This article is mentioned in:

Introduction

The autophagic flux is a catabolic biological process that maintains cellular homeostasis through various functions, including the degradation and removal of damaged substrates and macromolecules (1–3). Autophagic activity increases under stress, with the formation of autophagosomes regulated by several autophagy-related proteins (ATGs) (4). Autophagy may play a dual role, acting as either a tumor promoter or a tumor suppressor in different human tumors (5,6).

Chordomas (CHs) are rare bone tumors that originate from remnants of the notochord, likely driven by brachyury activation, and account for approximately 1.5% of primary malignant bone tumors (7). They can occur along the midline of the spine, extending from the clivus to the sacrum, and are located anterior to the spinal cord (8). CHs typically affect adults, although some cases have been also reported in children. The most common site for CHs is the sacrum-coccygeal region (50%), followed by the clivus and spheno-occipital area (30–35%). The remaining 10–15% are found in the vertebral body (9). The physical findings and clinical signs associated with CHs depend on their specific location. Clival and skull base CHs typically present with headaches, cranial neuropathies, endocrinopathies, and, rarely, rhinorrhea due to cerebrospinal fluid leaks (8). Vertebral and sacral CHs often exhibit nonspecific localized pain and may lead to pathological fractures, radiculopathy, or myelopathy, as well as dysfunction of the bladder and bowel due to involvement of the autonomic nervous system (8). The evaluation of CHs relies on imaging techniques such as computed tomography (CT) and magnetic resonance imaging (MRI). CT is more effective in demonstrating lytic bone destruction and irregular dystrophic calcification (8), while MRI provides better delineation of the extent of CHs, typically showing lower signal intensity on T1-weighted images and heterogeneous contrast enhancement or hyperintensity with gadolinium contrast (10).

The main macroscopic findings of CHs consist of firm masses that may contain fluid, gelatinous substances, and/or areas of hemorrhage and necrosis. Additionally, calcification and sequestered bone fragments may be present (11). Microscopically, CHs are characterized by the presence of typical ‘physaliphorous’ cells, which exhibit vacuoles in their bubbly cytoplasm (7,11). Currently, three subtypes of CHs are recognized: conventional, chondroid, and poorly differentiated/dedifferentiated, with the latter being the least common and associated with the poorest prognosis (8,11).

Although an association between the development of CHs and autophagic flux can be hypothesized, similar to findings documented in other intracranial and bone malignancies (2,12), the role of ATGs is yet to be fully elucidated. Therefore, this study aimed to investigate the immunoexpression of certain ATGs, including microtubule-associated protein 1 light chain 3 (LC3A/B), p62, and activating molecule in Beclin-1 regulated autophagy (AMBRA-1), in our series of sporadic adult conventional clival CHs.

Materials and methods

The analysis was conducted according to the Good Clinical Practice guidelines and the Declaration of Helsinki (1975, revised in 2013). Prior to the surgical procedures, all patients provided written, anonymized, and informed consent. Pathology reports and medical records were thoroughly reviewed. Patients' initials and other personal identifiers were removed from all images. The Institutional Review Board of the University Hospital of Messina (Messina, Italy) approved this study (prot. N. 47/19; May 2, 2019).

Case selection

A cohort of 10 cases of clival CHs removed through neurological resection and collected between 2011 and 2022 was obtained from the archives of the Department of Human Pathology of Adult and Developmental Age at the University of Messina, Messina, Italy. Clinical and pathological parameters, including age, sex, tumor site, growth fraction, neurological status, imaging appearance, surgical/radiotherapeutic treatment, and clinical course, were collected for all cases of clival CHs. Follow-up data were available for eight out of the 10 cases. All patients underwent gadolinium contrast-enhanced T1, T2, FLAIR, and MPR brain MRI, supplemented with CT scans for surgical navigation and when intralesional calcifications were observed.

