Stepwise to clarify differentiation between pleomorphic xanthoastrocytoma and giant cell glioblastoma using p53, CD34 and KI67; clinicopathological hospital based study

Rare astrocytic neoplasms include pleomorphic xanthoastrocytoma (PXA) and giant cell glioblastoma (GCGB). Their respective incidence rates are roughly less than 1% of all brain tumours [1, 2]. Gross demarcation, the presence of multinucleated giant cells, the deposition of copious reticulin fibres, and the presence of lymphocytic infiltrate are among the histological and gross characteristics that are similar between them [1]. GCGB tumours are classified as WHO grade IV and have a less than 50% one-year survival rate [1, 2]. On the other hand, PXA is considered a WHO grade II brain tumour and has a substantially better prognosis than GCGB, with a 5-year recurrence-free survival rate of 72% [3, 4].

Roughly 5% of glioblastomas (GB) are GCGB. The large cell variation of GB has better circumscription, a slightly better prognosis, and a wider age range (affecting younger individuals) than more usual forms of GB [1, 2]. Between the primary (de novo) GB and the secondary GB resulting from a precursor lesion, GCGB is in the centre [4].

Kepes and colleagues first reported PXA, a primary neoplasm of the central nervous system. Teenagers and young adults are typically affected by this tumour, which is typically superficially placed and preferentially affects the temporal lobe [5]. WHO grade II tumours make up the majority of PXA cases at presentation. Anaplastic PXA (APXA, WHO grade III) is a subset of PXA that is characterized by enhanced proliferative activity (high mitotic index with a mitotic index of approximately ≥ 5 mitoses/10 HPF) and may recur [2].

Anaplastic transformation traits have been found in 15–20% of PXAs, according to an expanding body of research [3]. Two main categories can be used to classify APXA: primary, or de novo, and secondary, or developing on top of standard WHO grade II PXA [2].

Since GCGB and PXA have many morphological characteristics, it might be difficult and complicated to distinguish them histologically [1]. For example, reticulin fibre depositions are more typical in PXA, but they can also be observed in GCGB [1]. PXA, APXA, and GCGB are immunophenotypically reactive for neuronal markers such synaptophysin and the glial fibrillary acidic protein (GFAP) [1, 6]. Their glial origin is compatible with this immunophenotypic profile.

Due to the rarity of PXA and GCGB, there has been a slow but steady increase in the literature over the past ten years, but it has only included well-documented case reports of classic cases and classic examples with unique clinical presentations or courses [6, 7]. A limited number of studies have revealed the elevation of the endothelial/progenitor cell markers CD34 [8, 9] and p53 tumour suppressor protein in GCGB and PXA, respectively, at the molecular level [10, 11]. To the best of our knowledge, the three entities (PXA, APXA, and GCGB) have only been compared in two prior investigations [1, 12]. The purpose of this study was to close this gap in the body of literature. In order to achieve this, we examined the expression patterns of GFAP, synaptophysin, p53, GCGB, Ki67, and CD34 in a group of PXA, APXA, and GCGB using immunohistochemical staining techniques. Additionally, these entities' clinicopathologic characteristics were analysed.

The clinicopathological and immunohistochemical studies were conducted at the Department of Pathology, Faculty of Medicine, Al-Azhar University (Assuit branch, Assuit, Egypt) from April 2015 to June 2020. The specimens consisted of paraffin-embedded tumour blocks from 10 cases of pleomorphic xanthoastrocytoma (PXA), 3 cases of anaplastic pleomorphic xanthoastrocytoma (APXA), and 10 cases of glioblastoma with giant cell features (GCGB) neuropathological evaluation was done by two independent neuropathologists regarding the hematoxylin and eosin-stained slides without previous knowledge of the clinical data. The revised WHO 2021 categorization criteria for CNS tumours, using established histological and immunohistochemical procedures were used as basis for histopathological diagnosis [2]. Retrospective clinical information was gathered by reviewing the medical records of patients who underwent surgery at the neurosurgery department of the Faculty of Medicine at Assiut University and neurosurgery department of the Faculty of Medicine, South Valley University. The study rigorously upheld patient anonymity. The patients provided written consent on the day of the surgical procedure for their tissue utilization in these investigations. The work underwent evaluation and received approval from the ethical committee of Assiut University.

