Difference in the hypoxic immunosuppressive microenvironment of patients with neurofibromatosis type 2 schwannomas and sporadic schwannomas.

BACKGROUND: Neurofibromatosis type 2 (NF2) patients uniformly develop multiple schwannomas. The tumor-microenvironment (TME) is associated with hypoxia and consists of immunosuppressive cells, including regulatory T cells (Tregs) and tumor-associated macrophages (TAMs). The hypoxic TME of NF2 schwannomas remains unclear. In addition, no comparative study has investigated immunosuppressive cells in NF2 and sporadic schwannomas. METHODS: In 22 NF2 and 21 sporadic schwannomas, we analyzed the immunohistochemistry for Ki-67, hypoxia-inducible factor-1α (HIF-1α), vascular endothelial growth factor receptor 1 (VEGFR1) and VEGFR2, platelet derived growth factor receptor-beta (PDGFR-ß), programmed cell death-1 (PD-1)/ programmed cell death ligand-1 (PD-L1), Foxp3, CD163, CD3, and CD8 to assess the immunosuppressive TME. RESULTS: Most vessels in sporadic schwannomas exhibited slight or negative VEGFR1 and 2 expressions with pericytes coverage. In contrast, large vessels in NF2 schwannomas exhibited strong VEGFR1 and 2 expressions without pericytes. The number of CD3+, CD8+, and CD163+ cells was significantly higher in NF2 schwannomas than in sporadic ones. The expression of PD-L1 and nestin positive cell ratio was higher in NF2 schwannomas than that in sporadic ones. The number of CD163+ cells, nestin positive cell ratio, and HIF-1α expression were significantly associated with shorter progression-free survival in NF2 schwannomas. CONCLUSIONS: This study presents the clinicopathological features of the differences in immunosuppressive cells and the expression of immune checkpoint molecules between NF2 and sporadic schwannomas. Hypoxic TME was first detected in NF2-schwannomas, which was associated with the tumor progression.

TIM-3 plays a more important role than PD-1 in the functional impairments of cytotoxic T cells of malignant Schwannomas.

Cancer immunotherapy using cytotoxic T cells demonstrates dramatic survival benefits in lymphomas, but its efficacy in solid tumors is limited. Here, we investigated the possibility of using cytotoxic T cells to treat malignant Schwannoma, a rare but aggressive nerve sheath tumor, by examining the native T-cell immunity in the host. We found that compared to CD8+ T cells from healthy controls or benign Schwannoma patients, the CD8+ T cells from malignant Schwannoma patients were present at normal frequencies but were substantially enriched with PD-1-TIM-3+ and PD-1+TIM-3+ cells. Compared to the PD-1-TIM-3- CD8+ T cells, the PD-1-TIM-3+ and PD-1+TIM-3+ CD8+ T cells presented significantly lower proliferation capacity, reduced interleukin 2 and interferon gamma expression, and/or dramatically decreased perforin and granzyme B secretion, indicating a whole-spectrum immunosuppression and reduced cytotoxicity. TIM-3 expression alone was associated with lower proliferation and less perforin and granzyme B secretion, whereas PD-1 expression alone was not associated with functional impairments, suggesting that TIM-3 expression was a better marker of exhausted CD8+ T cells. The expression of galectin 9, a TIM-3 ligand, in CD4+ Th cells was significantly elevated in malignant, but not benign, Schwannoma patients and were enriched in CD25+ Treg cells. Both the PD-1-TIM-3+ and PD-1+TIM-3+ CD8+ T cells responded to Treg-mediated and galectin 9-mediated suppression, whereas the PD-1+TIM-3- CD8+ T cells only responded to Treg-mediated suppression. In resected tumors, the malignant Schwannomas had more tumor-infiltrating CD4+ and CD8+ T cells than the benign Schwannomas, but a large fraction of these tumor-infiltrating CD4+ and CD8+ T cells expressed PD-1 and/or TIM-3, which indicated that their antitumor immunity was compromised. Together, our results suggested that PD-1 and TIM-3 blockade might be necessary in developing effective immunotherapeutic strategies in malignant Schwannoma, in which TIM-3 may play a more important role.

Schwannoma-like tumor in the anterior cranial fossa immunonegative for Leu7 but immunopositive for Schwann/2E.

