Utility of Immunohistochemistry in Subtyping Posterior Fossa Group A Ependymoma: A Retrospective Study

Abstract

Introduction

Posterior fossa ependymomas are classified into posterior fossa group A (PFA-EPN) and posterior fossa group B ependymomas. PFA-EPN shows immunohistochemical loss of H3 p.K27me3 expression. Molecular subgroup 1 (PFA-1) and 2 (PFA-2), encompass nine subtypes (PFA 1a–1f and 2a–2c). OTX2 and H3 p.K27M immunopositivity is noted in PFA-2c and 1f, respectively, with potential prognostic implications.

Aim

To assess the frequency of OTX2 and H3 p.K27M immunopositivity in PFA-EPN.

Materials and Methods

This retrospective study included PFA-EPN diagnosed at our institute from 2016 to 2022, based on immunohistochemical loss of expression of H3p.K27me3. Immunohistochemistry for OTX2 and H3 p.K27M was carried out on them.

Results

A total of 42 cases of PFA-EPN were encountered, ranging 10 months to 23 years, with a median of 4 years. OTX2 immunopositivity was seen in four cases (9.5%) and H3 p.K27M positivity in two cases (4.8%). Follow-up data were available partially and showed variable survival.

Conclusion

PFA-EPN can be segregated into OTX2 and H3 p.K27M immuno-positive tumors. Because of low frequency and few studies, long-term survival data are limited. Assessment of frequency of OTX2 and H3 p.K27M immunopositivity and their segregation into subsets with prognostic significance can help in diagnostics in routine laboratory settings. This can also potentiate open new therapeutic avenues in PFA-EPN.

 

MissingLinkText

Keywords

ependymoma - H3K27M - OTX2 - brain tumor - posterior fossa

Introduction

Ependymomas (EPNs) are a group of tumors showing heterogeneity across age groups, location, and prognosis. The World Health Organization (WHO) Classification of Tumors of the Central Nervous System divided EPNs based on molecular features and location into nine subgroups.[1] Among them, posterior fossa EPNs (PF-EPNs) comprise posterior fossa group A (PFA-EPN) and posterior fossa group B (PFB-EPN) EPNs, with very few cases of posterior fossa sub-EPNs. PFA tumors are predominantly seen in infants and children, median age being 3 years. They are seen to have poor prognosis than PFB-EPN.[2]

In 2018, Pajtler et al performed DNA methylation profiling of 675 cases of PFA-EPNs. They categorized PFA-EPN into two major subgroups: PFA-1 and PFA-2, which again subcategorizes into nine molecular subtypes—PFA-1a to 1f and PFA-2a, 2b, 2c. PFA-1 and PFA-2 have differences in gene expression profiles associated with different clinical variables and pattern of relapses. Specifically, PFA-2c was evidently distinct and showed an overall survival (OS) of >90%, which was significant and a contrast to the notion that PFA tumors have a poor survival.[2]

Hence, there was a stark contrast and heterogeneity of survival within PFA-EPNs, with PFA-2c EPN showing better OS and progression-free survival. We explored the utility of immunohistochemistry (IHC) with Orthodenticle Homeobox protein 2 (OTX2)—a transcription factor involved in development of different parts of the central nervous system (CNS)—for the detection of PFA- 2c EPN. OTX2 has earlier been used to subcategorize medulloblastoma and is detailed later in this article. Also, another subcategory with histone mutations (H3p.K27M), found to be enriched in PFA-1f subtype, could be detected by IHC for H3p.K27M. This mutation has been shown to confer particularly poor prognosis in glioma and may hold a key in prognosis of PFA-EPN as well.[1]

 

MissingLinkText

Materials and Methods

This was a retrospective study of 7-year duration from 2016 to 2022 in the Department of Neuropathology in a tertiary center of Neurosciences. Archival paraffin blocks diagnosed as PF-EPN in that time period were retrieved. H3 p.K27me3 immunostain (Medasys, MD48R. 1:200) was used to classify the cases into PFA cohort. Inclusion criteria comprised histologically proven PF-EPN and H3 p.K27me3 loss by IHC. The cases showing diagnoses other than EPN on review and/or not satisfying H3 p.K27me3 loss were excluded from the study. Applying these criteria for selection, a total of 42 cases were found. Tissue microarray (TMA) was carried out with Beecher Manual Tissue Array instrument. In this process, three cores of tumor tissue were taken from each sample, and one microarray block was prepared. 4-micron sections were taken from each such TMA block.