Immunohistochemistry

The immunohistochemical analysis was performed on 5-micron thick sections taken from paraffin-embedded tissue blocks. The sections were deparaffinized and washed using a descending alcohol gradient. A treatment with 3% hydrogen peroxide for 10 min was employed to eliminate endogenous peroxidase activity. After three rinses in deionized water, the sections were incubated for 30 min at room temperature with normal sheep serum to prevent nonspecific protein binding. Subsequently, the sections were incubated for 30 min at 37°C with primary polyclonal rabbit anti-human antisera against p62 (working dilution 1:250; Abcam), AMBRA-1 (working dilution 1:250; Abcam), and LC3A/B (working dilution 1:100; Abcam). Following three rinses with phosphate-buffered saline (PBS), the sections were incubated with a biotinylated goat anti-rabbit IgG secondary antibody (1:300; Abcam) for 20 min at room temperature. Finally, the sections were incubated with a horseradish peroxidase-labeled secondary antibody for 30 min, and the immunoreaction was visualized using diaminobenzidine tetrahydrochloride and counterstained with hematoxylin, utilizing the ULTRA Staining System (Ventana Medical Systems). Specific primary antisera were omitted and replaced with PBS to serve as negative controls.

The intensity of positive cells for p62, AMBRA-1, and LC3A/B was recorded according to previously reported immunohistochemical methods. The cytoplasmic staining intensity was rated as follows: 0, negative; 1, weak; and 2, strong. The percentage of positive cells was scored as follows: grade 0, 0–5%; grade 1, >5-25%; grade 2, >25-50%; grade 3, >50-75%; and grade 4, >75-100% for all ATGs. Immunohistochemical scores were independently assigned by two pathologists (AI and GT), who were blinded to patient information and other clinical characteristics. Scoring was conducted using a Zeiss Axioskop microscope (Carl Zeiss Microscopy GmbH) at 40× objective magnification. A kappa value ranging from 0.73 to 0.80 (substantial agreement) was documented for interobserver agreement.

Multiplying the staining intensity by the percentage of positive cells yielded a final score ranging from 0 to 6. Cases with an immunoreactive score of 0 to 3 were considered negative, while those with a score of 4 to 6 were classified as positive.

Ki-67 antiserum (clone MIB-1, dilution 1:100, Dako Corp., Glostrup, Denmark) was applied to slides for 30 min at room temperature to determine the growth fraction of meningiomas. This was preceded by antigen retrieval, which was performed three times in 0.01 M citrate buffer (pH: 6.0) using a microwave oven set to 750 W. The Ki-67 labeling index (LI) was calculated as the mean percentage of stained nuclei by counting 1,000 tumor cells across three representative neoplastic fields. A median Ki-67 LI value of 3% was established as the cut-off point to differentiate between low and high Ki-67 expression.

Statistical analysis

Statistical evaluation was performed using the SPSS version 13.0 software package (SPSS, Inc.). The Chi-square (χ2) test or Fisher's exact test was employed to analyze the relationship between autophagic immunoexpression markers (LC3A/B, p62, and AMBRA-1) and various clinicopathological parameters (age, sex, growth fraction, neoplastic volume, and surgical treatment), in relation to recurrence. A p-value of less than 0.05 was considered statistically significant.

Results

The clinico-surgical, pathological, and therapeutic features of clival CHs in this study are summarized in Table I.

Table I. Clinical, surgical, pathological and therapeutic features of CHs.

Table I.

Clinical, surgical, pathological and therapeutic features of CHs.

Sex	Age	Site	Associated diseases	Neurological status	MRI	Size (cm)	Volume (cm3)	Treatment Endoscopic endonasal	RT	Recurrence
F	70	Clivus	Hypertension,	Strabismus, dysarthria,	Non-homogenous,	2.2×2.3×2.5	17	Partial removal	No	Yes
			Diabetes II	left hemiparesis	discrete enhancement
M	48	Clivus	-	Dysphagia	Non-enhancing	1.7×1.5×1	3	Complete removal	No	No
F	77	Clivus with retro	Hypertension,	Left hemiparesis	Non-homogenous	2×2×1	8	Partial removal	Yes	Yes
		and intrasellar	Diabetes II, Goiter		weak enhancement
					with calcifications
M	26	Clivus with	-	Diplopia Right VI cn	Non-homogenous	3×3×2	18	Complete removal	No	NA
		retroclival and intrasellar		paresis	weak enhancement
M	40	Middle clivus,	Esophageal achalasia,	Negative	Non-homogenous	2×1×2	3	Complete removal	No	No
		retroclival and	sarcoma, bladder		weak enhancement
		intrasellar	carcinoma
M	63	Inferior clivus	Hypertension	Arms paresthesia	Isointense	4×3×4	48	Partial removal	Yes	Yes
			Diabetes II		non-enhancing,
					calcifications
F	58	Clivus	Charcot-Marie Tooth	Walking deficits Left	Non-enhancing	4×4×3	18	Partial removal	Yes	No
			1A Diabetes II Bell paralysis	hemiparesis
M	68	Clivus	-	Left hemiparesis	Non-homogenous	3×2×2	10	Partial removal	Yes	Yes
					weak enhancement
F	75	Inferior clivus	Hypertension	Negative	Isointense non-enhancing,	2×2×1	6	Complete removal	No	No
			Diabetes II		calcifications
M	65	Clivus	-	Diplopia	Non-homogenous weak	1.9×1.8×1	4	Complete removal	No	No
					enhancement

