Abstract
Aim: To characterize the growth pattern and antigen profile of peripheral nerve sheaths tumors (PNST) in a large series of tumors obtained from patients with neurofibromatosis type 1 (NF1). Materials and Methods: Tissue micro-array technique was applied to study 520 PNSTs of 385 patients with NF1 by immunohistochemistry for human epidermal growth factor receptors erb-b2 receptor tyrosine kinase 2 (ERBB2) and ERBB3, CD44, neuroregulin (NRG1) and proliferation index by Ki-67. PNSTs were classified as cutaneous neurofibroma (CNF) in 114 cases, diffuse neurofibroma (DNF) in 109, diffuse plexiform neurofibroma (DPNF) in 108, plexiform neurofibroma (PNF) in 110, and malignant PNST (MPNST) in 22. Results: The Ki-67 proliferation index was significantly higher in MPNST than in benign PNST (p<0.001). ERBB2 expression was significantly lower in PNST with diffuse growth than in PNF and MPNST (p<0.001). ERBB3 expression was also higher in PNF and MPNST (both p<0.001) than in diffuse PNST. NRG1 expression was significantly higher in PNF than in non-encapsulated benign PNST or MPNST (both p<0.001). Co-expression of ERBB2, ERBB3 and ligand NRG1 was rare, mainly observed in PNST with a plexiform component (in four PNFs, nine DPNFs, one CNF, and two MPNSTs). Expression of CD44 in contrast was significantly stronger in diffusely growing PNST than in PNF (p<0.001). Conclusion: Growth factor receptors ERBB2 and ERBB3 were significantly up-regulated in PNF and MPNST. The antigen expression pattern of DPNF resembled that of benign PNST with diffuse growth pattern rather than that of encapsulated PNF. Differentiating PNST may be important for the assessment of neurofibroma progression, and for the expected impact of drugs currently used for tumor reduction.
- Neurofibromatosis type 1
- neurofibroma
- plexiform neurofibroma
- diffuse plexiform neurofibroma
- neurofibroma classification
- malignant peripheral nerve sheath tumor
- ERBB2
- ERBB3
- NRG1
- Ki-67
Neurofibromatosis type 1 (NF1) is an autosomal dominant inherited tumor predisposition syndrome (1). The NF1 gene codes for neurofibromin, a tumor suppressor which is part of the rat sarcoma homolog (RAS) pathway. Loss of neurofibromin affects growth factor signaling, promoting the development of neoplasm (2-5). Peripheral nerve sheath tumors (PNSTs) are the hallmark of this multifaceted disorder (6-8). PNSTs in NF1 are further defined according to morphological and topographic criteria. The distinction between cutaneous neurofibroma (CNF) and plexiform neurofibroma (PNF) is an essential aid in NF1 diagnostics (6-8). The former often arise in large numbers in NF1, being diagnostically and esthetically relevant in the post-pubertal phase of life. The latter, probably already existing at birth, more frequently cause functional limitations. PNF is regarded as biological precursor of malignant peripheral nerve sheath tumors (MPNSTs) (6-8). MPNST represents a very important factor for the reduced life expectancy of patients with NF1 (9-11). PNSTs in NF1 arise from Schwann cells or Schwann cell precursors (3, 6-8, 10, 12). The morphological differentiation of NF1-associated PNST identifies additional subtypes of little-known functional and biological significance (3, 6-8, 10). Diffuse plexiform PNF (DPNF) can be distinguished from PNF by diffuse tumor growth in addition to encapsulated tumor regions (13). The distinction of plexiform neurofibroma subtypes may also be of therapeutic significance because DPNF may show defects of the capsular boundaries in its plexiform parts which facilitates exchange of substances, for example, drugs. In contrast, disruption of the capsular boundary is not observed in intraneural PNF.
The study follows on from an earlier report on immunohistochemical studies of NF1-associated PNST (14) and presents further characteristics concerning growth factor receptor expression in NF1-associated neurofibromas, with a focus on distinguishing DPNF within this group.