Ventana BenchMark XT was used for automated immunostaining. OTX2 (Invitrogen, IH12C4B5, 1:100) and H3 p.K27M (Medasys, RM192, 1:200) immunostains were done on the PFA-EPN cohort. Microscopy was done by two neuropathologists and based on qualitative interpretation, strong nuclear staining of OTX2 and H3 p.K27M was considered positive.

Follow-up was attempted for the patients and follow-up data were recorded for the ones available.

 

MissingLinkText

Results

A total of 42 cases of PFA-EPN were encountered. The age range was 10 months to 23 years, with median age being 4 years. OTX2 immunopositivity was seen in four tumors (9.5%). Two cases showed positivity for H3 p.K27M (4.8%; [Table 1] and [Fig. 1]).

Fig. 1 Bar diagram showing positivity of OTX2 and H3 p.K27M-positive tumors.

Table 1 Cases showing either OTX2 or H3 K27M immunopositivity

Case no.	Age (y)	Gender	OTX2	H3K27M	Current survival status
3	1.5	M	Positive	Negative	Expired post-radiotherapy
21	2	M	Positive	Negative	No progression, doing well
11	12	M	Positive	Negative	–
41	2	M	Positive	Negative	–
22	10 m	F	Negative	Positive	No progression, doing well
40	2	M	Negative	Positive	No progression, doing well

Abbreviations: F, female; M, male.

Four OTX2-positive tumors were seen in patients ranging from 1.5 to 12 years ([Fig. 2A]–[C]). All of the four tumors showed a high-grade morphology (CNS WHO grade 3), with ependymal pseudorosettes, sheets, and increased mitosis. These cases showed varied survival on follow-up, which was available in two out of four patients (patient #3 and #21).

Fig. 2 Representative case (#3): PFA-ependymoma showing perivascular pseudorosettes, sheets and high cellularity (H&E) (A), H3 p.K27me3 loss (B), and OTX2 positivity (C). PFA, posterior fossa group A.

Patient #3, an 18-month-old male child, deteriorated post-radiotherapy and succumbed in 5 months from surgery. Patient #21 was a 2-year-old child presented with headache and gait ataxia of 2-month history. Post gross total resection, and multiple shunt revisions postoperatively, the patient had developed meningitis, but recovered completely and magnetic resonance imaging scans showed no residual lesion. Adjuvant chemotherapy and radiotherapy were given; the patient had no recurrence and doing well.

There was no follow-up data available for patient #11 and #41, which showed OTX2 immunohistochemical positivity.

As mentioned, two cases were positive for H3 p.K27M ([Fig. 3A]–[C]). One of these was the youngest patient of the cohort (10-month-old girl, #22). Another was a 2-year-old male child, who showed positive H3 p.K27M. Both these patients were treated with gross total resection, followed by adjuvant radiotherapy, and currently showing no progression. In both cases, histology showed sheets of neoplastic ependymal cells along with perivascular pseudorosettes. Mitosis was high. The tumors corresponded to CNS WHO grade 3.

Fig. 3 Representative case (# 22): PFA-ependymoma showing pseudorosettes (H&E) (A), H3p.K27me3 loss (B), and H3 p.K27M positivity (C). PFA, posterior fossa group A.

For other cases (36 cases) where neither of the IHC stains were positive, follow-up data were available for 16 of them. In 11/16, there was no recurrence ([Table 2]).

