The Spectrum of Kit (CD117) Immunoreactivity in Lung and Pleural Tumors: A Study of 96 Cases Using a Single-Source Antibody With a Review of the Literature

Despite significant progress in understanding the pathogenesis and molecular biology of lung cancer, it remains the leading cause of cancer deaths in the United States, with an overall 5-year survival rate around 15%.1 Pleural malignant mesothelioma (MM), another aggressive intrathoracic malignancy, has a survival rate similar to or worse than that of lung cancer.2 The success of imatinib (STI571, Gleevec, Novartis, Basel, Switzerland), a novel tyrosine kinase inhibitor, in the treatment of such disparate neoplasms as chronic myelogenous leukemia and metastatic gastrointestinal stromal tumors (GISTs) has generated considerable interest in determining the efficacy of this agent in other neoplasms, particularly intractable tumors such as small cell lung carcinoma (SCLC) and MM.3,4 

Kit (CD117), the product of the proto-oncogene c-kit, is a tyrosine kinase transmembrane receptor. Binding of Kit with its ligand stem cell factor induces an array of intracellular responses critical to regulating the development and growth of some human cell types.5,6 Immunohistochemical staining for Kit has been described in a variety of tumors, including myelogenous and mast cell leukemias, seminoma, germ cell tumors, angiosarcoma, Ewing and clear cell sarcomas, neuroblastoma, and melanoma, as well as thyroid, endometrial, ovarian, and breast carcinomas.7–12 Kit positivity is in fact a defining feature of GIST.13 Kit immunoreactivity has also been documented in a number of normal human cell types, including hematopoietic cells, mast cells, germ cells, melanocytes, adrenal medulla, and breast epithelium, as well as interstitial cells of Cajal, from which GISTs are postulated to originate.8,12,14–16 

While Kit staining is well recognized in SCLC, neither normal human lung parenchyma nor bronchial epithelium has been shown to stain for Kit, implying Kit is aberrantly expressed in SCLC.8,14,17 While a few reports describe limited Kit immunoreactivity in non–small cell lung cancers (NSCLCs), the proportion of positive cases varies considerably, as do the immunohistochemical methods used.8,12,14,17–21 In this article, we report the expression of Kit in the major types of pulmonary and pleural neoplasms by standard avidin-biotin complex (ABC) immunohistochemical technique with a commonly used, commercially available polyclonal Kit antibody and present a comprehensive review of the literature on Kit staining in pulmonary and pleural tumors.

Cases were retrieved from the routine surgical pathology files of Duke University Medical Center, Durham, NC, and the consultation files of one of the authors (S.P.H.). Material was obtained from formalin-fixed, thoracoscopic wedge biopsies or resection specimens. Hematoxylin-eosin–stained sections were available for review in all cases. Diagnoses were confirmed using standard and widely accepted criteria.22 For SCLC, large cell neuroendocrine carcinoma (LCNEC), and pulmonary carcinoid tumor (PCT), immunohistochemical stains for cytokeratins, chromogranin, and synaptophysin were routinely performed to support the diagnosis, and CD34 immunoreactivity was assessed in cases of localized fibrous tumor (LFT). Similarly, a panel of markers (cytokeratins, calretinin, cytokeratin 5/6, TTF-1, and carcinoembryonic antigen) was used in cases of MM.

Immunohistochemical staining on paraffin-embedded tissue was performed using an avidin-biotin-peroxidase complex method. No heat-induced epitope retrieval (HIER) or proteolytic enzyme treatment was used for the majority of cases. In non–enzyme-treated, non-HIER cases, sections from the paraffin blocks were cut at 4 to 5 μm, placed on positively charged slides, deparaffinized in organic solvents, treated with methanolic hydrogen peroxide to quench endogenous peroxidase activity, and rehydrated. Sections were reacted with antibody against Kit (CD117) (rabbit IgG polyclonal; Dako Corporation, Carpinteria, Calif; 1:200). Nonimmune murine immunoglobulins (1:500; Southern Biotechnologies Association, Inc, Birmingham, Ala) were used as a negative control, and appropriate positive tissue controls were tested. The unlabeled, bound primary antibody was linked with biotinylated goat anti-rabbit IgG (Vector Laboratories, Burlingame, Calif; 1:300) and was detected with horseradish peroxidase-labeled streptavidin (Jackson Immuno Research Inc, West Grove, Pa; 1:800). Immunoreactivity was visualized using diaminobenzidine as the chromogen, with Harris modified hematoxylin as the counterstain.

Heat-induced epitope retrieval was performed in 4 cases, all LFTs. Sections from the paraffin blocks were cut at 4 to 5 μm, placed on positively charged slides, and deparaffinized. Tissue sections were subsequently placed in a 10 mM concentration of citrate buffer, pH 6.0, steamed for 20 minutes, and allowed to stand for an additional 20 minutes in hot buffer. Slides were subsequently placed on a Ventana automated stainer (Ventana Medical Systems Inc, Tucson, Ariz) using a titration protocol and aminoethylcarbazole as the detection system. Primary antibodies against Kit (CD117) (rabbit IgG polyclonal; Dako; 1:25) were placed on the tissue sections and incubated for 30 minutes. After the staining run was complete, the tissue sections were removed from the stainer and counterstained with hematoxylin.