Materials and Methods
Patient cohort and samples. Routinely formalin-fixed and paraffin-embedded human tissue samples were processed and diagnosed at the Institute of Neuropathology, University Medical Center Hamburg-Eppendorf. The tissue samples originated from surgery for PNST of patients with NF1 performed in the Department of Oral and Craniomaxillofacial Surgery, University Medical Center Hamburg-Eppendorf, between 1981 and 2012. All patients met the current diagnostic criteria for NF1 (14). The present cohort comprised 520 PNSTs from 385 patients. There were 228 samples from male patients and 283 from female patients. Data on sex were missing in nine cases. PNSTs were referenced into five diagnostic groups [cutaneous neurofibroma (CNF), diffuse neurofibroma (DNF), diffuse-plexiform neurofibroma (DPNF), plexiform neurofibroma (PNF), malignant peripheral nerve sheath tumor (MPNST)].
The samples comprised 136 CNFs, 123 DNFs, 113 DPNFs, 126 PNFs, and 22 MPNSTs. Atypical neurofibroma and other rare PNST entities were excluded from evaluation. Classification of tumors was performed prior to further immunohistochemical study by an experienced neuropathologist in neurofibromatosis diagnostics (CH) on eight standardized sections of each tumor sample staining with hematoxylin-eosin, periodic acid Schiff reaction, Elastica-van Gieson and Turnbull blue. In addition, immunohistochemistry was performed for S100 protein, epithelial membrane antigen, neurofilament, and Ki-67 protein. For the present study, a section was taken from each paraffin block and stained with hematoxylin and eosin. The tumors were re-classified according to the neurofibroma type (14) and typical areas were marked on the slides. The marked regions were then punched out from the block for construction of a tissue microarray (TMA). From each tumor, two punch samples were taken. A TMA consisted of up to 148 tissue punches. For immunohistochemical staining, TMA sections were processed in a Ventana Benchmark XT staining automat (Roche Deutschland, Grenzach-Wyhlen, Germany). Only samples that could be assessed without doubt were included in the evaluation. The classification of NF1-associated PNST according to morphological criteria has been described in detail elsewhere (14).
Results of CNF, DNF, and DPNF staining were combined into one group as ‘benign neurofibromas’ for several analyses because of their similar biological behavior. Benign neurofibromas were compared with PNF as a known precursor of MPNST in NF1 and with NF1-associated MPNST. The data on neurofilament expression used in this study has been described in detail elsewhere (14).
Immunohistochemistry. Technical equipment and performance of immunohistochemical examinations have been described in detail elsewhere (14). Details of the antibodies used in the present study are summarized in Table I. Evaluation of antibody binding was performed qualitatively, semi-quantitatively, or quantitatively according to the targeted structure (15).
Antibodies used in the study, in alphabetic order.
Quantitative evaluation of staining results. Proliferation and cell density were determined in the region of highest density of Ki-67-positive nuclei. All positive and negative nuclei were counted in a field of view at 40× magnification (0.12 mm2). The proliferation index (Ki-67 index) was determined as the percentage of Ki-67-stained nuclei in relation to all nuclei examined.
For the following four markers, the entire area of the TMA spot (0.79 mm2) was evaluated.
CD44 staining was graded according to cell positivity as: Negative, 0% positive tumor cells (score 0); weakly positive, 0-20% positive spindle-shaped cells (score 1); moderately positive, >20% to <100% positive cells (score 2); strongly positive, 100% positive cells (score 3).
Human epidermal growth factor receptor erb-b2 receptor tyrosine kinase 3 (ERBB3, HER3) staining was graded as follows: Negative, 0% positive nuclei (score 0); weak, 0-80% positive nuclei (score 1); moderate, >80% to <100% positive nuclei (score 2); strong, 100% positive nuclei (score 3).
Neuregulin-1 (NRG1) staining was graded as follows: No staining (score 0), weak staining of 100% of cells (score 1), weak to moderate staining of 33-66% of cells (score 2), moderate to strong staining of >66% to <100% of cells (score 3), and strong staining of 100% of cells (score 4).
Qualitative evaluation of staining results. ERBB2 (HER2) staining was graded as follows: None (score 0), perineural or perivascular (score 1), single tumor cells or reticular tumor staining (score 2).
Statistical analysis. The statistical evaluation was carried out with the IBM SPSS Statistics 21 program (IBM, Ehningen, Germany).
To describe the distribution of values, median and quartiles were calculated for metric variables, except for age, since the assumption of a normal distribution was not sufficiently fulfilled. These values are given in the results as follows: Median (25% percentile/75% percentile). For age at sampling, the assumption of a normal distribution was sufficiently fulfilled, therefore the mean value is given for this variable.