Table 2 All cases with dataset in the study

Pt. No.	Age (y)	Sex	Diagnosis	H3 K27me3	OTX2	H3 K27M	Follow-up	Recurrence	Adjuvant therapy taken	Current status
1	6	M	PFA-EPN grade 3	L			Y	No	RT and CT	Alive
2	2	M	PFA-EPN grade 3	L			Y	Yes, with CSF dissemination	RT	Alive
3	1.5	M	PFA-EPN grade 3	L	P		Y	–	RT	Dead after RT
4	7	F	PFA-EPN grade 3	L			Y	No	RT	Alive
5	2	M	PFA-EPN grade 3	L			Y	No	RT	Alive
6	6	M	PFA-EPN grade 3	L			Y	Recurrence	RT	Alive
7	4	F	PFA-EPN grade 3	L			Y	No	Not advised	Alive
8	6	M	PFA-EPN grade 3	L			N
9	4	F	PFA-EPN grade 2	L			N
10	3	M	PFA-EPN grade 3	L			N
11	12	M	PFA-EPN grade 3	L	P		N
12	5	F	PFA-EPN grade 3	L			N
13	3	M	PFA-EPN grade 3	L			N
14	12	F	PFA-EPN grade 3	L			Y	Small residual lesion, no growth	RT, CT	Alive
15	23	M	PFA-EPN grade 3	L			Y	No	Not taken	Dead
16	4	M	PFA-EPN grade 3	L			N
17	21	M	PFA-EPN grade 2	L			Y	No	RT	Alive
18	3	M	PFA-EPN grade 3	L			Y	Yes, with CSF dissemination	RT	Alive
19	1	M	PFA-EPN grade 3	L			N
20	3	M	PFA-EPN grade 3	L			N
21	2	M	PFA-EPN grade 3	L	P		N
22	0.8	F	PFA-EPN grade 3	L		P	N
23	2	M	PFA-EPN grade 2	L			N
24	17	M	PFA-EPN grade 3	L			N
25	2	F	PFA-EPN grade 3	L			N
26	8	M	PFA-EPN grade 3	L			N
27	2	M	PFA-EPN grade 3	L			N
28	1	M	PFA-EPN grade 3	L			N
29	9	M	PFA-EPN grade 3	L			Y	No	RT	Alive
30	5	M	PFA-EPN grade 3	L			Y	Recurrence	RT	Dead
31	8	M	PFA-EPN grade 3	L			Y	No	RT	Alive
32	9	F	PFA-EPN grade 3	L			N
33	4	F	PFA-EPN grade 3	L			Y	No	RT	Alive
34	4	F	PFA-EPN grade 3	L			Y	No	RT	Alive
35	9	M	PFA-EPN grade 3	L			N
36	14	M	PFA-EPN grade 3	L			Y	No	RT	Dead
37	2	F	PFA-EPN grade 3	L			N
38	8	M	PFA-EPN grade 3	L			N
39	3	F	PFA-EPN grade 3	L			N
40	2	M	PFA-EPN grade 3	L		P	N
41	2	M	PFA-EPN grade 3	L	P		N
42	6	F	PFA-EPN grade 3	L			N

Abbreviations: CT, chemotherapy; F, female; L, loss of expression; M, male; N, no; P, positive; RT, radiotherapy; Y, yes.

 

MissingLinkText

Discussion

PF-EPNs comprised about 40% of all EPNs in a recent study.[3] Lyons and Kelly in 1991 in their review article on PF-EPN suggested that children <5 years had a poor prognosis.[4] Later, PF-EPNs were divided into group A and Group B based on their demographic, transcriptional, genetic, and clinical differences.[5] PFA-EPNs were found to be associated with a pediatric age group and consistently associated with poor prognosis even after cytoreductive surgery and adjuvant chemotherapy.[6]

In 2018, a study conducted by Pajtler et al performed DNA methylation profiling of 675 cases of PFA-EPN. Methylation profile clustered in two different areas, categorizing the tumors into PFA-1 and PFA-2. Furthermore, PFA-1 was subcategorized into 1a till 1f and PFA-2 into 2a, 2b, and 2c. This led to classification of PFA into nine methylation subclasses. Among them, PFA-2c showed high levels of OTX2 expression and PFA-1f showed enriched H3 p.K27M mutations. Patients of PFA-2c subtype were found to be having OS of >90% at 5 years. This is significant for the fact that traditionally PFA-EPNs are known to show aggressive behavior and poor prognosis.[1]

In routine diagnostic setting, histomorphology and IHC play an important role in diagnosis. Hence, our quest was to estimate the frequency of OTX2 overexpression and H3 p.K27M mutations as they could be identified by IHC and used in routine laboratory setting. The study by Pajtler et al was a major background article of reference which imparted the felt need in our work. This was relevant because it could have prognostic implications on the PFA-EPN, mentioned above.

