Introduction: Studies on pancreatic neuroendocrine tumors (PanNETs) regarding loss of ATRX, DAXX, or frequency of microsatellite instability (MSI) show inconclusive results. So far, data on corresponding metastaseshave not been published. Methods: We performed immunohistochemistry (IHC) of ATRX, DAXX, MSH2, MSH6, MLH1, and PMS2 on 74 PanNETs and 19 metastases. ATRX- and DAXX-negative PanNETs were further sequenced for mutations. We used polymerase chain reaction for MSI on cases with IHC loss of MSH2, MSH6, MLH1, and PMS2. Results: Immunohistochemical loss of DAXX and ATRX was observed in 8/74 (11%) and 6/74 (8%) PanNETs. Loss of DAXX immunoreactivity was statistically associated with higher tumor grade and showed a tendency toward a decreased overall survival. Sequencing of DAXX- (7/11 [64%]) and ATRX-negative (5/11 [45%]) PanNETs revealed a mutation in 6/7 (86%) and 2/5 (40%). The specificity of immunohistochemical loss of DAXX and ATRX for mutation was 80% and 67%, respectively. The expression status of DAXX compared to primary tumor differs in 2/12 (17%) lymph node metastases. We further identified 3/74 (4%) tumors as MSI, associated with a poor prognosis. Discussion/Conclusion: Our study supports the hypothesis that a loss of DAXX immunoreactivity can identify a more aggressive subtype of PanNET with high confidence, while ATRX loss is a weaker indicator. Our results also strengthen the role of DAXX immunolabeling as a prognostic marker. We could show that ATRX might be less suitable as a surrogate for sequencing. Our results indicate that IHC of DAXX and ATRX may identify PanNET subtypes as targets for more aggressive therapy.

Due to the rarity and heterogeneity of pancreatic neuroendocrine tumors (PanNETs), it is challenging to identify and classify prognostic factors [1‒5]. So far, the prognosis of these tumors largely depends on clinical and histopathological factors, including age, tumor size, mitotic rate, tumor grade, and stage or complete surgical resection [6‒8]. To evaluate the prognosis, treatment, and classification of PanNETs, the WHO introduced a histological classification system of pancreatic neuroendocrine neoplasia (PanNEN) in 2010 [9‒11]. In 2017, there was an update on the WHO classification system dividing PanNENs into well-differentiated PanNEN and poorly differentiated PanNEN. Well-differentiated PanNENs include PanNET G1, PanNET G2, and PanNET G3. Poorly differentiated PanNENs are neuroendocrine carcinomas (NEC), either small cell NEC or large cell NEC [9‒14].

Through advances in sequencing technologies, the molecular heterogeneity of PanNETs, e.g., in somatic mutations, variations in copy numbers and DNA methylation changes, the alternative length of telomeres including the role of the ATRX and DAXX genes in these processes have been investigated [15‒22]. ATRX, a transcription regulator, and DAXX, a histone H3.3 chaperon, are chromatin remodelers that interact to maintain the integrity of telomeres and genome stability [17, 20, 23]. Recent studies have shown that mutations of ATRX and DAXX are most commonly found in PanNETs [2, 24, 25]. Investigations of Jiao et al. [18] identified 17.6% of ATRX and 25% of DAXX mutations in PanNETs.

It is well known that the integrity of genetic information depends on the fidelity of DNA replication and the efficiency of several different DNA repair processes. One repair system in particular, the DNA mismatch repair (MMR) system, is responsible for correcting base substitution mismatches and insertion-deletion mismatches. If these processes are disturbed, e.g., by a mutation of the DNA MMR system, a hypermutated phenotype occurs and a tumor may develop or progress faster [26‒28]. The MMR mechanism, therefore, has been well investigated in recent years and it is known that colorectal adenocarcinoma has a high prevalence for mutations of the MMR system [27, 29‒31]. Furthermore, recent studies suggest a prognostic and a pathogenetic value of microsatellite instability (MSI) in PanNENs. Nevertheless, the percentages of MMR deficiency in PanNENs are inconsistent and differ between 0% and 36% [32‒37]. In this study, we evaluated ATRX and DAXX and microsatellite status in primary tumor tissue and metastases of 74 patients with primary PanNETs.

Patients and Tissue

Samples were obtained from 74 patients with sporadic primary PanNETs undergoing surgical resection at the Department of General Surgery and Visceral Surgery at the St. Josef-Hospital – Katholisches Klinikum Bochum between 2010 and 2016. 60/74 (81%) patients had a sporadic primary PanNET without metastases and 14/74 (19%) patients had metastases. 71/74 (96%) were nonfunctional PanNETs and 3/74 (4%) were functional PanNETs, identified as insulinomas. An association to a hereditary disease did not appear. We investigated formalin-fixed and paraffin-embedded (FFPE) tissue from 74 primary tumor samples of PanNETs and tissue samples from 19 metastases, including 12 lymph node metastases and 7 liver metastases. Of the 14 patients with metastases, 5 cases showed both lymph node as well as liver metastases.