[i] MRI, magnetic resonance imaging; NA, not available; RT, radiotherapy.

The patient ages ranged from 26 to 77 years (mean 59 years), with a male-to-female (M:F) ratio of 6:4. All tumors were located in the clivus, occasionally extending into the middle or inferior portions with retroclival involvement. Upon admission, patients presented with a wide array of neurological symptoms, ranging from asymptomatic cases (where the lesion was discovered during a CT scan for oncological follow-up of urothelial carcinoma) to cranial nerve deficits (diplopia, dysphagia, and strabismus), arm paresthesia, and walking deficits.

Lesions appeared as solid masses isointense to brain parenchyma in both standard T1 and T2-weighted images (Fig. 1A and B). Following contrast administration, a wide range of enhancement was observed, from non-enhancement in three cases, weak enhancement in three cases, to discrete enhancement in the last case. Calcifications were noted in two cases. In one case, the CH remained confined to the occipital bone, while in the remaining cases, it extended extradurally in all directions, including the cervical spine at the level of the extracanalar laterocervical area, the posterior cranial fossa, cavernous sinus, Meckel's cave, sphenoidal sinus, and intrasellar region, resulting in compression of the pituitary gland. The mean tumor size was 2,61 cm (range: 1,7-4 cm), with the mean neoplastic volume measured at 13,5 cm3 (range: 3–48 cm3). MRI findings were available for all cases. The follow-up period for patients ranged from one to 60 months, with a mean follow-up of 18.8 months. The associated disease, neurological status, and magnetic resonance features are described in Table I. Surgical treatment was performed via endonasal endoscopy in all cases, achieving complete removal of the clival CH in five cases. Radiotherapy was not administered in six cases, while three cases received stereotactic radiotherapy with a 30 Gy isodose delivered in five fractions, and one case underwent proton beam therapy at 74 Gy. During the follow-up period, no patients died from the disease; however, one patient was lost to follow-up, and there was no record of ongoing therapy. Another patient refused radiotherapy and experienced an increase in the residual lesion a few months postoperatively. Four patients experienced disease recurrence at one, nine, and 17 months, respectively.

Figure 1. (A) Computed tomography scan with bone reconstruction algorithm showing a small non-enhancing bone lytic lesion with irregular borders at the level of the middle clivus. The lesion measured 1.7×1.5×1 cm and was centered in the right paramedian area. (B) Contrast-enhanced magnetic resonance study revealing a clival lesion with moderate enhancement, infiltrating and replacing the upper half of the clival cortical bone, extending widely toward the prepontine cistern. The lesion measures ~32×30×20 mm, infiltrates the dorsum sellae and shows close continuity with the pituitary gland, both carotid arteries, and the basilar trunk, reaching the sphenoid sinus.

All cases exhibited the classical appearance of CH with typical physaliphorous cells (Fig. 2A) and consistent immunoexpression of cytokeratin (CK) AE1/AE3 (Fig. 2B), CK 8, CK 18, CK 19, epithelial membrane antigen (EMA), and S100 protein. In contrast, they were negative for CK 7 and CK 20. Additionally, a clear nuclear expression of Brachyury was demonstrated (Fig. 2C), as reported in other studies on CHs (11,13).

