Abstract
Background/Aim: The aim of the study was to correlate the expression of mismatch repairs proteins (MMR), programmed-death-ligand1 (PDL-1), and estro-progestinic receptors (ER/PgR) in tissue samples from a series of cervical adenocarcinoma (ADC) patients with their clinicopathological features. Materials and Methods: Thirty-nine ADC specimens were retrospectively retrieved from the Division of Pathology of the University Hospital of Pisa from 2015 to 2021. Histological subtype, grade (G), Silva pattern, presence of lymph vascular space invasion (LVI), and perineural invasion (PNI) were annotated. On representative samples, immunostaining for ER/PgR, MLH1, PMS2, MSH2, MSH6, and PDL-1(sp142) was performed. Results: Thirty-five ADCs were HPV-associated usual type (24 invasive and 11 in situ), 2 were clear cell type, one was a minimal deviation adenocarcinoma (MDA), and one was an invasive stratified mucin-producing carcinoma (iSMC). ADC associated with LVI were mostly G2-3, whereas those associated also with PNI were G3 with Silva pattern C. No difference in the expression of ER/PgR was observed with a dichotomic age stratification (51 years) of patients. Only 6 ADCs were MMR-deficient, all of them were of the usual type (4 invasive and 2 in situ). The heterodimer MLH-1/PMS2 was the one most frequently altered (5/6), whereas only one case had MSH6 loss. None of ADCs express PDL-1, except iSMC which showed PDL-1 expression >1% in neoplastic cells. Conclusion: Both invasive and in situ usual type ADCs indicate MMR deficiency, highlighting how this could be an early event in tumorigenesis. None of the cases, except for iSMC, express PDL-1.
The prevalence of cervical adenocarcinoma (ADC) has increased over the past decades, reaching 20-25% of all cervical malignancies in developed countries (1-3). In Italy ADC represents 1.8% of all female malignancies with 2,700 new cases reported in 2019 (1, 2, 4).
ADC encompasses a heterogeneous group of tumours, associated with different biological behaviours and variable outcomes (5). According to the recent International Endocervical Adenocarcinoma Criteria and Classification (IECC) (5, 6), ADCs are divided in two major groups characterized by the presence or absence of human papilloma virus (HPV) infection (6, 7). HPV-related ADCs are generally associated with high-risk serotypes, such as HPV 18, 16 and 45 (5, 8). For this group, a risk stratification system has been proposed based on tumour morphology and stromal invasion –the Silva Pattern classification– able to improve the prognostic patient stratification and drive treatment decisions (9-11). This system categorizes HPV-related ADCs into three classes: Pattern A, B, and C.
The oncogenetic pathways of ADC have not been fully elucidated. Endometrioid type and endocervical “usual type” ADCs are associated preferentially with HPV 18 and HPV 16, whereas gastric type ADCs are HPV-independent (12-14). This latter, better known as minimal deviation adenocarcinoma (MDA) or adenoma malignum (15, 16), frequently harbours STK11 mutations (germline or sporadic) which represent one of its major molecular hallmarks. However, other genetic alterations have been involved, including somatic mutations of TP53, KRAS, CDKN2A, ERBB2/ERBB3, GNAS, DPC4, ARID1A, BRCA2, and PIK3CA (13, 14).
Most authors have reported that ADC has a greater tendency to lymph node, ovarian and distant spreading, and a worse prognosis compared to cervical squamous cell carcinoma (SCC), but whether the ADC histologic subtype is an independent prognostic factor is still debated (17-19).
The prognostic and predictive role of some biomarkers in ADCs, such as the mismatch repair proteins (MMR), cell surface tyrosine-kinases, and programmed death-ligand 1 (PDL-1), is still under investigation. MMR system, two redundant protein complexes that repair misincorporation errors occurring during DNA replication, gained particular interest in the last decade in several solid malignancies. Defects in this repair system can increase mutation rates and promote cancer initiation and progression, such as in patients with Lynch syndrome (20, 21). At least six genes, MSH2, MLH1, PMS2, MSH3, MSH6, and MLH3, are involved in the MMR system (22, 23). Evaluating and identifying MMR alterations has significant implications for cancer treatment and in the case of germline mutations also for family-based surveillance strategies.
PDL-1, an immune checkpoint protein normally expressed on activated immune cells, binds to regulatory surface receptors PD1 and B7 on CD8+ T-lymphocytes, thus modulating immune response (24, 25). Several cancer cells acquire the ability to express PDL-1, representing an adaptive immune resistance mechanism to reduce the host antitumor immune response. For this reason, the selective blockade of the PD1/PDL-1 axis has been considered a therapeutic strategy in different malignancies, including certain gynaecological cancers. Several studies demonstrated overexpression of PDL-1 in cervical SCC and its potential utility (26-29). On the other hand, few and contrasting data are available on ADC, due to the small number of patients included in these studies.
