AKTs. In multivariate analysis, only a history of metastasis (P = .0126) had a statistical impact (Table 3). We found no statistical correlation between subcellular localization and DSS.
4. Discussion 3.4. Clinicopathologic analysis Our overall results were analyzed statistically for their correlation with clinicopathologic profiles (Fig. 3 and Table 3). In all sarcoma cases, AKT1 gain was correlated with specific histologic types (P = .0314), clustering significantly in MPNST and osteosarcoma. AKT2 gene gain was correlated with tumor size (b5 cm or ≥5 cm, P = .0352). Moreover, Akt activation and AKT2 gain were predictors of subsequent metastasis (P = .0454 and P = .0344, respectively). When analyzed with respect to disease-specific survival (DSS), a significant correlation was observed with a history
The development of molecularly targeted therapies for bone and soft tissue tumors lags behind at present. However, Akt is a promising candidate for treating sarcomas because the involvement of Akt in myogenic, adipogenic, and neurogenic tumors is well known [4,9], and the association of Akt deregulation with aggressive behavior in RMS and LMS has recently been noted [1,8]. However, the expression profiles of the Akt isoforms and the significance of AKT gene increases have not been well characterized. We herein undertook the characterization of the expression profiles of Akt isoforms and AKT gains in bone and soft tissue tumors focusing on Akt1 and Akt2 because Akt3 displays the most
134 Table 4 Marker T-Akt N C N/C −/− p-Akt N C N/C −/− Akt1 N C N/C −/− Akt2 N C N/C −/−
Y. Dobashi et al. Correlation between subcellular localization and factors analyzed (P) n
T-Akt
p-Akt
Akt1
Akt2
AKT1
AKT2
Metastasis
DSS
–
.0031 ⁎ (C + N)/C∝ a
.0007 ⁎ N∝ b
.0007 ⁎ (C + N)/C∝ a
.3208
.1245
.1682
.2899
NE
–
.0035 ⁎ N∝ b
.6479
.4246
.4479
.0701
.7535
.5277
.5291
–
.5954
.0549
NA
.1325
.3587
NE
.136
.8381
–
NA
.2054
.6777
.0536
5 17 34 9 3 20 24 18 20 12 13 20 1 40 9 15
Abbreviations: DSS, disease-specific survival; NE, not evaluable because only negative groups were correlated. a Cytoplasmic and both nuclear/cytoplasmic positivity of T-Akt correlated with p-Akt and Akt2 expression. b Nuclear positivity of T-Akt and p-Akt correlated with Akt1 expression. ⁎ Statistically significant by the Mann-Whitney U test and Fisher PLSD test.
restricted tissue distribution, and only minor gene gains have been reported in a few tumor types [5,16]. Our results offer a number of insights: (1) although Akt2 protein has been recognized to be a predominant isoform abundantly expressed in mesenchymal tissues, frequent expression/activation of Akt1 and AKT1 gains was also observed. Indeed, physiologically, both Akt1 and Akt2 have been described to play distinct and essential roles in mesenchymal tissue: Akt1 regulates skeletal muscle growth and adipogenesis [1,11], whereas Akt2 regulates cellular metabolism of insulin-dependent cells and suppresses apoptosis [6]. (2) Benign lesions also showed frequent Akt activation, as previously described for leiomyoma, schwannoma, and giant cell tumors [8,17]. In particular, 1 case of
schwannoma revealed gene gains of both AKT1 and AKT2. (3) Overexpression of each isoform may alter the subcellular localization of T-Akt or p-Akt. (4) AKT gains and overexpression of their isoform products were found to be correlated with various clinicopathologic parameters and thus could be potential determinants of clinical behavior. In a previous study, we found that 6% of lung carcinomas exhibited amplification of AKT1 or AKT2, and 7% exhibited high-level polysomy of chromosome 14 or 19 [7]. In the current study, AKT1 and/or AKT2 gains were found in 28.3% of sarcoma cases, but no amplification was detected. AKT2 gains were found in a larger fraction of cases (23.3%) compared with AKT1 (11.7%). Furthermore, a correlation between AKT2 gain and tumor size suggests that AKT2 gain is
Fig. 3 Kaplan-Meier survival curves for DSS. Higher AJCC staging grade (A) and a history of metastasis (B) were significantly correlated with overall survival by the log-rank test.