It involved the application of hematoxylin and eosin staining to tissue sections of 4–5 μm in thickness, which were obtained from tumour specimens fixed in formalin and embedded in paraffin. The observed features include pleomorphic appearance characterized by the presence of mononucleated and multinucleated tumour giant cells, xanthomatous cells, nuclear inclusion, perivascular lymphocytes, vascular growth, mitotic counts, and necrosis. The differentiation between PXA (WHO grade II) and anaplastic PXA (WHO grade III) was determined by the grading criteria of having at least 5 mitoses per 10 high power fields, as outlined in the 2021 WHO Classification scheme [2]. Reticulin fibre networks were considered "positive" if they were found in over 30% of the tumour matrix, regardless of the staining of blood vessels [12].

The expression pattern of GFAP, synaptophysin, p53, CD34, and Ki-67 was analysed using very sensitive peroxidase-streptavidin staining techniques. In summary, 4-μm unstained paraffin sections that were placed on silanized slides underwent the process of dewaxing and rehydration. The endogenous peroxidase activity was blocked and antigen retrieval was conducted. The tissue sections were incubated with primary antibodies against the following antigens: GFAP (polyclonal, dilution 1:400; Dako, Carpinteria, CA, USA), synaptophysin (polyclonal, dilution 1:400; Dako, Carpinteria, CA, USA), CD34 (monoclonal, clone QBEnd 10, dilution 1:400; Dako, Carpinteria, CA, USA), p53 (monoclonal, Clone: DO-7, dilution 1:200; Dako, Carpinteria, CA, USA), and Ki67 (monoclonal, clone MIB-1, dilution 1:50, antigen retrieval with heat; Dako). Next, the secondary antibody (Ultravision detection system anti-polyvalent horseradish peroxidase/3,3’-diaminobenzidine, ready-to-use, Netmaker) was administered according to the instructions provided by the manufacturer. The staining reaction signals were generated utilizing diaminobenzidine as a chromogen. Next, the sections were stained again using Mayer’s hematoxylin. Immunostaining was conducted with suitable positive and negative controls. Negative control slides were generated by replacing the primary antibody with a blocking buffer.

The immunostaining of the indicators under investigation was assessed as either positive or negative. GFAP and synaptophysin were identified as exhibiting cytoplasmic staining, as previously described [11]. The term “p53 protein positivity” refers to the condition where there is a significant buildup of p53 protein in the nuclei of more than 10% of the tumour cells [13]. Tumours with more than 10% of tumour cells showing CD34-cytoplasmic positivity were observed [11]. The assessment of Ki67 was conducted using a high-power objective lens with a magnification of 400 × . Positive results were recorded when obvious nuclear staining was observed. The regions exhibiting the greatest concentration of tumour nuclei with positive results (referred to as “hot spots”) were chosen. Subsequently, an average of 1000 tumour cells from these regions with the highest labelling intensity were enumerated for further study [14]. The immunostaining results were assessed separately by two neuropathologists, who were unaware of the clinical data. Statistical analyses: The data analysis was conducted using the SPSS statistical software Package, namely version 20. The connection of categorical variables was determined using Pearson's Chi-squared test and Fisher's exact tests. Significant differences were determined for p-values less than 0.05.

Tumour samples were collected from 23 patients, comprising 10 GCGB cases and 13 PXA cases (10 PXA cases and 3 APXA cases). Two instances of recurrent PXA and one primary incidence of APXA were included in the APXA. Patients suffering from PXA, APXA, and GCGB ranged in age from 18 to 55 years (mean ± SD: 29 ± 10.6 years), 19 to 42 years (mean ± SD: 31 ± 11.5 years), and 21 to 75 years (mean ± SD: 49 ± 16.9 years), in that order. Six of the ten GCGB patients were female, and four were male. Two women and eight males made up the PXA cases. The three APXA instances were all male. The parietal (4 cases) and temporal (3 cases) lobes were the primary locations of GCGB. The frontal lobe, occipital lobe, and corpus callosum, respectively, housed the final three cases. Four cases with PXA revealed a preferential localization in the temporal lobe, with three cases showing a preference in the parietal lobe. One case included the corpus callosum, while two cases were found in the frontal lobe. The parietal lobe was impacted in all three APXA patients (Table 1). Large pleomorphic cell populations were frequently found in pleomorphic xanthoastrocytomas, intermingled with spindle cells (80%). In the former, nuclear inclusion was present in 100% of the instances (p value = 0.001), whereas cytoplasmic vacuoles compatible with lipidization were present in 80% of the cases (p value = 0.003). However, the majority of GCGB cases lack these distinctive PXA histological findings (Fig. 1A–C). Microvascular proliferation, necrosis, and high mitotic counts were observed in both GCGB and APXA, and these findings are statistically significant (p values of 0.044, 0.001, and 0.002, respectively). We observed elevated mitotic figures ranging from 5–8/10HPF (mean ± SD; 6.33 ± 1.53) and 3–20/10HPF (mean ± SD; 9.1 ± 5.99) in APXA (WHO, grade III) and GCGB, respectively. Necrosis (80%) and microvascular proliferation (60%) were observed in GCGB and 2/3 in APXA, respectively. Rare mitotic figures in PXA (WHO, grade II) ranged from 0 to 3/10HPF (mean ± SD; 1.3 ± 1.16), but there was no necrosis or endothelial vascular alteration. In contrast to the PXA tumour cells, which exhibited reticulin fibre deposition around 70% of the tumour cells, the GCGB tumour cells only (30%) displayed reticulin fibre deposition surrounding the tumour cells. The PXA tumour cells exhibited frequent deposition of numerous reticulin fibres between tumour cells. Table 2 presents the specifics of these statistics.