The aim of the investigation was to elucidate molecular aspects of ADC, evaluating the expression of PDL-1, ER-PgR, and the MMR status, towards highlighting their potential correlation with clinicopathological features
Materials and Methods
Clinicopathological data and representative samples from 39 patients were retrospectively collected from the Division of Pathology and the Division of Oncological Gynaecology of the University Hospital of Pisa from December 2015 to March 2021. The mean patient age was 49.4years (median=46.5 years; range=32-74 years). Patient treatment was managed as follows: 20 patients underwent primary radical hysterectomy, bilateral salpingo-oophorectomy, and pelvic lymphadenectomy with or without adjuvant radiotherapy; 5 underwent platinum-based neoadjuvant chemotherapy followed by radical hysterectomy, bilateral salpingo-oophorectomy, and pelvic lymphadenectomy; 5 underwent extra-fascial hysterectomy with or without bilateral salpingo-oophorectomy; 6 underwent conization alone; 3 were treated with definitive concurrent chemo-radiotherapy after cervical biopsy.
Staining. Representative samples of each case were stained with haematoxylin and eosin for standard morphologic evaluation, and histological subtyping, grade, and Silva Pattern were defined according to WHO/IECC 2020 criteria (5, 6).
Formalin-fixed and paraffin-embedded (FFPE) representative samples were stained for p16 (E6H4), ER (SP1), PgR (1E2), MSH2 (G219-1129), MSH6 (SP93), MLH1 (M1), PMS2 (A16-4), PDL-1 (SP142). All the antibodies have been acquired from Ventana Medical Systems, Inc. (Oro Valley, AZ, USA), and were “ready to use”; immunostainings have been performed on Ventana BenchMark ULTRA System, according to the vendor protocols. p16, was evaluated as positive or negative; ER and PgR were evaluated as the percentage of positive neoplastic cells. MMR proteins were evaluated according to CAP guidelines and cases labelled as MMR proficient (MMRp) or MMR deficient (MMRd). PDL-1 expression has been evaluated separately in neoplastic cells and in tumor-infiltrating immune cells and cut-off positivity was set at 1%. Biomarker expression was blindly evaluated by two pathologists (S.P. and G.N.F.); in case of discrepancies, a third pathologist (A.G.N.) reviewed the case.
Statistical analysis. Categorical variables were compared using Chi-Square and Fisher’s exact test, whereas continuous variables were compared using the Mann–Whitney U-test. Spearman’s rho test was used to assess the relationship between the biomarkers. All analyses were performed using the GraphPad Prism Software v9.0 (San Diego, CA, USA). Results were classified as statistically significant if their p-values were <0.05.
Results
In all examined patients, the most frequently observed histotype (35/39 - 91%) was the HPV-associated ADC usual type; of which 24 cases were invasive and 11 in situ (AIS). Moreover, some rare ADC subtypes were identified: 1 minimal deviation adenocarcinoma (MDA), 2 clear cell types, and 1 invasive stratified mucin-producing carcinoma (iSMC) (Table I). A mild preponderance of G1 (16 cases - 41%) were registered among all ADCs.
Histologic subtypes and grades in the examined series of ADC patients.
All ADCs associated with lymphovascular invasion (LVI) presented G2 or G3 atypia; whereas all ADCs that associated with perineural invasion (PNI) were G3 and showed a Silva Pattern C. In our cohort, we observed 9 cases with Pattern A, 4 cases with Pattern B, and 8 cases with Pattern C.
Among the 28 invasive ADCs, the Fédération Internationale de Gynécologie et d’ Obstétrique (FIGO) stage was: IA1 (n.4), IA2 (n.5), IB1 (n.7), IB2 (n.6), IB3 (n.1), IIA1 (n.2), IVB (n.1) and not defined (n.2).
All cases, but two (1 MDA and 1 AIS), were p16-positive. Most of the 39 ADCs were negative for ER/PgR: only 7 expressed both (17%), while 18/39 (40%) expressed ER with or without PgR. Moreover, the ER intensity and expression percentage were constantly higher than PgR expression (Figure 1).
Rate of ER expression compared to PgR expression in our series: the latter is generally absent or lower than the prior. PR: Progesterone receptor; ER: oestrogen receptor.
Stratifying patients according to age; younger or older than 51 years - the median menopausal age in Italy (30) - no significant differences were observed in ER/PgR expression. However, increased ADCs with LVI were found in the post-menopausal group (p=0.03). Remarkably all ADCs with LVI were negative for PgR (p=0.02) (Table II).
Correlation between LVI, and PgR status and patient age, in the examined cases.
Regarding the MMR status, 33/39 (85%) cases were MMRp. Even if no correlation was found between MMR status and clinicopathological features, interestingly the 6 MMRd ADCs were 4 invasive and 2 in situ (Figure 2). MLH-1 and PMS-2 were the most frequently absent proteins, whereas MSH-6 was absent in only one case. Two patients with MMRd ADC had metachronous malignancies (one with breast cancer, and one with melanoma, respectively).