Figure 2. (A) Characteristic microscopic appearance of physaliphorous cells present in CHs (hematoxylin and eosin; magnification, ×200); (B) the same elements exhibited evident cytoplasmic immunoreactivity for CK AE1/AE3 (Haemalum's counterstain; magnification, ×160); and (C) nuclear staining for Brachyury (magnification, ×160). CHs, chordomas; CK, cytokeratin.

Immunostaining for LC3A/B (Fig. 3A), p62 (Fig. 3B), and AMBRA-1 (Fig. 3C) was exclusively observed in neoplastic cells, with no expression detected in the surrounding stromal cells. Both LC3A/B and p62 were expressed in the cytoplasm and nucleus of neoplastic cells (Fig. 3A and B), while AMBRA-1 was predominantly localized in the cytoplasm (Fig. 3C). All CH cases exhibited a constant high immunoreactivity for p62. In contrast, low LC3A/B staining was found in five out of 10 cases (50%), while five cases demonstrated high LC3A/B immunoexpression; four of these cases were characterized by neoplastic recurrence and partial removal. Finally, AMBRA-1 low immunoreactivity was noted in seven out of 10 cases (70%), while three recurrent cases exhibited high AMBRA-1 immunostaining. In terms of growth fraction, five cases were considered high Ki-67 expression, with a Ki-67 LI >3%, four of which had undergone partial neoplastic removal and showed evidence of recurrence.

Figure 3. (A) LC3A/B (magnification, ×160) and (B) p62 (magnification, ×160) were clearly expressed in both the cytoplasm and nucleus of neoplastic elements in CHs, while (C) AMBRA-1 was preferentially and slightly localized in the cytoplasm (magnification 160X). LC3A/B, microtubule-associated protein 1 light chain 3; p62, ubiquitin-binding protein 62; AMBRA-1, Beclin 1 regulator 1.

Utilizing Fisher's exact test revealed significant P-values for LC3A/B (0.048), AMBRA-1 (0.033), Ki-67 (0.048), and surgical treatment (0.048) (Table II).

Table II. Clinico-pathological and immunohistochemical features of autophagic proteins (p62, LC3A/B and AMBRA 1) in relation to neoplastic recurrence in chordomas.

Table II.

Clinico-pathological and immunohistochemical features of autophagic proteins (p62, LC3A/B and AMBRA 1) in relation to neoplastic recurrence in chordomas.

Feature	No recurrence	Recurrence	P-value	OR (95% CI)
Age			1.00
<60	4	0		0.500 (0.037 to 6.680)
≥60	2	4
Sex			0.076
Male	4	2		0.062 (0.001 to 1.684)
Female	2	2
p62			1.00
Low	0	0		1.444 (0.024 to 87.250)
High	6	4
LC3A/B			0.048
Low	5	0		0.030 (0.001 to 0.940)
High	1	4
AMBRA 1			0.033
Low	6	1		0.032 (0.001 to 1.044)
High	0	3
Ki67			0.048
<3%	5	0		0.030 (0.001 to 0.940)
≥3%	1	4
Volume			1.00
<13.5 cm	4	2		0.500 (0.037 to 6.680)
≥13.5 cm	2	2
Resection			0.048
Partial	1	4		33 (1.063 to 1.204)
Total	5	0

[i] OR, odds ratio; CI, confidence intervals; AMBRA 1, autophagy and Beclin 1 regulator 1; p62, ubiquitin-binding protein 62; LC3A/B, microtubule-associated protein 1 light chain 3.

Discussion

The molecular mechanisms of autophagy and the role of autophagy-targeting agents in human brain neoplasms, particularly gliomas, have been previously investigated (1,2). However, the expression of ATGs has been reported in CHs (7), in which diffuse and strong immunohistochemical expression of p62 has been highlighted, consistent with the findings of the present investigation. The significance of such prominent p62 expression may be related to blocked autophagic degradation, as suggested by Karpathiou et al (7). On the other hand, another significant ATG, LC3A/B, is considered a reliable marker of autophagy and a prognostic factor in various malignant neoplastic conditions (13–15). In CHs, LC3A/B expression has been analyzed, showing negative, mild, moderate, and strong expression based on the number of dots per cell (7); however, no relationships emerged between immunohistochemical findings and neoplastic recurrence. Similarly, AMBRA-1 has been implicated in the autophagy machinery, and its poor prognostic value has been documented in cholangio-pancreatic and prostatic adenocarcinomas, where it is associated with depth of invasion and lymph node metastasis (16,17). Nevertheless, to date, no data regarding AMBRA-1 immunoexpression in CHs have been previously described.