The percentages of tumors for each expression score relative to the total number of tumors evaluated and tumor subtypes were calculated for nominally and ordinally scaled variables.
The statistical hypothesis “the results of the five tumor subgroups follow the same distribution on average” was tested for nominal- and ordinal-scale variables by the chi-squared test as standard. For metric variables, nonparametric tests (Kruskal–Wallis and Mann– Whitney U-tests) were used for evaluation by default because the assumption of normal distribution was not adequately met. For mean age at sampling, the Welch test was used instead of the t-test because a normal distribution but not equality of variance was assumed. The significance level was set at 5% for all analyses.
To analyze differences between tumor groups, cross-tabulations were considered, each containing only two groups to be compared. Groups were considered different when the p-value of the chi-squared test was less than 0.05.
The Kruskal–Wallis test considered the null hypothesis that the distribution of more than two groups would not differ in their means. If the null hypothesis was rejected, entities not differing in their means were combined into subgroups and pairwise differences were calculated with the Mann–Whitney U-test (CNF/DNF/DPNF vs. PNF or MPNST).
All statistical computations were performed as exploratory analyses.
Ethics and funding. This study was performed according to the rules of local authority of Eppendorf University Hospital (appointed by the University of Hamburg) as a prerequisite for the preparation of a medical dissertation (LKNN). All patients gave informed consent regarding the scientific evaluation of tumor data prior to their treatment in the hospital. All procedures performed in this study involving human participants were in accordance with ethical standards of the institutional and/or national research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethics standards. Data were anonymized prior to analysis, and the investigators studying the tissue samples were blinded for diagnosis, identity of individuals, and assignment of the single case to a diagnostic group. These investigations were carried out in accordance with the Hamburg Health Service Act (Hamburgisches Gesundheitsdienstgesetz). This type of study does not require an ethics approval.
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Results
Sex. Specimens (n=520) were examined from 385 patients. There were 228 samples from male patients and 283 from female patients. Data on sex were missing in nine cases.
Tumor subtypes and age at sample collection. The mean age of patients overall was 30.9 years. Considering tumor types separately, the CNFs (n=136) were excised at a mean patient age of 39.4 years, DNFs (n=123) at 32.0 years, DPNFs (n=113) at 25.6 years, PNFs (n=125) at 24.6 years, and MPNSTs (n=22) at 38.0 years. The Welch test revealed a significantly younger age at resection for patients with PNF than for those with non-encapsulated benign neurofibromas (CNF/DNF/DPNF) (p<0.001).
Tumor subtypes and localization. Tumor localization was known for 463 samples and was classified as head/neck in 33.0%, trunk in 40.5%, and extremities in 26.4%. Comparing tumor subtypes by location, CNFs (n=114) originated from the head/neck region in 19.3%, trunk in 57.0%, and extremities in 23.7%; DNFs (n=109) originated from the head/neck region in 41.3%, trunk in 22.9%, and extremities in 35.8%; DPNFs (n=108) originated from the head/neck region in 32.4%, trunk in 46.3%, and extremities in 21.3%; PNFs (n=110) originated from the head/neck region in 43.6%, trunk in 32.7%, and extremities in 23.6%; MPNSTs (n=22) were not observed in the head/neck region (0%), but mainly in the trunk (61.5%), and in the extremities (38.5%). The difference in location between MPNST and benign tumors was statistically significant (chi-squared, p<0.001).
Cellularity. Cellularity was determined in neurofibromas and MPNSTs (n=431). Median cellularity of all PNST was 166.0/0.12 mm2 (124.0/224.0). Considering tumor subtypes separately, Mann-Whitney U-test showed a significantly higher cellularity for benign tumors CNF/DNF/DPNF/PNF compared to MPNST (p<0.001, Figure 1).
Cellularity (cells counted in one visual field of 0.12 mm2) in benign human neurofibromas [cutaneous neurofibroma (CNF), diffuse neurofibroma (DNF), diffuse-plexiform neurofibroma (DPNF), plexiform neurofibroma (PNF)] and malignant peripheral nerve sheath tumor (MPNST). Benign neurofibromas demonstrated a significantly lower cellularity than MPNST. ***Significantly different at p<0.001 (Mann-Whitney U-test. Boxes represent 50% of the values, the line in the box is the median value. The upper and lower borders of the box represent the 3rd and 1st quartiles, respectively. Whiskers indicate the largest and smallest values (excluding outliers).