Schwannoma arising from the olfactory system, often called olfactory groove schwannoma (OGS), is rare, as the olfactory bulb and tract, belonging to the central nervous system, should lack Schwann cells. Another rare entity called olfactory ensheathing cell tumor (OECT) has been reported, which mimics clinical and radiological characteristics of OGS. Here, we report two rare cases of schwannoma-like tumor in the anterior cranial fossa that showed negative staining for Leu7, but positive staining for Schwann/2E, and discuss their origin. Two cases of mass lesions in the anterior cranial fossa in a 26-year-old man and a 24-year-old woman were successfully removed. Morphological examination of these tumors was compatible with a diagnosis of schwannoma. Immunohistochemically, both cases were negative for Leu7, yielding a diagnosis of OECT, but were positive for the schwannoma-specific marker, Schwann/2E. Immunohistochemical staining results in our two cases question the current assumption that OGS and OECT can be distinguished only by Leu7 staining pattern. In conclusion, the origins of OGS and OECT remain to be determined, and further studies in larger numbers of cases are needed to characterize these rare tumors in the anterior cranial fossa.

Phase I/II Study of a Vascular Endothelial Growth Factor Receptor Vaccine in Patients With NF2-Related Schwannomatosis.

PURPOSE: The humanized antivascular endothelial growth factor (VEGF) antibody bevacizumab (Bev) is efficacious for the treatment of NF2-related schwannomatosis (NF2), previously known as neurofibromatosis type 2. This study evaluated the safety and efficacy of a VEGF receptor (VEGFR) vaccine containing VEGFR1 and VEGFR2 peptides in patients with NF2 with progressive schwannomas (jRCTs031180184). MATERIALS AND METHODS: VEGFR1 and VEGFR2 peptides were injected subcutaneously into infra-axillary and inguinal regions, once a week for 4 weeks and then once a month for 4 months. The primary end point was safety. Secondary end points included tolerability, hearing response, imaging response, and immunologic response. RESULTS: Sixteen patients with NF2 with progressive schwannomas completed treatment and were assessed. No severe vaccine-related adverse events occurred. Among the 13 patients with assessable hearing, word recognition score improved in five patients at 6 months and two at 12 months. Progression of average hearing level of pure tone was 0.168 dB/mo during the year of treatment period, whereas long-term progression was 0.364 dB/mo. Among all 16 patients, a partial response was observed in more than one schwannoma in four (including one in which Bev had not been effective), minor response in 5, and stable disease in 4. Both VEGFR1-specific and VEGFR2-specific cytotoxic T lymphocytes (CTLs) were induced in 11 patients. Two years after vaccination, a radiologic response was achieved in nine of 20 assessable schwannomas. CONCLUSION: This study demonstrated the safety and preliminary efficacy of VEGFR peptide vaccination in patients with NF2. Memory-induced CTLs after VEGFR vaccination may persistently suppress tumor progression.

Primary brain tumours and specific serum immunoglobulin E: a case-control study nested in the European Prospective Investigation into Cancer and Nutrition cohort.

BACKGROUND: Case-control studies suggest that patients with allergic diseases have a lower risk of developing glioma but not meningioma or schwannoma. However, those data can be differentially biased. Prospective studies with objective measurements of immunologic biomarkers, like immunoglobulin E (IgE), in blood obtained before cancer diagnosis could help to clarify whether an aetiological association exists. METHODS: The present case-control study nested within the European Prospective Investigation into Cancer and Nutrition (EPIC) measured specific serum IgE as a biomarker for the most common inhalant allergens in 275 glioma, 175 meningioma and 49 schwannoma cases and 963 matched controls using the ImmunoCAP specific IgE test. Subjects with an IgE level ≥0.35 kUA/l (kilo antibody units per litre) were classified as sensitized by allergens. Odds ratios (OR) and 95% confidence intervals (CI) were estimated by adjusted conditional logistic regression models for each tumour subtype. The effect of dose-response relationship was assessed in five increasing IgE level categories to estimate P-values for trend. RESULTS: The risk of glioma was inversely related to allergic sensitization (OR = 0.73; 95% CI 0.51-1.06), especially pronounced in women (OR = 0.53; 95% CI 0.30-0.95). In dose-response analyses, for high-grade glioma, the lowest OR was observed in sera with the highest IgE levels (P for trend = 0.04). No association was seen for meningioma and schwannoma. CONCLUSION: The results, based on serum samples prospectively collected in a cohort study, provide some support for the hypothesis that individuals with allergic sensitization are at reduced risk of glioma and confirm results from previous case-control studies.