Bild vergrößern

A robust expression of GFAP in the cytoplasm was seen in 9/10 of PXA, 3/3 of APXA, and 10/10 of GCGB. Regarding synaptophysin, it was found in 9/10 of GCGB, 3/3 of APXA, and 10/10 of PXA. P value = 0.507 indicates that there were no statistically significant variations in the expression of these proteins between PXA, APXA, and GCGB. The current investigation found statistically significant differences between GCGB and PXA in the expression profiles of CD34, p53, and Ki67 (Table 3). Seventy percent of PXA cases had high levels of CD34 protein expression (Fig. 2 B). On the other hand, GCGB tumour cells showed no expression of CD34 at all (p value = 0.001). On the other hand, Fig. 2 E shows that while only 30% of PXA cases expressed the p53 protein, the majority of GGGBM cases (80%) did (P value = 0.024). It is interesting to note that APXA's CD34 and p53 protein expression levels were closer to those of GCGB. The three APXA cases that were analysed all showed negative p53 immunoreactivity and no expression of the CD34 protein (Fig. 2 A and C; D-F). The current investigation showed that GCGB, PXA, and APXA have different Ki67 labelling indices (Fig. 2 G-I). GCGB exhibited the highest labelling indices, with a mean ± SD of 16.0 ± 5.0 and a range of 6.0% to 22.0%. PXA's labelling indices, which varied from 5.0% to 7.0% (mean ± SD6.0 ± 1.0%), were statistically considerably lower than APXA's, which ranged from 12.0% to 13.0% (mean ± SD12 ± 1.0%) (p value < 0.001).

Bild vergrößern

The neuroimaging studies of different pathological entities including pleomorphic xanthoastrocytoma, anaplastic pleomorphic xanthoastrocytoma and giant cell glioblastoma including preoperative MRI brain and postoperative CT brain are displayed in Figs. 3, 4 and 5).

Bild vergrößern

Bild vergrößern

Bild vergrößern

Rare glial astrocytic tumours like PXA and GCGB typically affect children’s and young adults’ superficial cerebral hemispheres [1]. Histopathological characteristics and reticulin stains mostly successfully address the differential diagnosis of PXA and GCGB. However, some instances are still unclear because of their small sample size, unusual histologic patterns, and expression of immunohistochemical markers [1], necessitating further diagnostic testing to reach the right diagnosis. Reticulin fibre depositions and CD34 expression, which are frequently used in PXA and GCGB, were found to be particularly useful as diagnostic markers in PXA. However, Lohkamp and colleagues [12] cautioned that these markers should not be used carelessly for the purpose of differentiating between the two entities, since dense reticulin fibre networks and CD34 expression were also found in over 25% of GCGB.

The term “gliomas” refers to a fairly broad category of intrinsic brain tumours that are often categorized based on microscopic similarities with potential source cells that belong to glial precursor cell lineages [15]. Our findings corroborate a number of other investigations that found elevated levels of GFAP and neural markers like synaptophysin in various tumour forms [10, 12, 16]. Because of PXA's high, diffuse GFAP staining, it was first thought to be glial in lineage [1].

Studies using immunohistochemistry and ultrastructural analysis have also revealed the presence of neuron-like fine features and neuronal immunophenotype [17, 18]. Consequently, our research lends credence to the theory that these tumour entities belong to the glioneuronal lineage.

All forms of cancer have TP53 mutations, which are genetic events that are assumed to be inactivated early in the carcinogenesis process [19]. Furthermore, p53 contributes to genomic stability through its roles in cell-cycle arrest and centrosome duplication [18, 20].To identify underlying TP53 gene alterations, p53 protein immunostaining is a trustworthy proxy method [21].