Immunoreactivity was assessed independently by 2 of the authors (K.J.B. and V.L.R.). Cases were divided into 4 categories: 3+ for diffuse staining (>75% of tumor cells) with intensity comparable to positive control tissue; 2+ for cases with similarly intense staining in 25% to 75% of tumor cells, as well as cases with diffuse but moderately intense staining; 1+ for moderate to intense staining in 10% to 25% of tumor cells; or absent.

In addition, the English-language literature was reviewed for previously reported cases of lung and pleural tumors in which Kit immunostaining had been performed.

We examined 60 lung carcinomas (11 SCLCs, 4 LCNECs, 22 NSCLC–squamous cell tumors [NSCLC-SQs], and 23 NSCLC-adenocarcinomas [NSCLC-ADs]), 11 PCTs (9 typical, 2 atypical), 19 pleural MMs (9 epithelial, 8 biphasic, and 2 sarcomatoid), and 6 LFTs for Kit immunohistochemical staining, the results of which are summarized in the Table. Additionally, a review of the literature yielded several studies in which Kit immunoreactivity of lung and pleural tumors had been determined (Table). In the present study, all positive cases demonstrated cytoplasmic staining with membranous accentuation in the majority of intensely staining cases. Nuclear staining was not observed. Small cell lung carcinomas showed Kit staining in 9 of 11 cases; staining was moderate to intense in 8 cases (Figure 1). One of 4 LCNECs demonstrated moderate immunoreactivity. Of the NSCLCs, moderate staining in approximately 10% to 25% of tumor cells was observed in 2 of 22 NSCLC-SQs and 4 of 23 NSCLC-ADs (Figures 2 and 3). None of the 11 PCTs were immunoreactive for Kit. Moderate immunoreactivity was present in 3 of 6 LFTs (Figure 4). No staining for Kit was identified in 18 of 19 MMs, including 9 epithelial, 7 biphasic, and 2 sarcomatoid MM (Figure 5). All 3 of the Kit-positive LFTs had been subject to HIER. One biphasic MM demonstrated focal moderate cytoplasmic Kit staining. Positive and negative controls stained appropriately.

Kit immunoreactivity was present in a high proportion of SCLCs in this study. This finding is concordant with most previous reports of Kit staining in SCLC.8,14,18,19,23,24 Two SCLCs in this study were nonimmunoreactive for Kit. Antibody against 34βE12 was used to explore whether these cases represented basaloid carcinomas rather than SCLC. Staining for 34βE12, which is present in a high percentage of basaloid carcinomas, was not observed in our 2 cases, suggesting that a small population of truly Kit-negative SCLCs exists.25 

A few studies have reported substantially lower rates of Kit positivity in SCLC.12,20,26,27 Because SCLC is not typically managed by surgical resection, diagnostic tissue is usually in the form of small specimens, such as transbronchial biopsies. Because tumors can exhibit heterogenous Kit staining, sampling is an important factor for immunodetection. We sought to minimize this problem by including only thoracoscopic wedge biopsies or larger specimens in this study. Although detailed information regarding the types of specimens examined was not provided in most previous studies of Kit staining in SCLC, the inclusion of cases obtained by transbronchial biopsy may have played a role in the rate of Kit detection.

Another factor that may impair immunodetection of Kit in SCLC, particularly in transbronchial biopsies, is crush artifact. It is well known that autolysis alters the molecular structure and thereby the antigenicity of tissues. It seems likely that crush artifact might also disrupt cellular structure, such that antibody binding is decreased.

Other important factors influencing Kit staining are antibody selection and immunohistochemical technique. Studies of Kit immunoreactivity in soft tissue tumors have shown highly variable results, depending on the antibody and immunostaining method used.9–11,13 Recently, Dako rabbit polyclonal Kit antibody has been used increasingly, undoubtedly influenced by its being the antibody specified for use in clinical trials of imatinib.9 Of the studies reporting relatively low Kit staining in SCLC, 1 used antibody from a noncommercial source.12 

The distinction between SCLC and LCNEC is sometimes challenging, particularly in small specimens. From initial studies demonstrating frequent Kit staining in SCLC, it was hoped that Kit would provide a useful adjunct in distinguishing SCLC from LCNEC. However, staining for Kit has been detected in a sizable proportion of LCNECs.14,18–20 Kit, therefore, should not be used for discerning SCLC from LCNEC.