In the present study, we utilized the three aforementioned ATGs and employed a quantitative scoring system (0–3=negative; 4–6=positive) to explore the potential autophagic signature in CHs. Notably, p62 was highly positive in all CH cases, while LC3A/B and AMBRA-1 expressions were positive in 50 and 33% of CH cases, respectively. Interestingly, cases with elevated LC3A/B and AMBRA-1 immunoexpression were significantly associated with tumor recurrence, while p62 was evenly distributed between recurrent and non-recurrent cases. Therefore, the association between LC3A/B and AMBRA-1 allowed us to identify CH cases characterized by tumor recurrence, which also revealed an increased growth fraction (Ki-67 >3) and partial neoplastic resection. Consequently, a negative prognostic role for these two ATGs may be hypothesized in the development of CHs. Although the exact mechanism of activation in CHs remains poorly defined, a possible association between autophagic flux and brachyury activation (13,18) is a well-known molecular abnormality in CHs, similar to that reported in brain gliomas, where autophagy represents a significant phenomenon (1,2). Additionally, we demonstrated that autophagic factors, such as LC3A/B and AMBRA-1, are frequently present in CHs, associated with a strong and diffuse expression of p62, suggesting a blocked autophagic flow in contrast to their normal tissue counterparts.

In general, it is important to note that the present study had certain biases and limitations, primarily due to its retrospective nature. The main constraint was the investigation of specific ATGs exclusively through immunohistochemistry, though autophagy may represent a flux when more adequately and functionally analyzed. Furthermore, autophagy plays a crucial role in the immune tumor microenvironment and is considered significant for all tumors, including rare neoplasms such as sarcomas and CHs. However, only controversial data have been reported regarding the immune microenvironment in CHs (19–21), despite the potential for immunotherapy as a treatment option for these tumors (22). In detail, the analysis of the immune microenvironment in CHs has only recently begun, revealing that macrophages and T-lymphocytes are the most prevalent immune cells and may play critical roles in tumor immune regulation (23,24). Additionally, various cytokines and chemokines, along with other immune checkpoints such as PD-1/PD-L1, CD47/SIRPα, TIM3, and CTLA4, are expressed on the surface of both CH cells and immune cells (23,24). However, some contradictory results require further investigation and interpretation.

Further evidence characterizing the role of the immune tumor microenvironment is essential to enhance our understanding of identifying more aggressive subtypes of CHs after resection. This knowledge could inform decision-making regarding repeated resections for partial removal and adjuvant chemoradiotherapy. In conclusion, while our study represents one of the few reports on the immunohistochemical expression of ATGs in CHs, we hope that a larger series of CH tissue samples will be investigated for autophagic proteins. This investigation should compare their expression with the tumor immune microenvironment, particularly focusing on B cells, considering that PD-L1+ immune cells also express LC3A/B.

Acknowledgements

Not applicable.

Funding

Funding: No funds were received.

Availability of data and materials

The data generated in the present study may be requested from the corresponding author.

Authors' contributions

CP, AI and GT developed the study design and the manuscript draft. CP, AC, VF, MM and AG were involved in data acquisition and interpretation. AI and GT reviewed the manuscript. AG and GT confirmed the authenticity of all the raw data. All authors read and approved the final version of the manuscript.

Ethics approval and consent to participate

The analysis was conducted according to the Good Clinical Practice guidelines and the Declaration of Helsinki (1975, revised in 2013). Prior to surgical procedures, all patients provided written, anonymized and informed consent. Pathology reports and medical records were thoroughly reviewed. Patients' initials or other personal identifiers were removed from all images. The Institutional Review Board of the University Hospital of Messina (Messina, Italy) approved this study (approval no. N. 47/19; May 2, 2019).

Patient consent for publication

Written informed consent was obtained from all patients for the publication of their data.

Competing interests

No competing interests have been declared.

References