Ki-67 index. Anti-Ki-67 identified nuclei in proliferating cells (Figure 2). In all cases, immunoreactive (positive) nuclei were determined in one visual field of 0.12 mm2. The median Ki-67 index for PNSTs (n=426) was 1.3% (0.7%/2.5%). No results were available for 94 cases (18.1%).
Proliferation marker Ki-67 in neurofibromatosis type 1-associated peripheral nerve sheath tumors. A: Diffuse neurofibroma, low proliferation. B: Malignant peripheral nerve sheath tumor, high proliferation; Ki-67 immunohistochemistry with diaminobenzidine as chromogen, counterstained with hemalum. Bar: 20 μm for A and B.
Considering the tumor subgroups separately, Figure 3 shows the values as a boxplot. Mann-Whitney U-test revealed a significantly lower proliferative activity in benign tumors (CNF/DNF/DPNF) including PNF compared to MPNST (p<0.001).
Ki-67 index in benign human neurofibromas [cutaneous neurofibroma (CNF), diffuse neurofibroma (DNF), diffuse-plexiform neurofibroma (DPNF), plexiform neurofibroma (PNF)] and malignant peripheral nerve sheath tumor (MPNST). Benign neurofibromas had a significantly lower proliferation index than MPNSTs (cells counted in one visual field of 0.12 mm2, n=426). Boxes represent 50% of the values, the line in the box is the median value. The upper and lower borders of the box represent the 3rd and 1st quartiles, respectively. Whiskers indicate the largest and smallest values (excluding outliers). ***Significantly different at p<0.001 (Mann-Whitney U-test).
ERBB2. The majority of neurofibromas and MPNSTs (n=438) were negative for ERBB2 (89.0%); in 5.7%, perineurial and perivascular structures stained, and in 5.3%, single cells in tumor tissue were stained (Figure 4). ERBB2 also stained endothelial cells and glands in the tumors. No results were available for 82 cases (15.8%).
Erb-b2 receptor tyrosine kinase 2 (ERBB2) immunohistochemistry in human neurofibromas. A: Plexiform-diffuse neurofibroma, negative. B: Diffuse neurofibroma, perivascular staining. C: Plexiform neurofibroma, perineurial staining. D: Malignant peripheral nerve sheath tumor, staining of single cells. ERBB2 immunohistochemistry with diaminobenzidine as chromogen, counterstained with hemalum. Bars: 20 μm.
ERBB2 staining by tumor subtype is depicted in Figure 5. Comparing tumor subtypes regarding ERBB2 staining, chi-squared test showed a significantly lower expression of ERBB2 in the tumor tissue of benign non-encapsulated tumors (CNF/DNF/DPNF) than in PNFs (p<0.001) and MPNSTs (p<0.001), with a moderate labelling in the tumor tissue in 2/82 CNFs, 1/102 DNFs and 1/108 DPNFs compared to 16/124 PNFs and 3/22 MPNSTs.
Erb-b2 receptor tyrosine kinase 2 (ERBB2) expression in neurofibromatosis type 1-associated peripheral nerve sheath tumors [cutaneous neurofibroma (CNF), diffuse neurofibroma (DNF), diffuse-plexiform neurofibroma (DPNF), plexiform neurofibroma (PNF)] and malignant peripheral nerve sheath tumor (MPNST). ERBB2 expression in tumor cells, was significantly lower in benign non-encapsulated neurofibromas (CNF, DNF, DPNF) than in PNF or MPNST (n=438, p<0.001 for both comparisons, chi-squared test).
ERBB3. ERBB3 stained the cell membrane and cytoplasm in human neurofibromas with expression ranging from weak to strong (Figure 6). ERBB3 also stained glands, endothelial, and epithelial cells.
Erb-b2 receptor tyrosine kinase 3 in neurofibromatosis type 1-associated peripheral nerve sheath tumors. A: Cutaneous neurofibroma, weak staining, patient 313/1; inset depicts positive cytoplasm of tumor cell. B: Plexiform neurofibroma, moderate staining; inset depicts positive cells. C: Malignant peripheral nerve sheath tumor, strong staining; inset shows positive cells. Bar: 20 μm, applies to all main images.