In contrast to high-grade neuroendocrine carcinomas, Kit immunoreactivity is infrequent in PCT, another tumor with neuroendocrine features. In this study, no Kit staining was observed in either typical or atypical PCTs. In the few previous studies examining Kit in PCT, staining was limited.18–20,26 

Non-small cell lung cancer–SQ and NSCLC-AD had similar low rates of Kit positivity in this study.8,14,17 Although Kit staining was moderate to strong in intensity, immunoreactivity was restricted to single tumor cells or sporadic clusters. The significance of this localized pattern of staining is unclear, as Kit-positive tumor cells did not appear morphologically distinct from neighboring nonimmunoreactive cells.

Several other studies have reported higher proportions of Kit staining in NSCLC. This discrepancy likely reflects not only variations in immunohistochemical technique, but also differences in scoring systems. While some investigators are relatively conservative, others interpret weak- intensity staining and/or a low percentage of immunoreactive cells as “positive.” In the study showing the highest proportion of Kit staining in NSCLCs, cases were considered positive if more than 5% of tumor cells were immunoreactive.8 While there is no consensus regarding scoring criteria, we have found that quantifying the percentage of cells stained, as well as staining intensity, is helpful for making comparisons between studies.

Kit immunoreactivity in pleural MM was exceedingly uncommon in this study. To the best of our knowledge, only 2 previous reports have examined Kit staining in pleural MM.8,28 In one study, Kit staining was restricted to the nucleus of tumor cells in all but 1 case.28 This staining pattern has been reported previously, but only in normal and neoplastic medullary cells of the adrenal gland.14 In the other study, Kit immunoreactivity was detected in 26% of pleural MMs.8 However, staining intensity was weak in nearly all cases. Interestingly, these findings were from the same study reporting Kit staining in a high proportion of NSCLCs.8 

Localized fibrous tumor, like solitary fibrous tumor arising at other sites, shares histologic and some immunohistochemical similarities with GIST, including CD34 positivity.29 Whereas GISTs are consistently immunoreactive for Kit, variable staining has been observed in LFTs and extrapleural solitary fibrous tumors. In the present study, LFTs were the only tumors to undergo HIER. Interestingly, of the 6 LFTs in our study, 3 of the 4 tumors that had been so treated were Kit-positive. None of the LFTs in the present study exhibited features associated with aggressive behavior, such as large size, increased mitotic activity, nuclear pleomorphism, or necrosis. Others have found that while so-called malignant LFTs are uniformly positive for Kit, cases with cytologically benign features exhibit only rare staining.30 The few other studies on Kit in LFT have reported only occasional or absent staining.31,32 The mechanism underlying Kit expression in LFT is unknown. The discrepancy between the Kit staining characteristics of LFTs and extrapleural solitary fibrous tumors requires further study, particularly with regard to the role played by different antibodies and epitope-retrieval techniques.

Immunoreactivity for Kit is generally regarded as evidence of c-kit gene expression. Using immunoblot assays, a number of studies have confirmed the presence of Kit messenger RNA and protein in SCLC tumor cell lines known to stain for the Kit antibody.12,14,17,33,34 Kit protein has also been detected by Western blot analysis in cases of NSCLC-ADs immunoreactive for Kit.17 However, this has not been the experience of others.33,34 Studies using molecular techniques to evaluate Kit expression in other lung and pleural tumors have not been performed.

The success of imatinib in the treatment of metastatic GIST has stimulated significant interest in understanding the functional significance of Kit immunoexpression in other tumors. Kit expression in GIST results from the acquisition of activating mutations.35 However, such gain-of- function mutations may not be necessary for imatinib to be effective.36 Kit expression in SCLC is believed to arise from autocrine stimulation of tyrosine kinase receptor by its ligand, stem cell factor.37 Several in vitro studies have demonstrated inhibitory activity of imatinib on SCLC cell lines.38,39 Clinical trials are currently underway to assess the efficacy of this agent in SCLC patients.40 

Focal Kit immunostaining is considered by some to be of little therapeutic relevance.9 Tumors of the lung and pleura in this category include NSCLC, PCT, and MM. Nevertheless, some investigators postulate that imatinib might be effective in the treatment of MM.41 In addition to Kit receptor, imatinib inhibits the highly homologous platelet-derived growth factor receptor tyrosine kinase.42 Platelet-derived growth factor is thought to be an autocrine growth factor for MM.43 Clinical trials of imatinib for the treatment of MM are ongoing.41 It will be interesting to assess the effect of a tyrosine kinase receptor inhibitor on tumors with limited Kit staining.

As with soft tissue tumors, studies of Kit immunostaining in lung and pleural tumors have used different antibodies and a variety of epitope-retrieval methods, producing discrepant results. In our experience, Kit staining in lung and pleural tumors is observed consistently only in SCLC. Limited evidence from previous studies suggests that SCLCs positive for Kit behave more aggressively than those that are negative for Kit.23,27 Whether Kit immunostaining becomes a routine part of the pathologic evaluation of SCLC is likely to depend on the efficacy of tyrosine kinase receptor–specific agents in treating SCLC.