KEY POINTS FOR CD117 IMMUNOASSAYS
- Select one of the rabbit monoclonal antibody (rmAb) clones YR145 and EP10, both available as concentrated (Conc) and ready to use (RTU) formats.
- Do not use the polyclonal antibody (pAb) cat. no. A4502 or rmAb clone 9.7.
- For laboratory-developed assays, use an optimized protocol based on carefully calibrated antibody titer, efficient heat-induced epitope retrieval (HIER) in an alkaline buffer and a sensitive 3-step polymer/multimer detection system.
- As control tissue, use appendix in which Cajal cells of the muscularis propria must show an at least moderate, specific staining reaction.
CD117, also known as mast cell growth factor receptor and proto-oncogene c-Kit, is as a cell membrane receptor tyrosine protein consisting of an extracellular domain that functions as a receptor, and an intracellular domain that, upon receptor ligand binding, forms a homodimer and activates kinase activity. The ligand for CD117 is stem cell factor (KIT-ligand), which induces cell survival, proliferation, and differentiation.1
Among normal tissues, CD117 is particularly expressed in hematopoietic stem cells, mast cells, melanocytes, and interstitial cell of Cajal. These cells mostly display a strong membranous and cytoplasmic immunohistochemical (IHC) staining reaction. In addition, CD117 is expressed in various epithelial cells (breast, sweat and salivary glands, renal tubular cells, and thyroid follicular cells), where a weak to moderate cytoplasmic staining reaction is seen. Finally, neurons of the central nervous system express CD117.
IHC detection of CD117 is important in classification of mesenchymal tumors of the gastrointestinal tract, where the vast majority of gastrointestinal stromal tumors (GIST) are positive for CD117.2 This is of clinical importance, as the tyrosine kinase inhibitor imatinib may be used in treatment of this tumor. In germ cell neoplasias, CD117 is expressed in the majority of seminomas/dysgerminomas and a minority of embryonal carcinomas.3 Precursor germ cell neoplasia in situ (GCNIS) lesions of testis display a distinct membranous staining reaction. In addition, Merkel cell carcinoma and follicular and papillary thyroid carcinoma are consistently positive for CD117. Epithelial tumors from lung, breast, colon, prostate, bladder, esophagus, and pancreas varyingly express CD117 (with little clinical consequence).
The Nordic immunohistochemical Quality Control (NordiQC) external quality assessment (EQA) program has assessed CD117 IHC assays 6 times during 14 years. A large proportion of submitted stains have been assessed as insufficient. The aim of the present paper is to give an overview of the EQA results for CD117 in the 2 latest assessments, run 47 (2016) and 51 (2017). Detailed assessment reports are available on the NordiQC homepage (http://www.nordiqc.org).
MATERIALS AND METHODS
Six EQA runs for CD117 have been accomplished: Run 4, 7, 14, 21, 26, 47, and 51. This article will focus on the results from the 2 most recent runs (47 and 51) only.
Tissue microarrays (TMAs) were constructed for each run and included: appendix, desmoid tumor, GIST (2 tumors with high-level and low-level CD117 expression, respectively) and testis with GCNIS (run 51 only). Before slide sets were circulated to the participants, sections from different levels of the TMAs were evaluated by the NordiQC reference laboratory to monitor the level of CD117 expression throughout the blocks. The GIST with high-level CD117 expression showed a strong distinct staining of virtually all neoplastic cells, while the GIST with low-level CD117 expression displayed a weak to moderate distinct staining reaction in the majority of neoplastic cells. In the testis with GCNIS a strong membranous staining reaction was seen in all neoplastic cells. In appendix, a strong membranous and cytoplasmic staining reaction of Cajal cells was shown in the lamina muscularis propria. In addition, faint staining reaction of the appendiceal crypts (basal compartment) and strong staining of mast cells was seen. The desmoid tumor was included as negative tissue control.
