Diagnostically misleading aberrant terminal deoxynucleotidyl transferase expression in germ cell tumors

Introduction

Terminal deoxynucleotidyl transferase (TdT) is a specialized, template-independent DNA polymerase catalyzing the addition of new nucleotides to the 3´ terminus of the DNA molecule [1]. It participates in junctional diversity in T-cell receptors and immunoglobulins since it adds random N-nucleotides to the VDJ exons. This process enables the production of a plethora of variants of immune receptors reacting with multiple antigens. It is normally expressed by immature pre-B and pre-T lymphocytes inhabiting lymphatic organs such as the thymus, lymph nodes, tonsils, and bone marrow. Nevertheless, TdT is an important diagnostic marker in hematopathology, since the immunohistochemical assessment of TdT expression is employed to detect the presence of immature lymphoid cells of acute lymphoblastic leukemia (ALL) [2, 3].
This marker is generally considered highly specific for ALL, but occasionally may be detected in quite diverse, unrelated groups of malignancies. Some nonlymphoid, poorly differentiated cancers, such as small cell lung carcinoma and Merkel cell carcinoma, may infrequently express TdT [4].
Recently a few studies have demonstrated TdT expression in seminoma and other germ cell tumors (GCTs). This phenomenon represents a potential diagnostic pitfall, especially in GCTs arising in the mediastinum, a site that is also commonly affected by ALL.
Nonetheless, its association with clinicopathological parameters has never been investigated. The current study aimed to assess the expression of TdT in a large cohort of testicular GCTs and a small subset of ovarian dysgerminomas and gonadoblastoma and compare it with other parameters.

Material and methods

Study group

The study enrolled 173 cases of testicular GCTs, 5 ovarian dysgerminomas, and one gonadoblastoma diagnosed between 2010 and 2020 with available formalin-fixed, paraffin-embedded tissue blocks. Considering each histology of testicular GCTs, the analyzed group consisted of 116 seminomas, 70 embryonal carcinomas, 31 post-pubertal yolk-sac tumors, 47 post-pubertal teratomas, 20 choriocarcinomas, 5 spermatocytic tumors, and 5 pre-pubertal yolk sac tumors. Basic clinicopathological parameters (age, TNM, tumor size, histopathological report) were extracted from patients’ medical records. Patients’ data were anonymized before the analyses.

Immunohistochemistry

Subsequently, tissue microarrays were constructed with Manual Tissue Arrayer (MTA-1) (Beecher Instruments, Inc., Sun Prairie, WI, USA) using 1.5 mm core needles. Representative tumor areas including each component of mixed tumors were sampled. The median number of cores was 3 (range 2–10). Subsequently, 5 µm sections were cut from each TMA and stained using anti-TdT monoclonal rabbit antibody EP266 (Dako) and polyclonal rabbit antibody (Cell Marque).
Stained slides were analyzed by two pathologists (MK and RP). Only nuclear staining of TdT was considered positive expression. The stainings were evaluated semiquantitatively utilizing an H-score calculated by a combined evaluation of staining intensity (0–3) and the percentage of positive cells (0–100). The following formula was used: 3  percentage of strongly stained nuclei + 2  percentage of moderately stained nuclei + 1  percentage of weakly stained nuclei; thus the final H-score was in the range 0–300. In the case of mixed tumors, each component was assessed separately.

Statistics

Basic study data were evaluated using descriptive statistical methods (frequency, median, interquartile ranges). Expression of TdT between groups was compared with Mann-Whitney U test or Kruskal-Wallis one-way analysis of variance when applicable. The correlation of TdT with contiguous variables was assessed by Spearman’s rank correlation coefficient. The threshold for statistical significance was set at p-values less than < 0.05. Statistical analysis was performed with Statistica 13.3 software (TIBCO Software Inc., Palo Alto, CA, USA) and R statistical environment. Boxplots and scatterplots were created using the ggplot2 package.

Ethics

The study was approved by the Bioethical Committee of Medical University of Gdańsk (approval No. NKBBN/485/2019).
This research was funded by internal grants from the Medical University of Gdańsk (02-0095/07/267; 01-0395/08/643).
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of Medical University of Gdańsk (NKBBN/485/2019). Patient consent was waived due to the retrospective nature of the study.
The data analyzed during the current study are available from the corresponding author on reasonable request.