AKTs in bone and soft tissue tumors a later event after tumorigenesis. A previous study reported that tumors overexpressing Akt2 were more metastatic in vivo [5,16]. Consistent with this, we observed that tumors with AKT2 gain or Akt activation exhibited a higher frequency of metastasis. Tumors overexpressing Akt2 also showed this trend, but this was not statistically significant (Table 3). Although past studies have suggested that amplification of AKT and Akt activation in tumors indicate a poor prognosis [5,6], we did not find such correlations in lung carcinoma [7], and others found favorable prognosis in colorectal [18] and lung carcinomas [13] exhibiting Akt activation. Thus, the mode of Akt involvement may differ depending on the organ involved and/or may depend on the site of Akt phosphorylation: p-Akt-Thr308 is correlated with poor prognosis, whereas p-Akt-Ser473 is not in lung carcinomas [1,19]. In the current study, we did not find any correlation between AKT gain or Akt activation and prognosis. Akt isoforms have been shown to function differentially, either in complementary or opposing manners [16,20]. The complex interplay between these isoforms is reflected in the varied descriptions of their roles and correlation with clinicopathologic parameters. For example, several studies reported that Akt2 and p-Akt expressions are associated with significantly shorter survival in soft tissue tumors [1,9]. However, this may depend on the histologic type because Akt isoforms play diverse roles in mesenchymal cells, depending on the tissue [21]. In adipose tissue, Akt1 and Akt2 are both important for the regulation of insulin-mediated metabolic signaling, whereas Akt2 is particularly indispensable for the regulation of adipocyte number [22]. Consistently, zebrafish expressing constitutively active Akt2 developed WD-LS, and Akt activation was noted in 32% of human WD-LS [23]. Because, in the current study, Akt activation and gene gains were also frequently found in myxoid/round-LS, the Akts are likely to be involved not only in WD-LS but also in LS in general. In skeletal muscle, Akt1 and Akt2 differentially regulate muscle-specific gene transcription, and Akt2 predominantly promotes differentiation of myoblasts [11,16,24]. Consistently, we observed frequent overexpression of both Akt1 and Akt2 in RMS, and AKT2 gain was found in 1 case. In neurogenic tumors, T-Akt overexpression was observed in all 4 cases of schwannoma and MPNST, and p-Akt, in 3 cases each, with a high frequency of polysomy. Therefore, Akt may play a more functional role in nerve sheath tumors. Furthermore, although Akt1 was overexpressed in all cases of schwannoma, Akt2 was more predominantly overexpressed in MPNST, suggesting that Akt2 plays a critical role in the malignant phenotype. The specific functions of Akt1 and Akt2 [25] are regulated by proteins that modulate their activities [26], which are determined by differential subcellular localization. With respect to soft tissue tumors, one group reported that nuclear p-Akt expression was correlated with a favorable prognosis than cytoplasmic p-Akt [1], but another group reported that detection of p-Akt in both the nucleus and the
135 cytoplasm was correlated with a worse prognosis [9]. In the current study, cytoplasmic localization of T-Akt was correlated with Akt activation. Furthermore, the linkage between Akt2 overexpression and cytoplasmic T-Akt or between Akt1 overexpression and nuclear T-Akt/p-Akt may signify that Akt1 plays a critical role in the nucleus and Akt2 in the cytoplasm. This may explain the isoform-dependent subcellular localization of T-Akt and p-Akt. In tumors exhibiting Akt activation, inhibition of Akt induces apoptosis and thereby impedes cell proliferation, and this observation has led to the development of several Aktspecific inhibitors as cancer therapeutics [4,5,26]. Potent Akt1/Akt2 dual inhibitors have been developed at the preclinical stage [16,27]. However, given the different functions of the 2 isoforms, there may be potential for treatments targeting isoform-specific events [16,28]. In conclusion, we describe that a very substantial fraction of bone sarcoma and soft tissue tumors exhibited not only Akt activation but also AKT gene gains. The presence of the latter suggests that involvement of Akt in these particular kinds of tumors may not be a passive consequence of oncogenesis or merely nonpathogenic “passengers” in this process. Although it may be a physiological phenomenon, these results support the idea that activation and molecular alterations of Akt could serve as therapeutic biomarkers, which would allow the implementation of more coordinated molecular approaches. Using appropriate isoform-specific Akt inhibitors may further enhance the efficacy of chemotherapy for bone and soft tissue tumors in combination with conventional regimens.
Acknowledgments The authors thank the staff in the Department of Pathology, Saitama Medical Center, for their helpful support.
References [1] Valkov A, Kilvaer TK, Sorbye SW, et al. The prognostic impact of Akt isoforms, PI3K and PTEN related to female steroid hormone receptors in soft tissue sarcomas. J Transl Med 2011;9:200. [2] Fletcher CDM, Bridge JA, Hogendoorn P, Mertens F, editors. WHO Classification of Tumours of Soft Tissue and Bone. Lyon: IARC Press; 2013. [3] Vemulapalli S, Mita A, Alvarado Y, Sankhala K, Mita M. The emerging role of mammalian target of rapamycin inhibitors in the treatment of sarcomas. Target Oncol 2011;6:29-39. [4] Bellacosa A, Kumar CC, Di Cristofano A, Testa JR. Activation of AKT kinases in cancer: implications for therapeutic targeting. Adv Cancer Res 2005;94:29-86. [5] Altomare DA, Testa JR. Perturbations of the AKT signaling pathway in human cancer. Oncogene 2005;24:7455-64. [6] Osaki M, Oshimura M, Ito H. PI3K-Akt pathway: its functions and alterations in human cancer. Apoptosis 2004;9:667-76. [7] Dobashi Y, Kimura M, Matsubara H, et al. Molecular alterations in AKT and its protein activation in human lung carcinomas. HUM PATHOL 2012;43:2229-40.