OTX2 is a transcription factor, gene for which is in chromosome 14. It is essential for development of eye and optic nerve. OTX2 also was found to implicate a role in altering the dynamics of neuronal progenitor cell proliferation. Accumulation of proliferative clusters of cells in the cerebellum and brain stem which resulted from OTX2 overexpression showed medulloblastoma characteristics. OTX2 directly inhibits cell differentiation of medulloblastoma cells. Indeed, WNT activated and non-WNT, non-SHH medulloblastomas are found to overexpress OTX2, which is detectable routinely by IHC. Also, since OTX2 was found to have a role in the trimethylation of H3K27 by sustaining a bivalent-like state of OTX2-binding promoters, it is possible that OTX2 may have a role in reduced H3K27me3 in PFA-EPN.[7] [8] OTX2 has also been touted as a therapeutic target in retinoblastoma, which can also be potentially significant for PFA-EPN.[9] Hence, identification of OTX2 by an easier laboratory diagnostic modality like IHC would firstly help to potentially categorize PFA-EPN on prognosis, and also open new possibilities in its management.

Pajtler et al showed OTX2-overexpressing tumors to be localized in the PFA 2c subclass. In our study, OTX2-expressing tumors were found to be 9.5%. To the best of our knowledge, any other such studies are lacking and identification of these tumors by IHC is important for prognosis.[1]

H3 p.K27M mutations are mostly noted in diffuse midline gliomas (DMGs; 96%) and very little data are available on their occurrence in PFA-EPN.[10] Pajtler et al showed 4.2% of PFA-EPNs harbor K27M mutations. The trimethylation of H3K27 is hindered by EZHIP overexpression in most cases of PFA-EPN. However, in their study, a subset did not show EZHIP overexpression. These were predominantly enriched in PFA-1f, which showed H3K27M mutations. Nambirajan et al also found EZHIP-overexpressed tumors to be mutually exclusive from H3K27M-mutated ones. This exclusivity is also seen in DMG, where the minor subset showing EZHIP overexpression and the predominant subset with H3K27M mutations are mutually exclusive. In this study, frequency of K27M mutations was found to be 1/41.[1] [11] Mariet et al[12] carried out a recent study on K27M mutant PFA-EPN, where they performed a comprehensive clinico-radiological, histopathological, genetic, and epigenetic characterization of PFA-EPN H3 p.K27M-mutant. The frequency of K27M mutations was 6%. They found that these tumors are more in the midline and histologically similar to EZHIP overexpressing counterparts. In our study, out of 42 cases, two were found to be H3 K27M-positive (4.8%). This frequency is like the above-mentioned studies, and closer to the findings of Pajtler et al. As very little data on K27M mutations in PFA-EPN are there, its implications on prognosis or histological correlation are yet to be ascertained. Our study could not only add to the existing scant data of H3 p.K27M mutant PFA-EPN, but also more importantly would aid to the immunohistochemical categorization of these tumors.

In low-income or developing countries, DNA methylation profiling can be an expensive affair and facilities related to same are unavailable in most. Although frequency is low, a definite subset of cases has emerged in our study, showing OTX2 and H3 p.K27M immunopositivity in PFA-EPN. Correlated with the study by Pajtler et al, we emphasize the need for such segregation especially where DNA methylation is out of reach for majority of the population. However, surrogate markers need to be used with caution. To ensure proper representation of the mutations and gene profiles by IHC; interdisciplinary work by correlation between clinical radiological and histological features is of utmost importance.

