KEY POINTS FOR SOX10 IMMUNOASSAY
- Careful calibration of the primary antibody is required.
- Sufficient HIER in an alkaline buffer is required.
- Skin and colon are recommended, but not perfect, tissue controls.
- Do not use polyclonal SOX10 antibodies.
The protein SOX10 is a transcription factor belonging to the SOX (SRY-related HMG-box) family involved in regulation of embryonic development and cell differentiation.1 In embryogenesis, SOX10 is first detected in the emerging neural crest. It is involved in maturation and maintenance of Schwann cells and melanocytes. SOX10 is a structural protein in chromatin activating transcription of a number of genes, including microphthalmia transcription factor. In adults, SOX10 is expressed in melanocytes, Schwann and satellite cells, and oligodendrocytes.
SOX10 is expressed in virtually all cases of nevi, malignant melanoma, Schwannoma, neurofibroma, and granular cell tumor. SOX10 is moreover expressed in the majority of cases of oligodendroglioma, astrocytoma, acinic cell carcinoma, myoepithelial carcinoma, and pleomorphic adenoma. SOX10 is expressed in about half of cases of malignant nerve sheath tumor. Apart from a group of breast ductal carcinomas, epithelial, lymphomas, and mesenchymal tumors are SOX10 negative.
In diagnostic settings, SOX10 is particularly used to identify tumors of melanocytic lineage. Studies suggest that SOX10 has superior sensitivity and specificity in detecting malignant melanoma when compared with S1002 and Melan A.3
The Nordic Immunohistochemical Quality Control (NordiQC) external assessment scheme has assessed SOX10 in 2015 (run 45) and 2016 (run 48) (www.nordiqc.org).
MATERIALS AND METHODS
Tissue microarrays (TMAs) for the 2 NordiQC runs 45 (2015) and 48 (2016) included: 1 Schwannoma, 2 malignant melanomas, 1 colon adenocarcinoma, breast tissue with hyperplastic glands (run 45 only), normal skin, and normal colon. A series of melanomas was initially screened and 2 tumors, based on the high-level and moderate-level expression profiles, respectively, were selected for the SOX10 TMAs. The TMAs were evaluated by NordiQC reference laboratories and also showed that the Schwannoma displayed high-level expression of SOX10, while the colon adenocarcinoma was negative. Normal skin and colon served as both positive and negative tissue controls, where all melanocytes and myoepithelial cells of sweat glands in the skin and Schwann cells of the colon expressed SOX10, whereas all other cells were negative.
Specific staining criteria for each of the assessment marks were defined before the assessment. The criteria for an optimal SOX10 included: a generally high intensity of cells expected to be stained (melanocytes, Schwann cells, all malignant melanomas, and the Schwannoma), whereas nonspecific staining should be as faint as possible and not interfere with the interpretation. A complete overview of the criteria can be found on the NordiQC homepage (www.nordiqc.org). All participants submitted their assay protocol on the NordiQC homepage and received a set of slides cut from the TMA described above. After performing the assay (using the protocol submitted to NordiQC), slides were returned to NordiQC. An anonymized consensus assessment was performed by an expert panel of pathologists and technicians, examining each slide on a microscope linked to a projector. Each slide was marked as “optimal,” “good,” “borderline,” or “poor” based on the technical quality and adherence to the staining criteria.
Run 45, 2015
In total, 86 laboratories returned slides to NordiQC. A total of 23 (27%) were assessed as optimal, 16 (18%) as good, 35 (41%) as borderline, and 12 (14%) as poor. Sufficient technical quality (optimal or good) was obtained for 45% of the submitted protocols in this run.
Of the insufficient results, 29 of 47 (62%) were related to a weak or false negative staining in ≥1 cores expected to be positive, especially the low-level SOX10 expressing melanoma. Of the remaining 38% insufficient results, both too weak specific and excessive background staining was observed. This pattern was especially seen when polyclonal antibodies (pAbs) were used (Figs. 1F2–3).
The mouse monoclonal antibody (mmAb) BC34 and pAb 383A were the most commonly used clones. BC34 outperformed 383A, as none of slides stained with 383A were assessed as “optimal.” The pass rate of BC34 was also significantly higher compared with pAb 383A (60% and 21%, respectively). For the BC34 clone, optimal results required heat-induced epitope retrieval (HIER), preferably in an alkaline buffer.