All neurofibromas examined were positive for ERBB3 (n=222, 100%). However, expression differed between diffuse benign tumors and PNF as well as MPNST (Figure 7). Chi-squared test demonstrated that expression of ERBB3 was significantly lower in benign non-encapsulated neurofibromas (CNF/DNF/DPNF) than in PNF and MPNST (p<0.001 for both comparisons, median labelling intensity for CNF/DNF/DPNF=2, median for PNF and MPNST=3).
Erb-b2 receptor tyrosine kinase 3 expression in neurofibromatosis type 1-associated peripheral nerve sheath tumors [cutaneous neurofibroma (CNF), diffuse neurofibroma (DNF), diffuse-plexiform neurofibroma (DPNF), plexiform neurofibroma (PNF)] and malignant peripheral nerve sheath tumor (MPNST). Benign non-encapsulated neurofibromas (CNF, DNF, DPNF) showed significantly lower expression of ERBB3 than PNF or MPNST (n=222). ***Significantly different at p<0.001 (chi-squared test).
NRG1. Neurofibroma cells demonstrated a variable expression of NRG1 in the cell membrane (Figure 8). The majority of neurofibromas and MPNSTs (n=411) were NRG1-positive (87.1%, negative: 12.9%). Semi-quantitative evaluation yielded 50.5% weakly stained tumors, 14.1% moderately stained, and 22.4% strongly stained tumors overall. No results were available for 109 cases (21.0%).
Neuregulin-1 expression in human neurofibromas. A: Cutaneous neurofibroma, negative. B: Cutaneous neurofibroma, weak. C: Malignant peripheral nerve sheath tumor, moderate. D: Plexiform neurofibroma, strong. E: Plexiform neurofibroma, strong staining in plexiform parts. Bar in A-C: 20 μm; D: 50 μm; E: 200 μm.
Comparing tumor subtypes (Figure 9), PNFs (n=117) were the only group to show strong staining. In all other subtypes, weak expression predominated [median labelling for non-encapsulated neurofibromas (CNF/DNF/DPNF) and MPNST, respectively=1 vs. median in PNF=3; PNF vs. (CNF/DNF/DPNF), p<0.001; PNFs vs. MPNSTs, p<0.001; chi-squared test for both comparisons].
Neuregulin-1 (NRG1) expression in neurofibromatosis type 1-associated peripheral nerve sheath tumors [cutaneous neurofibroma (CNF), diffuse neurofibroma (DNF), diffuse-plexiform neurofibroma (DPNF), plexiform neurofibroma (PNF)] and malignant peripheral nerve sheath tumor (MPNST). Benign non-encapsulated neurofibromas (CNF, DNF, DPNF) and MPNST showed significantly lower expression of NRG1 than PNF (n=411). ***Significantly different at p<0.001 (chi-squared test).
Relating the numbers of axons previously determined (14) to NRG1 immunohistochemistry revealed that tumors with high numbers of axons expressed significantly more NRG1 than tumors with low numbers of residual axons (p<0.001, Mann-Whitney U-test, Figure 10).
Number of axons according to neuregulin-1 expression in neurofibromatosis type 1-associated peripheral nerve sheath tumors (PNST). The number of axons was determined per tumor microarray spot comprising an area of 0.79 mm2 (n=182). Boxes represent 50% of the values, the line in the box is the median value. The upper and lower borders of the box represent the 3rd and 1st quartiles, respectively. Whiskers indicate the largest and smallest values (excluding outliers). ***Significantly different at p<0.001 (Mann-Whitney U-test).
CD44. CD44 stained the cell membrane in human neurofibroma cells (Figure 11). Regarding all samples investigated for CD44 expression (n=487), 24.4% of benign neurofibromas and MPNSTs were negative for CD44, 45.8% expressed the antigen weakly, 27.9% moderately, and 1.8% strongly.
CD44 expression in peripheral nerve sheath tumors. A: Plexiform neurofibroma, negative. B: Plexiform-diffuse neurofibroma, weak. C: Diffuse neurofibroma, medium. D: Plexiform-diffuse neurofibroma, strong; the inset shows positive cytoplasm. Bar: 20 μm (all main images).
Considering the tumor subtypes separately (Figure 12), chi-squared test revealed significantly lower expression in PNF than in diffusely growing tumor subtypes (benign lesions and MPNST vs. PNF p<0.001, median for DNF, DPNF, CNF and MPNST=1, Median for PNF=0; Figure 13). CD44 also stained inflammatory cells (mast cells, monocytes, and macrophages), perivascular cells and glands in tumor.