Laparoscopic resection of a retroperitoneal degenerative schwannoma: a case report and review of the literature.

Degenerative schwannomas are rare benign tumors. The patient presented in this case report complained of a dull left upper quadrant pain for several months. A computed tomography scan revealed a low-density lesion at the level of T12. The lesion was laparoscopically resected and pathologic examination revealed a degenerative schwannoma.

The interaction mode of premalignant Schwann and immune effector cells during chemically induced carcinogenesis in the rat peripheral nervous system is strongly influenced by genetic background.

Contrary to rats of the highly sensitive inbred strain BDIX, BDIV rats are resistant to the induction of malignant schwannomas by N-ethyl-N-nitrosourea, arising predominantly in the trigeminal nerves. A point mutation of the neu/erbB-2 gene diagnostic of N-ethyl-N-nitrosourea-induced rat schwannomas is an early marker of Schwann precursor cells at high risk of subsequent malignant transformation. Neu-mutant cells initially arise at a similar frequency in sensitive and resistant animals. However, these cells disappear from the trigeminal nerves of resistant rats while giving rise to highly malignant schwannomas in susceptible animals. The resistance of BDIV rats obviously includes mechanisms to recognize and eliminate premalignant cells. The involvement of a cellular immune response was investigated in trigeminal nerves of both strains at different times after neonatal carcinogen exposure. An inflammatory reaction involving sequentially CD4(+) macrophages and T helper cells, CD8(+) cytotoxic T cells, and ED1(+) and ED2(+) macrophages was detected as a consequence of N-ethyl-N-nitrosourea treatment as early as postnatal day 40, briefly after the emergence of premalignant neu-mutant Schwann cells. It persisted throughout the observation period (40-250 days). However, there were no gross differences in immune cell counts between tumor-susceptible and tumor-resistant rats, except for a moderate increase of ED2(+) macrophages in N-ethyl-N-nitrosourea-treated BDIX rats only. Differential interactions of immune effector cells with premalignant Schwann cells may thus be involved in genetically determined tumor susceptibility or resistance, which could include functional differences of immune effector cells and/or a differential capability of premalignant Schwann cells to escape or counteract the cellular immune response.

Prognostic Analysis of Clinical and Immunohistochemical Factors for Patients with Spinal Schwannoma.

BACKGROUND: Schwannoma comprises approximately 25% of all spinal tumors, but there is little information published in the literature regarding this subject. Our aim in this study was to discuss diagnostic and prognostic factors for spinal schwannoma. METHODS: A retrospective study was performed to analyze the clinical and immunohistochemical data of patients with spinal schwannoma surgically treated in our center between 2005 and 2013. RESULTS: A total of 524 patients with spinal schwannoma were included in the study. The mean follow-up period was 58.3 months. Forty-eight patients developed recurrence, and 26 died. Findings from the statistical analyses suggested duration of preoperative symptoms, Sridhar classification, tumor size, bone damage, Ki67 labeling index, and S100 expression were different between benign schwannoma and the malignant subtype. Recurrence was associated with resection mode, segments of involvement, pathology grade, CD57 expression, Ki67 labeling index, and S100 expression. The overall survival was closely related with recurrence, location in sacrum, pathology grade, Ki67 labeling index, and P53 expression. CONCLUSIONS: Compared with the benign subtype, malignant schwannoma has a shorter duration of preoperative symptoms, larger tumor size, greater Sridhar classification, and poorer prognosis. Total resection can significantly reduce recurrence but not guarantee a better survival, which is associated location and pathology grade. A Ki67 labeling index >5% was not only an index for malignant subtype but also a prognostic indicator for recurrence and poor survival. Moreover, S100-negative was a prognostic indicator for recurrence, whereas P53-positive was associated with a poor prognosis.

Immune Pathobiology of Schwannomas: A Concise Review.

Schwannomas are benign tumors treatable with neurosurgery or radiosurgery, yet a small subset may exhibit aggressive growth. Hence illuminating their immune features can help develop better treatments. A tumor-promoting inflammation exists in schwannomas. Transcription factor NF-κB triggers synthesis of inflammatory cytokines and chemokines. NF-κB is suppressed by NF2/merlin, yet it is mutated or repressed in schwannomas, and therefore MCP-1/CCL2, MIP-1α/CCL3, CXCL16, and CXCR6/Bonzo are likely expressed in these tumors. CD68+ and CD163+ macrophages may infiltrate schwannomas and promote their growth. Anti-inflammatory salicylates inhibit schwannomas in cell culture and clinically. Schwannomas that cannot be completely removed by neurosurgery or controlled by radiosurgery may be suitable targets of pharmacologic interventions focusing on immune mechanisms.