In contrast to low expression in PXA, the current investigation found strong P35 expression in GCGB and APXA. Our results more closely resemble the two comparative studies by Martinez-Diaz and colleagues [1], who found that p53 staining was found in many tumour cells in 5 out of 8 GCGB tested, while this antigen staining was negative or focally positive in 6 out of 8 PXAs, and Lohkamp and colleagues. [12], who found that nuclear p53 positivity was found in 74% of GCGB and that p53 positivity was less common in PXA cases (30%). Hirose and colleagues. (2008) reported, in contrast to our observations, that all tumour samples (12 PXA cases: six with conventional features and six with anaplastic features) showed evidence of p53 immunoreactivity. These differences could have to do with methodological evaluation and sample size. The results support the theory introduced recently by Cantero and colleagues [9] that the high rate of TP53 change in GCGB indicates that this gene is involved in these tumours and may be one of the first events in the gliomagenesis of GCGBs. The P53 protein is not involved in the pathophysiology of PXA, despite being well-known to exist in diffusely invading astrocytic gliomas, such as glioblastomas and anaplastic astrocytomas. These genetic variations appear to be the cause of PXA's unique biological characteristics.

Hematopoietic stem cells (HSCs) and hematopoietic progenitor cells were initially thought to be identified by CD34 [18]. Our findings showed that while CD34 was not expressed in GCGB or APXA, it was highly expressed in PXA. Our findings are consistent with the earlier reports by Hirose and colleagues [10] and Reifenberger and colleagues [11], which reported that tumour cells expressing CD34 were more frequently found in WHO grade II PXAs than in PXAs exhibiting anaplastic features. As far as we are aware, CD34 expression has not yet been investigated in relation to GCGB. Strong, widely distributed positivity was observed in giant cell glioblastomas (55%), according to Galloway and colleagues [22]. Our results are consistent with the findings of Lohkamp and colleagues [12], who observed low CD34 expression in GCGB (27%), in large series.

Relatively little is yet understood regarding CD34's role. Hematopoietic cell studies point to functions in cytoadhesion and the control of cell proliferation and differentiation [18, 23]. Hence, despite APXA and GCGB having less frequent and weaker CD34 immunoreactivity, PXAs had higher expression of CD34, which may help explain why these tumours' growth is more restricted. Moreover, could be a sign of dedifferentiation and suggest a function for decreased or lost CD34 expression in the development of tumours [11].

In many tumours, the proliferative activity, also known as the Ki67 proliferation/labelling index, is a significant prognostic indicator. High-grade gliomas and other high-grade tumours have high Ki67 indices [24].

We discovered comparatively high Ki67 labelling indices in GCGB and APXA, but not in PXA, as other researchers have [25‐27]. When combined with other research findings, our findings bring up a number of ideas. High-grade gliomas can be diagnosed with the help of the Ki67 labelling index. It can assist in identifying the anaplastic transition in PXA and has a significant prognostic function.

According to published research, between 9 and 20% of PXAs may develop a malignant transformation upon recurrence or may exhibit it at the time of first presentation [11]. Two secondary APXA cases were shown in this investigation; one patient experienced a tumour recurrence after a year, and the other after a year and a half. The intriguing immunohistochemistry results showed that CD34 expression had decreased and p53 expression had increased in both of the recurrent PXA patients. As a result, APXA showed an immunological profile (negative CD34 and positive P53) that is comparable to GCGB, with positive synaptophysin and GFAP in both tumours.

Certain immunohistochemical patterns (positive p53 and negative CD34) for these neoplasms were observed in a number of published case reports of PXA transforming into APXA or GB [18, 25, 28]. The idea that p53’s positive expression and CD34’s negative expression are distinctive features of GB and APXA is further supported by our data. Another theory is that PXA carcinogenesis may not be initiated by a mutation in the p53 gene, but rather may be accelerated by it. Loss of CD34 expression concurrently could be a sign of dedifferentiation.

However, the precise biological processes behind the accumulation of p53 proteins in PXA as well as their predictive significance are yet unknown. The information supports the theory that APXAs are thought to display the interphase between PXA and lipidized GCGB [18].

To conclude, our work revealed frequent expression of CD34 in PXA, despite the complete absence of its expression in GCGB. In contrast, p53 is frequently expressed in GCGB and infrequently expressed in PXA. This study indicated that a panel of immunohistochemical stains, CD34, and p53 may provide robust diagnostic markers for PXA (positive CD34 and negative p53) and GCGB (negative CD34, and positive P53). Our results revealed that APXA immunohistochemically negative CD34 and positive p53 which is similar to GCGB. This data is in favour of the sequential transformation of APXA into glioblastoma (giant and/or epitheloid), presumably similar molecular alterations of p53 and CD34 have possible roles. The diagnostic and prognostic ramifications of our findings are open for future investigations.