Results

No specific nuclear reaction was found with the use of a polyclonal anti-TdT antibody in any case, but the cytoplasmatic reaction was commonly observed. In case of the EP266 antibody, TdT H-score values varied between main histological subtypes. Any expression of TdT (H-score > 0) was found in 155 cases (89.59%) of testicular GCTs. Positive nuclear staining was consistently seen in pure seminomas and seminoma components of mixed GCTs, with 113/116 (97.4%) displaying a positive reaction (median H-score: 170, IQR: 100–240). Representative examples of various TdT staining intensities in seminoma are presented in Figure 1.
Similar observations were made for embryonal carcinoma (60/70 positive cases, 85.7%), but the immunohistochemical reaction tended to be weaker with a median H-score of 70 (IQR: 20–110). Other testicular tumors displayed very weak positive staining in single cases – yolk sac tumor in 4/31 cases (12.9%), teratoma in 2/47 cases (4.2%), and choriocarcinomas in 1/20 cases (0.5%). Importantly, TdT expression was consistently expressed by germ cell neoplasia in situ (GCNIS).
In all cases, histologically normal elements, i.e. healthy epithelium of seminiferous tubules, Leydig cells, and Sertoli cells, were TdT-negative. Similarly, no reaction was noted in tumor-associated immune cells.
Tumors not associated with GCNIS, i.e. spermatocytic tumor and prepubertal yolk sac tumor, did not express TdT. The ovarian counterpart of seminoma, dysgerminoma, expressed TdT in all cases (5/5). In the single analyzed case of gonadoblastoma, strong TdT expression was noted in nests of dysgerminoma-like germ cells. Representative examples of TdT staining in selected nonseminomatous samples are shown in Figure 2.
Most embryonal carcinomas expressed TdT at low or intermediate levels, whereas seminomas had a wide distribution of TdT expression but tended to have high-level expression (Fig. 3). In general, seminomas demonstrated higher TdT expression compared to embryonal carcinoma. On the other hand, both these tumors had greater TdT expression than other types of testicular GCTs (Fig. 3).
A comparison of TdT levels in seminoma and embryonal carcinoma concerning TNM characteristics is presented in Table I. Expression of TdT in seminoma was not associated with any clinicopathological variables except tumor size, with which it was weakly negatively correlated (p = 0.0004, Spearman’s rho = –0.2654, Fig. 4A).
Both seminomas and embryonal carcinomas tend to express higher TdT levels in their pure forms than in their respective components of mixed GCTs (Fig. 5A). On the other hand, TdT expression in embryonal carcinoma is higher in tumors metastatic to lymph nodes (p = 0.014, Fig. 5B). Moreover, TdT expression in embryonal carcinoma shows a weak positive correlation with age (p = 0.009, Spearman’s rho = 0.3086, Fig. 4B).

