KPNA2 expression is a potential marker for differential diagnosis between osteosarcomas and other malignant bone tumor mimickers

Background Karyopherin α 2 (KPNA2), a member of the Karyopherin α family, has been observed in several cancers but lack substantial investigation in malignant bone tumors. The purpose of the current study was to evaluate KPNA2 expression level and its utility as a novel diagnostic biomarker in osteosarcomas and their malignant bone tumor mimickers, such as chondrosarcomas and Ewing sarcomas. Method We investigated the expression of KPNA2 protein by immunohistochemistry on paran embedded surgical specimens from 217 patients with malignant and benign tumors of bone, including 81 osteosarcomas, 42 chondrosarcomas, 9 Ewing sarcomas, 28 osteoid osteoma, 20 osteochondroma and 37 Chondroblastoma. Immunoreactivity was scored semi quantitatively based on stain extent and intensity. Seventy one of 81 (87.7%) osteosarcomas, zero of 42 (0%) chondrosarcomas and one of 9 (11.1%) Ewing sarcomas showed immunoreactivity for KPNA2. Negative KPNA2 expression was observed in all of benign bone tumors. Much more positive expression of KPNA2 was found in osteosarcomas as compared with chondrosarcomas and Ewing sarcomas. The sensitivity and specicity of KPNA2 immuno-expression for osteosarcoma was 87.7% and 100%, respectively. In several subtypes of osteosarcomas, immunohistochemical expression of KNA2 was more frequent in osteoblastic (94.5%), with 39 (70.9%) showing strong-intensity staining. KPNA2 positivity was observed in eleven of 13 (84.6%) chondroblastic, three of 6 (50%) broblastic, three of 4 (75%) telangiectatic and two of 3 (66.7%) giant cell-rich osteosarcoma. Stronger-intensity staining was observed in osteoblastic osteosarcoma.

Conclusion KPNA2 is most frequently expressed in osteosarcomas, particularly in osteoblastic and chondroblastic tumors, but is rarely positive in chondrosarcomas and Ewing sarcomas. This feature may have diagnostic value since it is very useful for distinguishing between osteosarcomas and other bone sarcomas mimickers. This report supports KPNA2 as a novel marker for the diagnosis of osteosarcoma. Background Osteosarcoma is de ned by the presence of malignant cells producing osteoid or immature bone, representing the most common primary skeletal sarcomas with high prevalence in children, adolescents and young adults [1]. Surgery combined with neoadjuvant chemotherapy is the main therapeutic strategy for the treatment of osteosarcoma patients, and the ve-year survival rate has markedly improved to over 60% in patients with localized tumor [2]. Accurate distinction between osteosarcoma and other sarcomas of bone is very important, as chemotherapeutic and surgical approaches differ signi cantly that dependent mainly on the histopathological diagnosis [3].
The identi cation of osteoid matrix is elemental for osteosarcoma diagnosis. However, this diagnostic feature may be extremely limited in other cases and the diagnosis becomes challenging when minimal or scant osteoid matrix formation is identi ed on biopsy. Immunohistochemistry has been evaluated as an adjunct to offer additional information to support the nal diagnosis [4]. Early investigations suggested that osteonectin was a sensitive and speci c marker of osteoblastic differentiation that could be helpful in the diagnosis of bone tumors, particularly osteosarcomas [5]. Subsequent studies, however, have demonstrated that osteonectin can be detected in other primary bone sarcomas, including Ewing sarcoma (ES) and chondrosarcoma [6]. Some proteins have been identi ed such as DMP-1,CADM1, galectin-1 and NDRG1 as potential markers of osteosarcoma, but their practical utility in the diagnosis of bone tumor is unclear [7][8][9][10]. Given the role of SATB2 in regulation of osteoblast differentiation, immunohistochemistry for SATB2 has shown promise as a highly sensitive marker of osteoblasts [11].
However, SATB2 positivity is not speci c for osteosarcoma as compared with other primary bone sarcomas has been well documented by Davis and Horvai [12]. Therefore, it is urgently needed to identify highly sensitive and speci c biomarkers for osteosarcoma differential diagnosis.
Karyopherin α-2 (KPNA2), an adaptor protein, is a member of the karyopherin-α protein family that plays a crucial role in transportation of proteins from the cytoplasm into the nucleus [13]. Work together with importin-β, KPNA2 mediates numerous nuclear translocations of target proteins aimed by nuclear localization signal passing through the nuclear pore complex [14]. Previous studies have demonstrated that KNPA2 is a potential biomarker in multiple forms of cancer, including breast cancer, lung cancer, gastric cancer, colon cancer, prostate cancer and upper tract urothelial carcinoma and so on [15][16][17][18].
Expression of KPNA2 has been reported to be tied with poor prognosis in patients with oesophageal squamous cell carcinoma, epithelial ovarian carcinomas and small hepatocellular carcinoma [19][20][21]. However, in human malignant bone tumors, the expression level of KPNA2 is unclari ed.
The purpose of the present study was to evaluate the expression of KPNA2 in osteosarcoma, chondrosarcomas and ES samples using immunohistochemistry in order to con rm its potential diagnostic utility as a novel molecular marker for discriminating osteosarcoma from other primary bone sarcomas. Histopathological diagnosis of each biopsy was made according to the WHO classi cation. The clinicopathological features of patients with malignant bone tumors were recorded in Table 1.