Immunoreactivity of bovine schwannomas.

Thirty schwannomas from 22 cows were examined immunohistochemically. All were positive for vimentin and Ki-67 but negative for pancytokeratin, neurofilament, and desmin. S-100 immunolabelling varied between and within lesions. The numbers of tumours giving positive results for S-100, neuron-specific enolase (NSE) and glial fibrillary acidic protein (GFAP) were 16, 30 and 25, respectively. It was concluded that vimentin-positive tumours suspected to be schwannomas should also be immunolabelled for NSE and GFAP to confirm the diagnosis.

Effect of levamisole on malignant experimental neurinoma grown subcutaneously in a young rat: brief communication.

Levamisole, a drug that stimulates the immunologic defenses of the host, was tested on the experimental malignant neurinoma grown subcutaneously in young female inbred CDF rats in 3-month trial. The drug resulted in prevention of tumor growth in 80% of the treated animals, whereas tumors grew in 100% of the controls. After discontinuation of the drug, none of the animals in the treated group showed evidence of tumor growth upon clinical and pathologic examination 1 month after drug treatment was terminated. Treated animals failed to show any side effects related to levamisole upon clinical and pathologic examination.

Sequential adsorption analysis of antibodies to neuroectodermal cells by an indirect quantitative method.

An indirect radioimmunoassay was developed for the sequential adsorption analysis of rabbit antibodies raised against rat neuroectodermal cells. The quantitativee relationship between primary adsorbed rabbit antitumor antibodies and secondary adsorbed goat antibodies to rabbit IgG was explored by paired radioiodine analysis. In the final indirect method, the amount of unlabeled rabbit antibody removed by cultured monolayers of cells at each step in a sequence of cells could be determined from an equation relating the unlabeled amount to values of bound 125I-labeled goat antirabbit IgG. To obtain the total quantity of rabbit antibody in a particular antiserum reagent, a sequential adsorption analysis was done with successive steps of homologous cells. To obtain the amount of antibody that was cross-reactive for other cell lines, we included those lines in the first several steps of the sequence. The sequential adsorption profile was considered as a more important indicator of the quality of a particular antiserum reagent than was the total amount of antibody contained in it. Neuroectodermal cell lines used as illustrative examples included subclones of the C6 astrocytoma and of the RN-2 schwannoma.

Characterizing the immune microenvironment of malignant peripheral nerve sheath tumor by PD-L1 expression and presence of CD8+ tumor infiltrating lymphocytes.

BACKGROUND: Malignant peripheral nerve sheath tumor (MPNST) is an aggressive sarcoma with few treatment options. Tumor immune state has not been characterized in MPNST, and is important in determining response to immune checkpoint blockade. Our aim was to evaluate the expression of programmed death-ligand 1 (PD-L1), programmed cell death protein 1 (PD-1), and presence of CD8+ tumor infiltrating lymphocytes (TILs) in MPNST, and correlate these findings with clinical behavior and outcome. RESULTS: PD-L1 staining of at least 1% was seen in 0/20 nerves, 2/68 benign lesions and 9/53 MPNST. Two of 68 benign lesions and 7/53 (13%) MPNST had at least 5% PD-L1 staining. CD8 staining of at least 5% was seen in 1/20 (5%) nerves, 45/68 (66%) benign lesions and 30/53 (57%) MPNST. PD-L1 was statistically more prevalent in MPNST than both nerves and benign lesions (p=0.049 and p=0.008, respectively). Expression of PD-1 was absent in all tissue specimens. There was no correlation of PD-L1 or CD8 expression with disease state (primary versus metastatic) or patient survival. METHODS: A comprehensive PNST tissue microarray was created from 141 surgical specimens including primary, recurrent, and metastatic MPNST (n=53), neurofibromas (n=57), schwannoma (n=11), and normal nerve (n=20). Cores were stained in triplicate for PD-L1, PD-1, and CD8, and expression compared between tumor types. These data were then examined for survival correlates in 35 patients with primary MPNST. CONCLUSIONS: MPNST is characterized by low PD-L1 and absent PD-1 expression with significant CD8+ TIL presence. MPNST immune microenvironment does not correlate with patient outcome.