CD44 expression in neurofibromatosis type 1-associated peripheral nerve sheath tumors by tumor subtype. CNF: Cutaneous neurofibroma. DNF: Diffuse neurofibroma. DPNF: Diffuse -plexiform neurofibroma. PNF: Plexiform neurofibroma. MPNST: Malignant peripheral nerve sheath tumor. Chi-squared test revealed significantly lower expression in PNF than in diffusely growing tumor subtypes (benign lesions and MPNST vs. PNF p<0.001).
CD44 expression in diffusely growing peripheral nerve sheath tumors (PNST) vs. plexiform neurofibroma (PNF). PNST with diffuse growth pattern (cutaneous neurofibroma, diffuse neurofibroma, diffuse-plexiform neurofibroma, malignant peripheral nerve sheath tumor) showed significantly higher expression of CD44 than PNF (n=487). ***Significantly different at p<0.001 (chi-squared test).
Discussion
In this study, PNST was differentiated into five subgroups: CNF, PNF, DPNF, DNF, and MPNST. This differentiation of PNST into tumor subtypes is widely accepted (6, 7). However, these subtypes naturally also exhibit various similarities. In the present study, similar antigen expression patterns made grouping for statistical analysis meaningful, in particular of CNF, DNF and DPNF on the one hand, and PNF or MPNST on the other. The protein expression pattern of tumors shows that DPNF has more in common with CNF and DNF than with PNF. In the following paragraphs the different variables investigated are discussed.
Cellularity. High cellularity is often a feature of malignant tumors. MPNSTs are described as highly cellular tumors (6-8, 10, 11). Accordingly, MPNSTs in the present study showed a significantly higher cellularity (median=482.0) than all benign tumor subtypes. Within the benign lesions, CNFs demonstrated the highest cell density (median=242.3) which was about half of that in MPNSTs. PNF showed the lowest cellularity of about one fifth of that in MPNST (median=104.2).
Ki-67. The study confirms the expected low proliferation rate of tumor cells in benign PNST (16-21). Significant differences in the Ki-67 index were found only between MPNST and benign PNST. The determination of the proliferation index is an important and reliable step for the morphological differentiation of PNST (18).
The Ki-67 index is correlated with tumor malignancy in human PNSTs (20). It is low in benign subtypes and high in MPNSTs (20). Various studies report the mean proliferation index to be 10-65% for MPNSTs and 1-5% for benign neurofibromas (16, 18, 20). In the present analysis, the findings were similar. For CNFs, DNFs, DPNFs, and PNFs, the mean Ki-67 indices were 1.1-1.5%, and for MPNST, it was significantly higher at 16.9%.
Growth factors and receptors: ERBB2, ERBB3, NRG1. Epidermal growth factors and their receptors (EGFR) play an essential role in peripheral nerve development, maintenance, repair, chronic diseases, and PNST (22-36).
NRG1 binds to the ERBB3, which subsequently forms a heterodimer with ERBB2 or ERBB4 (30). There are numerous interactions between NRG1 and ERBB in normal peripheral nerve development and disorders (37-45). Stonecypher et al. detected ERBB2, ERBB3, and NRG1 in human neurofibromas and MPNST. Inhibition of growth factor receptors in cell culture resulted in a reduction of ERBB phosphorylation and DNA synthesis in MPNST cells (30). Further studies highlight the important role of this signaling pathway in tumor growth and migration of neurofibroma and MPNST cells in vitro and in vivo (1, 45). In the present work, the expression of ERBB2, ERBB3, and NRG1 was tested in surgical tumor samples. Most tumors (89%) examined in this study were negative for ERBB2. In ERBB2-positive tumors, the cytoplasm of a small fraction of tumor cells was stained. Immunostaining of the cytoplasm rather than the cell membrane was noted because the antibody used (c-erbB2 oncoprotein, A0485; Dako) binds to an intracytoplasmic portion of the transmembrane receptor ERBB2. Despite the low expression of ERBB2 in neurofibromas, statistical analysis revealed that PNFs and MPNSTs expressed this receptor significantly more frequently than benign non-encapsulated neurofibromas (CNF, DNF, DPNF). This is in accordance with a recent study describing ERBB2 (HER2) amplification in 10/25 (40%) of MPNST cases, whereas no mutations or gene amplification were detected in neurofibromas (p<0.001) (34). In a previous study, no somatic mutations were found within tyrosine-kinase-encoding exons of EGFR and ERBB2. However, expression of EGFR and ERBB2 was frequently detected in MPNSTs at the protein level (32).