Immunohistochemical analysis
Immunohistochemistry was performed on 4-μm para n-embedded whole tissue sections with standard techniques. Brie y, following depara nization and rehydration, charged slides with 4-μm thick sections of tissue were treated with 3% hydrogen peroxide to eliminate endogenous peroxidase activity, then processed for antigen retrieval with 10 mm citrate buffer pH 6.0 for 15 min at 95 °C, followed by incubation in 5% bovine serum albumin (BSA) for 20 min at room temperature. All sections were rinsed with phosphate-buffered saline (PBS) and incubated overnight at 4 °C with monoclonal mouse antihuman antibody against KPNA2 (sc-55538, 1:100 dilution, Santa Cruz Biotechnology, Santa Cruz, USA). The slides were washed with PBS and then incubated at room temperature with the secondary antibody for 20min. Diaminobenzidine (DAB) staining was then performed. Cervical cancer cases were used as positive controls. Negative control sections were prepared by substituting non-immune IgG for the primary antibody.

Analysis of immunohistochemical staining
The percentage of KPNA2 positive cells and the staining intensity were scored in a semiquantitative manner. Immunohistochemical slides were scanned and evaluated by two experienced researchers (Lucen Jiang and Qingzhu Wei) who were blinded to the clinical data of the patients. The proportion of cells with nuclear KPNA2 staining was scored as follows: 0, no staining; 1+, <5% positive; 2+, 5-25% positive; 3+, 26-50% positive, 4+, 51-75% positive; and 5+, 76-100% positive. Staining intensity was also graded as weak, moderate, or strong.

Patient characteristics
Clinicopathological characteristics of patients with malignant bone tumors are shown in Table 1 The immunohistochemical ndings in malignant and benign bone tumors were summarized in Table2. Of the osteosarcomas, 71 of 81 (87.7%) were positive for KPNA2. KPNA2 expression was predominantly observed in the nuclei of tumor cells, with little expression observed in the cytoplasm of either normal or tumor cells (Fig. 1 A&B). To evaluate the utility of KPNA2 in distinguishing tumors of osteoblastic origin from histological mimics, we examined chondrosarcomas and ES. As shown in Table 2, KPNA2 was expressed in a very low proportion of chondrosarcomas (0) and ES (11.1%) cases than osteosarcomas, very weak and diffuse staining was detected in sole case of tumors ( Fig. 1 C-F). All cases of benign bone tumors encompassing osteoid osteoma, osteochondroma and chondroblastoma were negative for KPNA2 ( Table 2; Fig. 2 A-F). Therefore, sensitivity and speci city of KPNA2 immuno-expression were 87.7% and 100% respectively in osteosarcomas. Obviously, higher proportion and stronger intensity of  Table 3. Of the osteoblastic, fty-two of 55 (94.5%) were positive for KPNA2, with 39 (70.9%) showing strong-intensity staining (Fig. 3 A&B). KPNA2 was expressed in eleven of 13 (84.6%) chondroblastic, with 7 (53.8%) strong-intensity staining (Fig. 3 C&D). Positive KPNA2 expression was observed in three of 6 (50%) broblastic osteosarcomas, with 2 (33.3%) strong-intensity staining (Fig. 3 E&F). Three of 4 (75%) in telangiectatic and two of 3 (66.7%) giant cell-rich osteosarcoma KPNA2 expression were positive staining, with 50% and 33.3% staining respectively (Fig. 3 G-J). When KPNA2 positive expression was evaluated separately in the various components of conventional osteosarcoma, areas with osteoblastic and chondroblastic showed the greatest extent of staining, whereas more limited staining was observed in broblastic components. Telangiectatic and giant cell-rich osteosarcoma were also reactive for KPNA2, but the pattern of staining was generally not as extensive as in osteoblastic osteosarcomas. Overall, both extent and intensity of KPNA2 immuno-expression in osteoblastic osteosarcomas is the greatest among types of bone tumors.