Participants submitted their assay protocols on the NordiQC homepage and received a pair of slides cut from the TMA described above. After staining, the participants performed the IHC assays for CD117 according to the protocol submitted, and returned one of the slides to NordiQC. An expert assessor panel consisting of experienced pathologists and technicians performed the assessment. The specific staining criteria for each of the assessment marks were defined in advance. Criteria for an optimal staining included a reaction pattern corresponding to that obtained by the reference laboratory (see above). A complete overview of the criteria for each tissue core can be found in the assessment reports on the NordiQC homepage (http://www.nordiqc.org). The assessor panel, evaluating each slide on a microscope linked to a projector, performed an anonymized consensus assessment. Each slide was marked as “optimal,” “good,” “borderline,” or “poor” based on the technical quality and concordance to the staining pattern and criteria outlined above. Optimal and good are considered sufficient, while borderline and poor are insufficient for diagnostic use.
In total, 549 laboratories participated in run 47 (n=272) and run 51 (n=277). For this overview, data from the 2 runs have been compiled. In total, 108 (20%) slides were assessed as optimal, 197 (36%) good, 176 (32%) borderline, and 68 (12%) poor. Grouped together, 44% of submitted slides were assessed as insufficient for clinical diagnostic use.
In 77% (187/244) of the insufficient results the main cause was a too weak or false negative staining reaction of cells expected to stain (especially Cajal cells of the appendiceal muscularis propria and GISTs with low-level CD117 expression). The remaining 23% (57/244) was characterized by poor signal to noise ratio and/or false-positive staining of smooth muscle cells and neoplastic cells in the desmoid tumor. Figures 1 and 2 illustrate optimal and insufficient staining reactions.
Seven different clones were used in Conc formats, while the number of RTU formats increased from 10 in run 47 to 11 in run 51, based on 6 and 5 antibody clones, respectively. Three Ab clones (YR145, EP10, and T595) were available both as Conc and RTU formats. Table 1 summarizes the assessment result for the most commonly used (>10 participants) antibody clones. Conc and RTU products based on the same clone (but from different vendors) gave highly concordant results and were therefore pooled together.
Used within a laboratory develop assay (Conc), the 2 rmAb clones YR145 and EP10 provided high pass rates of 91% and 100% (of which 75% and 50%, respectively, were optimal). The main reason for suboptimal staining reaction was use of less sensitive 2-step detection systems.
The most commonly used Ab was the Conc pAb A4502 that gave a significant lower pass rate of 50% with only 7% optimal. It was possible to obtain optimal staining results with pAb A4502 on all the major staining platforms Autostainer (Dako), Omnis (Dako), Bond (Leica), and Benchmark (Ventana) although the proportion of optimal results was very low. The characteristics for insufficient staining results were background staining, or weak or negative staining of cells expected to be positive. In addition, the pAb A4502 in some cases provided false-positive staining of the desmoid tumor. As no single protocol parameter could be identified explaining this aberrant staining result, lot to lot variation may be a plausible explanation.
The rmAb clones YR145 and EP10 as RTU formats provided high pass rates of 79% and 81% of which 53% and 62%, respectively, were optimal. This was somewhat lower than the Conc formats. Insufficient staining results were (just as for the Conc format) mostly related to use of low-sensitivity 2-step detection systems.
The majority of RTU assays were based on the rmAb 9.7 which provided a low pass rate of 33% of which only 2% (1 protocol) was optimal. Despite using protocol settings similar to protocols obtaining optimal results with the robust rmAbs (eg, YR145 and EP10), optimal performance could not be accomplished indicating that this clone may be technically challenging and difficult to optimize.
The 2 latest NordiQC EQA runs for CD117 revealed significant and consistent challenges for diagnostic laboratories. Several factors influenced the overall performance, although the most important was choice of the primary Ab. Among the most commonly used Abs, the rmAbs YR145 and EP10 provided significantly higher pass rate and proportion of optimal results both as Conc and RTU formats compared with the commonly used pAb A4502 and rmAb 9.7.