Discussion

The current study confirmed the common expression of TdT in some types of GCTs, especially in seminoma, dysgerminoma, and embryonal carcinoma. Moreover, for the first time, the association between elevated TdT expression and nodal metastases in embryonal carcinoma was demonstrated. The predominant expression was observed in seminoma and embryonal carcinoma, two types of GCT directly originating from pluripotent GCNIS cells. On the other hand, GCT types believed to develop as the result of differentiation of embryonal carcinoma cells, i.e. choriocarcinoma, yolk sac tumor, and teratoma, were uniformly negative or only weakly positive for TdT. It suggests the potential role of TdT in tumors that do not differentiate into extra-embryonic elements, and precursor lesions (GCNIS).
To date, several studies have investigated TdT expression in testicular cancer. The first report of a positive immunohistochemical reaction for TdT with EP266 antibody in classical seminoma was published in 2017 [5]. The authors observed TdT expression in all 10 analyzed cases. They excluded the possibility of false-positive results due to the tissue processing, as seminoma cells retained the TdT immunoreactivity independent of previous cryosectioning or direct formalin fixation. In contrast, aberrant cytoplasmic TdT staining may occur in previously frozen, formalin-fixed paraffin-embedded samples of lymphoblastic lymphoma [6]. A subsequent study by Gayhart et al. utilizing the same clone revealed TdT expression in 4/25 seminomas (16%), 6/12 germinomas (50%), and 1/3 dysgerminomas (33%) [7]. Another study on a larger population (n = 192) reported frequency of TdT expression very similar to our cohort [8]. None of the aforementioned studies analyzed the association between the level of immunohistochemical expression of TdT and clinicopathological characteristics. One study postulated that long-term storage of paraffin-embedded GCT tissues may lead to losing their TdT detectability [8]. In contrast, we observed that deeper parts of the tumor show lower (if any) reactions than superficial areas, where the formalin penetration is more efficient. Accordingly, larger tumors tend to have lower TdT levels. Therefore, in our opinion, the effectiveness of formalin fixation is a crucial determinant of TdT detectability in germ cell tumors.
Jaconi et al. aimed to comprehensively investigate the origin of TdT positivity in GCTs [9]. In a cohort of 22 GCTs, predominantly seminomas, TdT expression was commonly detected with EP266 and polyclonal rabbit antibody, but the latter tended to produce weaker staining. However, this study failed to demonstrate TdT protein presence in seminoma with Western blot and liquid chromatography/mass spectrometry, which may suggest that immunohistochemical expression of TdT is a result of cross-reaction with a 45– 50 kDa protein homologous to TdT, e.g. polMu [9]. Nevertheless, non-neoplastic testicular tissue is one of the few sites with reported DNTT mRNA expression in healthy organs, together with thymus, tonsil, and bone marrow according to the Human Protein Atlas consensus dataset [10]. It is also possible that truncated TdT of lower molecular weight is produced in GCTs, but the available UniProt data do not support this hypothesis, as emphasized by Jaconi et al. [9]. One of the most interesting findings in our study was the association between higher TdT H-scores and the presence of nodal metastases in embryonal carcinoma. It may suggest that TdT or a homologous protein detectable by immunohistochemistry may play a role in tumor progression. Additionally, we observed higher expression of TdT in pure seminomas and embryonal carcinomas when compared to their counterparts in mixed GCTs. In our cohort, the frequency of nodal metastases in pure embryonal carcinomas was higher than in mixed tumors; hence the aforementioned effect is most likely caused by this correlation. Gene expression microarray data suggest that pure and mixed types of GCTs are molecularly different [11], which is supported by the current study. Additionally, we observed TdT expression in GCNIS and the two most common invasive GCTs originating from GCNIS (seminoma, embryonal carcinoma). In contrast, GCTs not associated with GCNIS, i.e. spermatocytic tumor and prepubertal yolk sac tumor, were negative in all cases.
Due to the very limited number of events in our cohort we were unable to investigate the prognostic impact of TdT expression in GCTs. Of note, a recent study demonstrated that TdT expression is an independent prognostic factor in Merkel cell carcinoma (MCC), but it is associated with favorable cancer-specific survival [12]. In the case of MCC, the aberrant TdT expression is presumably a consequence of Merkel polyomavirus infection which alters affected stem cells and induces expression of pre-and pro-B-cell markers along with neuroendocrine and epithelial markers [13]. It is unknown whether TdT is directly involved in any biological processes in MCC, but it may participate in DNA repair, enhancing the survival of malignant cells [14]. Its role in the biology of GCTs remains even more elusive.
In our everyday pathological practice, we observed TdT expression in GCTs for the first time in a case of mediastinal GCT morphologically mimicking lymphoblastic lymphoma, and other authors had similar experiences [5]. However, lymphoblastic leukemia/lymphoma may also involve the testis, and thus pathologists should be very careful in the interpretation of TdT staining. Perplexingly, TdT was also detected with polyclonal rabbit antibody in a subset of pediatric small round cell tumors, i.e. medulloblastoma, Ewing sarcoma, and rhabdomyosarcoma [15]. Moreover, reactive, perivascular aggregates of TdT-positive lymphocytes were observed in some cases. It represents one more potential diagnostic pitfall in the differential diagnosis of pediatric acute lymphoblastic leukemia. Occasionally, a positive reaction with polyclonal anti-TdT antibody may be detected in pulmonary small cell carcinoma [4]. Recently, a positive TdT reaction obtained with EP266 antibody was demonstrated in normal and neoplastic sebaceous and myoepithelial cells. In the former, this phenomenon may occur due to the differentiation programming prior to apoptosis initiated by the holocrine secretion mode [16]. On the other hand, no DNTT mRNA was detected in myoepithelial cells [17]; thus, cross-reactivity with another antigen should be considered (similarly to GCT).

Conclusions

Immunohistochemical staining of TdT is frequently detected in seminoma and embryonal carcinoma cells. The inverse correlation between tumor size and TdT levels may be related to the poorer fixation of larger tumors. The higher expression of TdT in embryonal carcinoma components of GCTs metastatic to lymph nodes may suggest its role in testicular cancer biology. Finally, our study confirms that aberrant expression of TdT by testicular and ovarian GCTs exemplifies a potential diagnostic pitfall in histopathological diagnostics. The authors declare no conflict of interest.

References