Discussion
In this study, we carried out a comparative analysis of KPNA2 expression pro les in osteosarcoma, chondrosarcomas, ES and several benign bone tumors based on clinical samples, and discovered KPNA2 expression was signi cantly obvious in osteosarcoma patients. KPNA2 can be served as a novel promising biomarker for diagnosing osteosarcoma because of this characteristic. In addition, we observed that KPNA2 positive staining was markedly higher proportion and stronger intensity in osteoblastic cases than other subtypes. Osteosarcoma is the most common primary malignant bone tumor in children, adolescents, and young adults. Cancer registry data with histological strati cation indicate that osteosarcoma cases accounted for approximately 35 percent of cases, followed by chondrosarcoma (25%), and ES (16%). Accurate diagnosis of each type of sarcoma is critical to ensure appropriate therapy for clinical patients. Among the three common malignancies of bone, selection of treatment regimens with adjuvant or neoadjuvant therapy in conjunction with surgical resection varies dramatically, depending on the pathological diagnosis. The current standard of therapy for localized osteosarcoma is neoadjuvant chemotherapy followed by complete surgical resection of the primary disease and then adjuvant chemotherapy. Multiple agents include methotrexate, doxorubicin and cisplatin (MAP) are recommend for osteosarcoma patient chemotherapy as the rst-line treatment [22][23][24]. In contrast, multiple studies have found conventional chondrosarcoma was chemotherapy-resistant, possibly because the systemic agents are di cultly delivered to malignant cells in cartilaginous matrix, and no clear evidence for neoadjuvant or adjuvant therapy would be suggested. Radiation therapy is sometimes recommended for positive surgical margins [25][26][27]. Most patients with ES start treatment with neoadjuvant chemotherapy for approximately 12 weeks followed by local control with radiation, surgery, or both. Further chemotherapy is administered to reach a cumulative total of at least 14 cycles. Radiation and chemotherapy with a different combination of agents have been shown prominent effect on ES patients [28][29][30]. The need for immunohistochemical to con rm a diagnosis of osteosarcoma over other high-grade bone sarcomas is very important when bone or osteoid production are not overtly apparent on biopsy. Some of the osteosarcoma can be accurately classi ed based on a sensitive biomarker, STAB2. As a nuclear antigen directly involved in osteoblast lineage commitment, SATB2 is an attractive target for immunohistochemical identi cation of osteoblasts. However, the results obtained by Davis and Horvai suggested that SATB2 positivity is not speci c for osteosarcoma as compared with other primary bone sarcomas [12]. Therefore, an additional biomarker to assist in accurately stratifying patients into the correct diagnostic category of bone sarcomas would be bene cial.
In recent years, KPNA2 has emerged as a potential biomarker in multiple solid tumor types and its aberrant expression is often associated with poor prognosis in the patients. There was a trend toward lower degree of differentiation and higher pathological states in the high KPNA2 expression group. Alshareeda et al. found that KPNA2 signi cantly contributed to aberrant localization of key proteins and predicted poor prognosis of breast cancer [15]. Altan et al. proved KPNA2 expression in primary lesions and metastatic lymph nodes was associated with poor prognosis and progression in gastric cancer [17]. Zhang et al. reported that KPNA2 was a novel prognostic marker and a potential therapeutic target for colon cancer [18]. In our study, regarding the scenario osteosarcoma versus chondrosarcoma and ES, we observed KPNA2 immunoreactivity in 81 osteosarcoma, 42 chondrosarcomas and 9 ES sections. The results of the current study indicated that a large proportion of osteosarcoma is KPNA2 positive expression, only one pattern KPNA2 staining scored as positive, with weak intense in ES and none was found in chondrosarcoma. Our research revealed that the KPNA2 expression is a sensitive and speci c marker for osteosarcoma as compared with chondrosarcomas and ES. This is the rst report of a correlation between KPNA2 expression and sarcoma. The results of the present study showed that positive expression of KPNA2 was signi cantly observed in osteosarcoma compared with benign bone tumors tissues. Our data demonstrated that KPNA2 is su ciently reliable to be used as a diagnostic marker in OS biopsies.
We did note that proportion of positive KPNA2 expression was higher and strong immunostaining was more common in osteosarcoma than in other high-grade bone sarcomas. However, variations in the extent and intensity of staining were observed within the osteosarcoma subtypes, from weak to diffuse and strong nuclear reactivity. Notably, KPNA2 expression is observed at higher proportion in osteoblastic cases that generally showed stronger staining than other subtypes. In chondroblastic, all but two cases expressed KPNA2 (84.6%) with 53.8% strong-intensity staining that is next only to osteoblastic. Moreover, despite the cases of broblastic, telangiectatic and giant cell-rich osteosarcoma are less than six, the tendency of KPNA2 positive expression keeping pace with osteoblastic and chondroblastic. This result appears promising and further broad studies with larger cohorts to validate the ndings are warranted. All cases of osteoid osteoma, osteochondroma and chondroblastoma were KPNA2-negative. It is consistent with those of previous studies that the benign tissues have been demonstrated as negative to KPNA2 expression [31].

Conclusion
In summary, we found for the rst time that KPNA2 immunohistochemical expression was highly sensitive and speci c in osteosarcoma compared with chondrosarcoma and Ewing sarcoma. This study reveals that KPNA2 immuno-expression may be a potential marker for differentiating osteosarcoma, particularly osteoblastic and chondroblastic subtypes, from chondrosarcoma and Ewing sarcoma. Although the identi cation of malignant osteoid matrix formation and the integration of clinical and radiological data remain the corner stone of osteosarcoma diagnosis, this case strongly supports that the KPNA2 expression as an additional diagnostic marker can improve the diagnosis of osteosarcoma combination with histological features.