GFA protein reactivity in nerve sheath tumors: a polyvalent and monoclonal antibody study.

We studied glial fibrillary acidic (GFA) protein immunoreactivity in 30 schwannomas, including two intracerebral examples, 26 neurofibromas and 12 neuromas using the immunoperoxidase method with a polyvalent antiserum (PVAS) and three well-characterized monoclonal antibody (MAb) preparations. Twelve of the schwannomas, including both intracerebral tumors, two of the neurofibromas and none of the neuromas immunostained with PVAS. Except for one schwannoma, all the PVAS-positive tumors were positive with two of the MAb preparations. While both of the intracerebral schwannomas were positive with the third MAb, none of the extracerebral tumors were. Our results suggest that: 1) human nerve sheath tumors contain cells having polypeptides that share epitopes with GFA protein, but 2) these polypeptides differ from astrocytic GFA protein by at least one epitope, and 3) the location of the tumors in relation to the central nervous system may influence GFA protein immunoreactivity.

Immunity to transplantable nitrosourea-induced neurogenic tumors. I. Potentiation of tumor immunity with Corynebacterium parvum.

Various injection schedules of C. parvum and tumor cells of the nitrosourea-induced malignant neurinoma, TR-481, were used to induce tumor immunity in syngeneic CDF rats. Although subcutaneous injection of the poorly immunogenic TR-481 cells alone or with C. parvum caused retardation of growth of 2 x 10(5) TR-481 cells injected 1-3 weeks later, no significant difference in tumor size or incidence was obtained, as judged by tumor growth at 8 weeks. In contrast, injection of TR-481 with C. parvum into C. parvum-presensitized rats caused a more significant degree of tumor immunity with complete inhibition of the challenge tumor growth in 17-33% of the animals. Repeated subcutaneous injection of gamma-irradiated TR-481 tumor cells mixed with C. parvum also proved effective, resulting in absence of tumor growth in 40% of the rats. Tumor immunity was specific, since growth of an unrelated tumor was unaffected. It is suggested that local immunological reactivity to C. parvum in the immunizing tumor promotes development of specific tumor immunity.

Mitogenic lectin receptors of nervous system tumors. Study of gliomas, neural crest tumors, and meningiomas in vitro using phytohemagglutinin and concanavalin A.

The effects of mitogenic lectins Phytohemagglutinin (PHA), and Concanavalin A (Con A) on the growth rate of cells derived from glial tumors (astrocytoma, ependymoma, glioblastoma, medulloblastoma, and C6 rat glioma), neural crest tumors (neuroblastoma and schwannoma), and meningiomas were studied. The cell lines were of human and animal origin. The specificity of lectin binding to mitogenic receptors was evaluated using complementary monosaccharides. In all glial- and some neural-crest tumor-derived cell lines, there was a lectin concentration-dependent and cell density-dependent, biphasic growth rate response with stimulation at low and inhibition at high lectin concentrations. This response did not depend on the type of glial tumor, species of origin, or passage level in vitro. Although, in meningioma-derived cell lines, lectins did not induce a growth rate response, they caused morphological changes ("whorling"). Lectin stimulation in glial tumor-derived cell lines resembles that occurring in peripheral blood lymphocytes. Lectin-induced mitogenesis may lay the groundwork for the establishment of a model of glial cell proliferation, and that permits the evaluation of cell surface effects, intracellular mechanisms, and epigenetic factors in studies of tumors, neural development, and neuroimmunology.

Immunity to transplantable nitrosourea-induced neurogenic tumors. II. Immunoprophylaxis of tumors of the brain.

An effective method was sought to immunize rats against the growth of intracerebrally (IC) injected T9 tumor, a gliosarcoma cell line. Rats which were immunized with either 10(6) T9 cells mixed with 0.14 mg C. parvum, or 10(7) irradiated T9 cells showed tumor immunity to intradermal (ID) transplantation. However, to obtain tumor immunity to an IC challenge of T9 cells, rats initially had to reject an ID challenge of T9 cells. After sequential rejections of ID challenges of as many as 10(7) cells, a high degree of immunity was obtained, allowing rejection of up to 5 X 10(6) T9 cells injected IC. Spleen cells from highly immunized rats mixed with T9 cells at a ratio of 1:25 (tumor:spleen) destroyed T9 tumor cells in a Winn test. Normal spleen cells had no effect on tumor growth. We conclude that: 1) T9 cells are moderately immunogenic, 2) an effective method of immunization of CDF rats against ID transplanted T9 cells is 10(6) T9 cells with 0.14 mg C. parvum or 10(7) irradiated T9 cells, 3) a higher degree of tumor immunity is necessary to reach the brain than is needed to reach the periphery, and 4) spleen cells of highly immunized rats are cytotoxic to T9 tumor cells.