MMR-deficiency in usual type adenocarcinoma: A) Haematoxylin & eosin; B) MSH2 and C) MSH6 show normal expression; D) MLH1 and E) PMS2 is lost in neoplastic cells. Magnification of all images (100×). MMR: Mismatch repair proteins; MSH2: MutS homolog 2; MSH6: MutS homolog 6; MLH1: mutL homolog 1; PMS2: PMS1 homolog 2.
PDL-1 was negative in 38/39 ADCs (29 cases with no expression in both compartments and 9 with <1% of tumour cell). Only one case, the iSMC, showed PDL-1 expression in >1% of tumour cells (Figure 3). Seven ADCs were associated with high-grade squamous intraepithelial lesion (HG-SIL). As an ancillary result, we have observed a positive reaction for PDL-1 in HG-SIL in 4 of these 7 cases.
Expression of PD-L1 in iSMC. A) Haematoxylin & eosin; B) Punctate PD-L1 expression in neoplastic cells. Magnification of all images 200×. PD-L1: Programmed-death-ligand1; iSMC: invasive stratified mucin-producing carcinoma.
Due to the small sample size, no correlation was found between FIGO stage and MMR status, ER/PgR, p16, and PDL-1 expression.
Discussion
It is well-known that treatment of cervical cancer is largely based on the FIGO staging system at the time of presentation, regardless of the histotype. However, treatment is based on studies that include essentially SCCs and does not take into account several morphological parameters, such as the Silva Pattern classification system and the molecular background.
Our cohort included only ADCs, which could represent a limitation due to the small sample size. However, in our opinion, it could represent the correct approach for the investigation of this particular subgroup of cervical cancer.
Most ADCs of our cohort were HPV-related, in agreement with existing literature (5). Among the several morphological aspects evaluated, our results remarkably highlight an inverse correlation between LVI and PgR status. This relationship seems even stronger in the group of post-menopausal patients.
Bonneville et al. found that only 2.6% of patients with cervical cancer harbour MMRd, but they did not analyse separately SCC from ADC (31). However, in our cohort, which includes only ADC, 15% of cases were MMRd. The significantly higher percentage of MMRd cases could suggest a pivotal role of this dysregulation in ADC carcinogenesis. Moreover, either invasive or in situ lesions can host this molecular alteration, supporting this assumption. Importantly, distinguishing MMRd due to germline mutations or somatic mutations remains crucial, particularly because of the increased lifetime risk of several solid malignancies in patients harbouring a germline mutation. Indeed, in this subset, we observed two patients with multiple metachronous malignancies: one with melanoma and one with breast cancer. This underpins the importance of screening for Lynch syndrome also in ADC, confirming the importance of genetic counselling.
Nowadays, cancer immunotherapy represents an innovative and game-changing treatment approach for several solid malignancies. Among gynaecological malignancies, particularly in endometrial cancer, other authors have observed a higher presence of tumour-infiltrating lymphocytes (CD8+) and PDL-1 expression in MMRd compared to MMRp tumour. For this reason, it could be of interest to evaluate this same relationship also in ADC; however, so far, the relevance of PDL-1 protein expression in cervical cancers has been investigated only in a few studies, generally referred to SCC with different and contrasting results (25, 32-36). Our results showed a substantial absence of PDL-1 expression (<1%) in all cases, independent of MMR status or other clinicpathological variables. The only exception was the iSMC, which has been recently described as a peculiar morphological histotype with evident HPV infection-related features, apical mitotic figures, and apoptotic bodies, and frequently associated with a brisk tumour inflammatory infiltrate (37).
In conclusion, we found a useful and easily detectable predictive marker of vascular involvement and node spreading in the PgR status. Moreover, according to our data MMRd could represent an early event in ADC tumorigenesis, suggesting the possibility of the investigation of novel therapeutic strategies. On the other hand, PDL-1 seems not to play a pivotal role in ADC, except for iSMC. Though our results need to be validated in larger multi-institutional cohorts, this study is a foundation for further analyses underpinning the heterogeneity of this malignancy and the still under-investigated molecular features.
Footnotes
Authors’ Contributions
Conceptualization, S.P. and A.G.; methodology, G.N.F, F.G. and F.G.; data curation, F.G., F.G., S.C., and A.G.N.; writing – original draft preparation, S.P., G.N.F. and A.G.N.; visualization, S.P., and G.N.F.: writing – review and editing, S.P., G.N.F., and A.G.; supervision, A.G. All Authors have read and agreed to the published version of the manuscript.
Conflicts of Interest
The Authors declare no conflicts of interest.
- Received October 26, 2022.
- Revision received November 22, 2022.
- Accepted November 28, 2022.
- Copyright © 2023 International Institute of Anticancer Research (Dr. George J. Delinasios), All rights reserved.
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.