In the 2 latest NordiQC runs, the overall pass rate was 56%, and this finding is somewhat different from those published by another IHC EQA organization, UK NEQAS, journal number 119, where the assessor group rated 91% (107/118) of the submitted slides as “acceptable.”4 Noteworthy, 90% of the stains (n=84) based on the pAb A4502 were rated as acceptable. Moreover, as described in the NordiQC results, the assessor group of UK NEQAS found that the main feature for an insufficient performance was a weak/false negative or false-positive staining result. The assessment criteria used were similar to the assessment criteria applied by the NordiQC assessor group for CD117. In previous runs for CD117, the pass rates have been nearly identical in the 2 proficiency-testing programs (UK NEQAS and NordiQC). However, the discrepancy in pass rates obtained in the latest run for CD117 between these 2 assessment programs is difficult to elucidate on, as the included tissue was of same origin and diagnosis (not accounting for biological variations). In the initial phase in selection of appropriate tissue for assessments (Run 47 and 51, NordiQC), the reference method used to test included tissue was based on the rmAb YR145 in combination with HIER in an alkaline buffer and a sensitive 3-step multimer detection system, setting a new standard for optimal performance. Therefore, critical evalution of staining results may account for the overall downgrading of the “less efficient” primary Abs (eg, pAb A4502 or rmAb 9.7) by the assessor group of the NordiQC organization, bearing in mind that new robust rmAb clones produce stronger, more specific, and sensitive results.
The choice of primary Ab and the use of appropriate protocol settings are of immense importance. As shown in both assessment programs (UK NEQAS and NordiQC), false-positive staining was observed with the pAb A4502 in the desmoid tumors, which demonstrates the challenges that clinical diagnostic pathologists face with suboptimal IHC assays. These results are in line with the findings of Miettinen et al,2 but in conflict with the results obtained by Wong and Melegh, demonstrating that a variety of soft tissue tumors were positive for CD117 but none of desmoid tumors (n=6).5 It must be emphasized, that the results were based on less efficient/sensitive protocol settings as only no or HIER in acidic buffer was tested, which could explain the lack of reaction in desmoid tumors. One might speculate, what the final outcome would have been if HIER in an alkaline buffer was applied. Desmoid tumors have been consistently negative in NordiQC assessments for CD117 with the rmAbs YR145 and EP10, irrespectively of protocol settings applied, for example, HIER in an alkaline buffer.
Reliable positive and negative on-slide tissue controls are important in any IHC assay to ensure correct analytical sensitivity levels and to identify specificity problems and technical reaction errors.6–8
On the basis of the NordiQC results for CD117, appendix should be included in the control material, since insufficient staining results often were identified in this tissue. In appendix, an at least moderate staining reaction of the Cajal cells in the muscularis propria should be seen. In addition, strong membranous staining of mast cells and a weak staining reaction of epithelial cells in the basal compartment of the crypts is seen. All other cells should be negative. As CD117 can be defined as a IHC-type II marker, it can be argued, that GISTs (with a broad spectrum of antigen densities) and desmoid tumors (expected negative staining result) should be a part of the control material due to the diagnostic challenges mentioned above and clinical management of the patients.
1. Miettinen M, Lasota J. KIT (CD117
): a review on expression in normal and neoplastic tissues, and mutations and their clinicopathologic correlation. Appl Immunohistochem Mol Morphol. 2005;13:205–220.
2. Miettinen M, Sobin LH, Sarlomo-Rikala M. Immunohistochemical spectrum of GISTs at different sites and their differential diagnosis with a reference to CD117
(KIT). Mod Pathol. 2000;13:1134–1142.
3. Lau SK, Weiss LM, Chu PG. D2-40 immunohistochemistry
in the differential diagnosis of seminoma and embryonal carcinoma: a comparative immunohistochemical study with KIT (CD117
) and CD30. Mod Pathol. 2007;20:320–325.
5. Wong NA, Melegh Z. Antigen retrieval and primary antibody type affect sensitivity but not specificity of CD117 immunohistochemistry
. Histopathology. 2009;54:529–538.
6. Torlakovic EE, Francis G, Garratt J, et al. Standardization of negative controls in diagnostic immunohistochemistry
: recommendations from the international ad hoc expert panel. Appl Immunohistochem Mol Morphol. 2014;22:241–252.
7. Torlakovic EE, Nielsen S, Francis G, et al. Standardization of positive controls in diagnostic immunohistochemistry
: recommendations from the International Ad Hoc Expert Committee. Appl Immunohistochem Mol Morphol. 2015;23:1–18.
8. 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.
Keywords:Copyright 2019 Wolters Kluwer Health, Inc. All rights reserved.
immunohistochemistry; quality control; CD117; NordiQC