Run 48, 2016
In total, 128 laboratories returned slides. One participant used an unknown primary antibody clone staining other epitopes than SOX10 and was excluded from this analysis. Of the remaining 127 slides, 63 were assessed as “optimal” (50%), 24 as “good” (19%), 27 as “borderline” (21%), and 13 as “poor” (10%), meaning that 69% of the participants submitted a sufficient assay.
Of the insufficient results, 34 of 40 (85%) were related to a weak or false negative staining reaction in one more of the cores, especially the low-level SOX10 expressing melanocytes of the melanoma (core 6) and the normal melanocytes and myoepithelial cells of the normal skin.
The most commonly used primary antibody clones were mmAb BC34 (n=48) and rabbit monoclonal antibody (rmAb) EP268 (n=47), both available in concentrated and ready-to-use formats. The overall pass rate for the BC34 was 69% of which 56% were optimal. Optimal results could be obtained on all major immunohistochemistry (IHC) stainer platforms. Applied as the concentrated format by laboratory developed assay, 72% of slides were sufficient of which 59% were optimal. When applied as a ready-to-use format (different vendors) only 50% of protocols were sufficient. With concentrated formats, shorter mean HIER time (29 vs. 49 min.) and more diluted primary Abs (mean dilution 1:132 vs. 1:52) was seen in insufficient results compared with sufficient. Laboratory developed assays based on the rmAb clone EP268 provided a pass rate of 74% of which 44% were optimal.
Eleven of the participants used pAbs (Fig. 3). Eight of these were assessed as borderline, where 3 were assessed as poor. None was assessed as technically sufficient.
The number of submitted protocols presented in the study deviate from the data presented in the original assessment report for run 48. This is due to several participants returned their slide after the deadline for the report. Complete and in-depth analysis of protocol parameters and assessment marks as well as optimized protocols can be found on www.nordiqc.org.
Several staining protocol parameters influenced the overall performance: insufficient HIER (to short HIER time and low pH), inappropriate or too low concentration of the primary antibody, and use of less sensitive detection systems were the main issues.
None of the assays based on pAbs (n=11, 9%) obtained a sufficient assessment mark in run 48. These slides were characterized by an overall weak staining of nuclei expected to be positive, often combined with aberrant and excessive background staining. This combination of low analytical sensitivity and specificity would compromise correct interpretation in clinical setting.
In the first assessment for SOX10 (run 45, 2015) 45% (39 of 86) of all participants used pAbs. Although a decrease in the proportion of participants using pAbs was observed in run 48 (2016), the overall results indicates that the use of pAbs should be discontinued, especially considering the number of excellent monoclonal alternatives available.
Use of reliable positive and negative tissue controls is an important issue that must be addressed in any immunohistochemical staining reaction.4 SOX10 is a challenging marker in this sense, as no easy accessible tissues with consistent low-level expression have been identified at this time. Several potential external control tissues were included in run 45 and 48, including skin, colon, and hyperplastic breast tissue (run 45 only). In skin, moderate to strong staining reaction was seen in melanocytes in optimal protocols. Myoepithelial cells lining sweat glands showed a moderate staining reaction. This was also observed in the myoepithelial cells of the benign breast hyperplasia. In colon, Schwann cells showed a strong nuclear staining reaction. At present the recommendation is to include both skin and colon as external tissue controls. The cells listed must show a distinct nuclear Sox10 expression at a level minimum as mentioned above, whereas other cells and cellular compartment must be unstained.
1. Mollaaghababa R, Pavan WJ. The importance of having your SOX on: the role of SOX10 in the development of neural crest-derived melanocytes and glia. Oncogene. 2003;22:3024–3034.
2. Vrotsos E, Alexis J. Can SOX-10 or KBA.62 replace S100 protein in immunohistochemical evaluation of sentinel lymph nodes for metastatic melanoma? Appl Immunohistochem Mol Morphol. 2016;24:26–29.
3. Willis BC, Johnson G, Wang J, et al. SOX10: a useful marker for identifying metastatic melanoma in sentinel lymph nodes. Appl Immunohistochem Mol Morphol. 2015;23:109–112.
4. Torlakovic EE, Nielsen S, Vyberg M, et al. Getting controls under control: the time is now for immunohistochemistry. J Clin Pathol. 2015;68:879–882.