136 [8] Setsu N, Yamamoto H, Kohashi K, et al. The Akt/mammalian target of rapamycin pathway is activated and associated with adverse prognosis in soft tissue leiomyosarcomas. Cancer 2011;118:1637-48. [9] Tomita Y, Morooka T, Hoshida Y, et al. Prognostic significance of activated AKT expression in soft-tissue sarcoma. Clin Cancer Res 2006;12:3070-7. [10] Kirkegaard T, Witton CJ, Edwards J, et al. Molecular alterations in AKT1, AKT2 and AKT3 detected in breast and prostatic cancer by FISH. Histopathology 2010;56:203-11. [11] Vandromme M, Rochat A, Meier R, et al. Protein kinase B beta/Akt2 plays a specific role in muscle differentiation. J Biol Chem 2001;276: 8173-9. [12] Edge SB, Byrd DR, Compton CC, Fritz AG, Greene FL, Trotti A. AJCC Cancer Staging Manual. 7th ed. New York: Springer; 2010. [13] Shah A, Swain WA, Richardson D, et al. Phospho-akt expression is associated with a favorable outcome in non-small cell lung cancer. Clin Cancer Res 2005;11:2930-6. [14] Cappuzzo F, Hirsch FR, Rossi E, et al. Epidermal growth factor receptor gene and protein and gefitinib sensitivity in non-small-cell lung cancer. J Natl Cancer Inst 2005;97:643-55. [15] Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977;33:159-74. [16] Heron-Milhavet L, Khouya N, Fernandez A, Lamb NJ. Akt1 and Akt2: differentiating the aktion. Histol Histopathol 2011;26:651-62. [17] Dobashi Y, Suzuki S, Sato E, et al. EGFR-dependent and independent activation of Akt/mTOR cascade in bone and soft tissue tumors. Mod Pathol 2009;22:1328-40. [18] Baba Y, Nosho K, Shima K, et al. Phosphorylated AKT expression is associated with PIK3CA mutation, low stage, and favorable outcome in 717 colorectal cancers. Cancer 2011;117:1399-408.
Y. Dobashi et al. [19] Al-Saad S, Donnem T, Al-Shibli K, et al. Diverse prognostic roles of Akt isoforms, PTEN and PI3K in tumor epithelial cells and stromal compartment in non-small cell lung cancer. Anticancer Res 2009;29: 4175-83. [20] Zhou GL, Tucker DF, Bae SS, et al. Opposing roles for Akt1 and Akt2 in Rac/Pak signaling and cell migration. J Biol Chem 2006;281:36443-53. [21] Mukherjee A, Wilson EM, Rotwein P. Selective signaling by Akt2 promotes bone morphogenetic protein 2–mediated osteoblast differentiation. Mol Cell Biol 2010;30:1018-27. [22] Fischer-Posovszky P, Tews D, Horenburg S, Debatin KM, Wabitsch M. Differential function of Akt1 and Akt2 in human adipocytes. Mol Cell Endocrinol 2012;358:135-43. [23] Gutierrez A, Snyder EL, Marino-Enriquez A, et al. Aberrant AKT activation drives well-differentiated liposarcoma. Proc Natl Acad Sci U S A 2011;108:16386-91. [24] Cenni V, Bavelloni A, Beretti F, et al. Ankrd2/ARPP is a novel Akt2 specific substrate and regulates myogenic differentiation upon cellular exposure to H(2)O(2). Mol Biol Cell 2011;22:2946-56. [25] Gonzalez E, McGraw TE. Insulin-modulated Akt subcellular localization determines Akt isoform-specific signaling. Proc Natl Acad Sci U S A 2009;106:7004-9. [26] de Frias M, Iglesias-Serret D, Cosialls AM, et al. Akt inhibitors induce apoptosis in chronic lymphocytic leukemia cells. Haematologica 2009; 94:1698-707. [27] Zhao Z, Leister WH, Robinson RG, et al. Discovery of 2,3,5trisubstituted pyridine derivatives as potent Akt1 and Akt2 dual inhibitors. Bioorg Med Chem Lett 2005;15:905-9. [28] Saji M, Narahara K, McCarty SK, et al. Akt1 deficiency delays tumor progression, vascular invasion, and distant metastasis in a murine model of thyroid cancer. Oncogene 2011;30:4307-15.