Malignant peripheral nerve sheath tumor (MPNST) showing perineurial cell differentiation.

We report a malignant peripheral nerve sheath tumor (MPNST) showing the unusual immunohistochemical feature of epithelial membrane antigen (EMA) immunoreactivity in a 52-year-old man. The tumor, located in the left paravertebral muscles, was composed of both cellular and myxoid areas. Spindle-shaped tumor cells were arranged in longitudinal cell cords, in a storiform pattern, and in whorled structures. Ultrastructurally, the tumor cells had many interdigitating cell processes, primitive cell junctions, and discontinuous external laminae. These histological and ultrastructural features were most consistent with those of an MPNST; but on immunohistochemistry, the tumor cells reacted for epithelial membrane antigen rather than for S-100 protein or Leu-7. Because EMA-immunoreactivity was recently demonstrated in perineurial cells, we concluded that the tumor was an MPNST mainly composed of tumor cells showing perineurial cell differentiation.

Immunohistochemical detection of epithelial membrane antigen in normal perineurial cells and perineurioma.

The use of epithelial membrane antigen (EMA) as an immunohistochemical marker for normal and neoplastic perineurial cells is described. Normal perineurial cells react strongly for this antigen, which is also expressed by the cells of perineurioma. Instead, neurofibromas and schwannomas only show some peripheral or entrapped layers of EMA-positive cells. In traumatic and Morton's neuromas, bundles of neural fibers are wrapped in layers of EMA-positive perineurial cells. Neurothekeoma and granular cell tumor show no EMA reactivity. The detection of an epithelial marker in perineurial cells is in agreement with the concept of a "perineural epithelium" and seems to support a common embryologic origin for the perineurial cell and the equally EMA-positive arachnoidal cap cell. The availability of an immunohistochemical marker for the perineurial cell provides an easy and convenient tool for the evaluation of the participation of this cell in a variety of pathologic processes.

Gastrointestinal stromal tumors--value of CD34 antigen in their identification and separation from true leiomyomas and schwannomas.

The term "gastrointestinal stromal tumor" (GIST) has been applied to mesenchymal tumors that represent neither typical leiomyomas nor schwannomas. In this study we analyzed immunohistochemically 67 histologically benign [< 2 mitoses/10 high-power field (HPF)], six borderline (3-5 mitoses/10 HPF), and 23 malignant GIST (> 5 mitoses/10 HPF) and compared them with 10 typical leiomyomas and 5 schwannomas of the gastrointestinal tract. The benign GISTs with spindle cell pattern (67 cases) were typically negative for muscle cell markers (only 3% positive for desmin and 25% for alpha-smooth muscle actin) and S100 protein, but 70% of the cases were positive for CD34, the myeloid progenitor cell antigen also present in endothelial cells and some fibroblasts. However, none of the cases was positive for CD31 (PECAM-1), a more endothelial cell-specific antigen. The absence of CD31 in GIST separates it from Kaposi's sarcoma, a tumor known to be positive for both CD34 and CD31. Fourteen cases of benign GIST of epithelioid cell type showed an immunophenotypic profile similar to the spindle cell tumors. The small intestinal tumors were more commonly actin positive and less commonly CD34 positive than were the gastric tumors. The malignant spindle and epithelioid GIST showed features essentially similar to those in corresponding benign tumors. In contrast, all typical leiomyomas were positive for muscle cell markers and were negative for CD34 and S100 protein. Gastrointestinal schwannomas were S100-protein positive, and negative for muscle markers and CD34. Our results show that gastrointestinal mesenchymal tumors can be immunophenotypically divided in categories that correlate with light microscopically defined diagnostic entities, namely typical leiomyomas, schwannomas, and GIST, most cases of the latter representing tumors of primitive mesenchymal cells that are CD34 positive.