In contrast to rare ERBB2 expression in PNST in general, all neurofibromas showed weak to strong cytoplasmic and membranous staining for ERBB3 in our study. In this case, membranous staining was to be assumed to be exclusive. However, cytoplasmic reactivity has been observed in other tumor types and reflects endocytosis of the receptor (46). The intensity of ERBB3 staining increased with increasing malignancy of neurofibroma. Again, the difference between PNFs and MPNSTs compared with CNF, DNF, and DPNF was significant.
NRG1, a ligand of growth factor receptors, was demonstrated immunohistochemically in the cytoplasm and cell membrane in 87.1% of neurofibromas. This cytoplasmic staining pattern has been also described in another study (45). The NRG1 antibody (NRG1, AP 06168PU-N; Acris) used in the present study binds to the N-terminus, which in transmembrane NRG1 isoforms is located in the cytoplasm (38). The intensity and distribution of NRG1 staining was strongest in PNF.
Most studies have investigated the expression of NRG1 in human MPNSTs to demonstrate NRG1 playing a role in tumor progression and metastasis. In this study, NRG1 showed a significantly higher expression in the benign progenitor (PNF) of MPNSTs and thus may play a role in the progression of PNF to MPNST (15) in conjunction with other events such as p53 gene mutations (44).
Under physiological conditions, NRG1 is produced by axons as a signaling molecule for communication with associated Schwann cells (47). In neurofibromas, the origin of NRG1 has not been fully elucidated. Some studies assume autocrine or paracrine secretion by tumorous Schwann cells (30, 45). In a previous study by us using the same tumor cohort, axons were visualized with an antibody against neurofilament (14). Combining the data of the two studies, it was possible to relate axon density to NRG1 expression. It was found that tumors with many axons had significantly stronger NRG1 expression than tumors with low axon density. This may suggest that axons in neurofibromas are at least partially responsible for the formation of NRG1. On the other hand, evaluation of immunohistology clearly showed that tumor cells, especially in PNFs, expressed NRG1.
CD44. CD44 is a cell-surface glycoprotein involved in cellular adhesion and cell migration. Binding of CD44 to extracellular matrix components such as glycosaminoglycan hyaluronan triggers intracellular signal cascades (47, 48). Hyaluronan is the most common and immediate ligand for CD44 (49). In Schwann cells, CD44 enhances NRG1-mediated signaling (50). The expression of CD44 splice variants in malignant tumors is associated with metastatic potential of tumor and has prognostic significance (51). It has been shown that in NF1-associated PNST, significant amounts of hyaluronan are present in the extracellular matrix, presumably a product of tumorous Schwann cells (52). The current investigation showed that CD44 is expressed in all PNSTs. Riddle et al. showed that CD44 indicates invasive and infiltrative activity of neurofibromas (53). The results of the present study are in line with these findings in that diffusely growing benign neurofibromas and MPNST had the highest expression of this antigen. In contrast, nodular PNF not exhibiting infiltrative or invasive activity expressed significantly less CD44. These distributional differences of the protein in diffuse and encapsulated tumors may be important for the efficacy of drugs used to treat PNST (54-57).
Conclusion
Protein expression patterns differed significantly in subgroups of NF1-related PNST. The results of the present study indicate that DPNF, which comprises both diffuse and intraneural growth, seem to have more in common with CNF and DNF than with PNF, which are actually considered precursor lesions of MPNST. NRG1 was shown to be expressed predominantly in PNF and thus may play a role in tumor progression in association with other factors. The differences in expression of signaling molecules between tumor subtypes should be considered in therapeutic trials of NF1-related PNST.
Footnotes
Authors’ Contributions
Histological and immunohistological evaluation: LKNN and CH; clinical diagnosis and treatment of patients: REF and CH; drafting the article: REF and CH; final approval of the article: all Authors.
Conflicts of Interest
The Authors have no conflicts of interest regarding the work presented.
- Received February 21, 2022.
- Revision received March 28, 2022.
- Accepted March 31, 2022.
- Copyright © 2022 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.