The study duration was of 7 years (2016–2022) and the follow-up data were inadequate because they were available for 17/42 cases only. Also, varied survival of OTX2-positive tumor cases were seen, one of which was death post-radiotherapy and other patient surviving after treatment—showing effect of treatment-related morbidity on patient outcome. Hence, an analysis on survival and progression for the OTX2, H3 p.K27M-positive cases and their comparison with negative cases could not be done. However, this study can be a stepping stone in a right direction for future studies.

 

MissingLinkText

Conclusion

PFA-EPN, though recognized as an aggressive tumor, can be classified on IHC to segregate a potentially favorable subgroup which is OTX2-positive. Although a previous single large-scale study though indicates favorable outcome in this subset, further trials are needed. This can help to prognosticate a favorable subset of PFA-EPN. H3 p.K27M mutant tumors in PFA-EPN comprise a very rare subset, which is also amenable to be diagnosed by IHC in routine laboratory setting, without stepping into molecular diagnostics. The significance regarding prognosis though unclear at present because of paucity of cases and studies, future research studies may open new avenues for prognostic and subsequent therapeutic ventures.

 

MissingLinkText

 

MissingLinkText

Conflict of Interest

None declared.

Acknowledgement

The authors acknowledge technical staff for their efforts in making this study possible.

• References

• 1 Ellison DW. Central nervous system tumours. WHO Classification of Tumours. 5th ed.. Vol 6. Lyon: International Agency for Research on Cancer (IARC); 2021
• 2 Pajtler KW, Wen J, Sill M. et al. Molecular heterogeneity and CXorf67 alterations in posterior fossa group A (PFA) ependymomas. Acta Neuropathol 2018; 136 (02) 211-226
• 3 Kresbach C, Neyazi S, Schüller U. Updates in the classification of ependymal neoplasms: The 2021 WHO Classification and beyond. Brain Pathol 2022; 32 (04) e13068
• 4 Lyons MK, Kelly PJ. Posterior fossa ependymomas: report of 30 cases and review of the literature. Neurosurgery 1991; 28 (05) 659-664 , discussion 664–665
• 5 Witt H, Mack SC, Ryzhova M. et al. Delineation of two clinically and molecularly distinct subgroups of posterior fossa ependymoma. Cancer Cell 2011; 20 (02) 143-157
• 6 Ramaswamy V, Hielscher T, Mack SC. et al. Therapeutic impact of cytoreductive surgery and irradiation of posterior fossa ependymoma in the molecular era: a retrospective multicohort analysis. J Clin Oncol 2016; 34 (21) 2468-2477
• 7 Bunt J, Hasselt NE, Zwijnenburg DA. et al. OTX2 directly activates cell cycle genes and inhibits differentiation in medulloblastoma cells. Int J Cancer 2012; 131 (02) E21-E32
• 8 Zagozewski J, Shahriary GM, Morrison LC. et al. An OTX2-PAX3 signaling axis regulates Group 3 medulloblastoma cell fate. Nat Commun 2020; 11 (01) 3627
• 9 Li J, Di C, Jing J, Di Q, Nakhla J, Adamson DC. OTX2 is a therapeutic target for retinoblastoma and may function as a common factor between C-MYC, CRX, and phosphorylated RB pathways. Int J Oncol 2015; 47 (05) 1703-1710
• 10 Gessi M, Capper D, Sahm F. et al. Evidence of H3 K27M mutations in posterior fossa ependymomas. Acta Neuropathol 2016; 132 (04) 635-637
• 11 Nambirajan A, Sharma A, Rajeshwari M. et al. EZH2 inhibitory protein (EZHIP/Cxorf67) expression correlates strongly with H3K27me3 loss in posterior fossa ependymomas and is mutually exclusive with H3K27M mutations. Brain Tumor Pathol 2021; 38 (01) 30-40
• 12 Mariet C, Castel D, Grill J. et al. Posterior fossa ependymoma H3 K27-mutant: an integrated radiological and histomolecular tumor analysis. Acta Neuropathol Commun 2022; 10 (01) 137

Publication History

Article published online:
07 July 2025