The patients’ clinical and general practitioner records retrieved clinical characteristics such as age, sex, survival status, and tumor localization. All tumors were categorized after reclassification according to recent WHO 2019 and the current TNM system (see online suppl. Table 1; for all online suppl. material, see www.karger.com/doi/10.1159/000524920). The neoplastic cell nests are cuboidal and eosinophilic, with finely granular cytoplasm. Between the neoplastic cell nests, there are areas of dense, hyalinized collagen. An immunohistochemical expression of synaptophysin and chromogranin A was analyzed to verify a neuroendocrine differentiation of the neoplastic cells. The three-tier grading system, including low (G1), intermediate (G2), or high grade (G3), was based on the Ki-67 labeling index. PanNET is graded as G1 (Ki-67 labeling index <3%), G2 (Ki-67 labeling index 3–20%), and G3 (Ki-67 labeling index >20%). The Institutional Review Board approved the study (REC reference number 15-5431).

Immunohistochemistry of MMR Proteins, DAXX, and ATRX

For the immunohistochemical staining, 1-μm paraffin-embedded sections were taken by the medical-technical assistant with the microtome (Thermo Scientific Sliding Microtome, Microm HM 430; Walldorf, Germany). After deparaffinization and rehydration, the immunohistochemical staining for all antigens was performed on an automated staining system (Leica Biosystems BOND-III; Wetzlar, Germany). The primary antibodies, host, dilution, incubation, clone, and antigen retrieval solution are presented in online supplementary Table 2. Visualization was performed using an alkaline phosphatase-linker antibody conjugation system (Bond Polymer Refine Red Detection Kit of Leica Biosystems, Wetzlar, Germany [Catalog No.: DS 9390]), which yielded a purple-red staining signal. For all scorings, only nuclear protein staining was considered positive. To exclude false-negative samples, an immunohistochemical control was performed on non-neoplastic cells. Tissue from the appendix vermifomis with a known intact expression of MLH1, MSH2, MSH6, and PMS2, as well as tissue of the testicle with a known intact expression of DAXX and tissue of the brain with known intact expression for ATRX served as a positive control.

Positive nuclear staining of ATRX and DAXX tumor cells (Fig. 1, 2) was considered wild type, whereas a complete loss of nuclear staining was considered a surrogate marker for a pathogenic mutation of the tumor (Fig. 3, 4). Lack of MMR proteins was defined for tumor samples with more than 90% of nuclear expression loss for at least one of MLH1, MSH2, MSH6, or PMS2. Of those potential MMR-deficient tumor samples, MSI status was analyzed by fragment length analysis afterward.

Fig. 1.

Retained expression of ATRXIHC in primary, low-grade PanNET. ATRXis presented with strong nuclear staining within the tumor cells.

Fig. 1.

Retained expression of ATRXIHC in primary, low-grade PanNET. ATRXis presented with strong nuclear staining within the tumor cells.

Close modal
Fig. 2.

Retained expression of DAXXIHC in primary, low-grade PanNET. DAXXis presented with strong nuclear staining within the tumor cells.

Fig. 2.

Retained expression of DAXXIHC in primary, low-grade PanNET. DAXXis presented with strong nuclear staining within the tumor cells.

Close modal
Fig. 3.

Loss of IHC expression of ATRXin primary, intermediate-grade PanNET. There is no nuclear expression of ATRXin tumor cells compared to the positive internal control. The positive internal control of ATRXis seen as positive staining in the fibroblasts and lymphocytes.

Fig. 3.

Loss of IHC expression of ATRXin primary, intermediate-grade PanNET. There is no nuclear expression of ATRXin tumor cells compared to the positive internal control. The positive internal control of ATRXis seen as positive staining in the fibroblasts and lymphocytes.

Close modal
Fig. 4.

Loss of IHC expression of DAXXin primary, intermediate-grade PanNET. There is no nuclear expression of DAXXin tumor cells compared to the positive internal control.

Fig. 4.

Loss of IHC expression of DAXXin primary, intermediate-grade PanNET. There is no nuclear expression of DAXXin tumor cells compared to the positive internal control.

Close modal

DNA Extraction from Paraffin-Embedded Tissue

Tumor samples with immunohistochemical loss of nuclear expression for at least one of the four MMR proteins were taken for further DNA analysis of MSI. Therefore, 10-μm-thick paraffin sections of NET and tumor-free tissue of the same patient were microdissected with a sterile scalpel. The tumor-free margins of the sections of the NETs were separated grossly. NET and tumor-free tissue sections were digested overnight at 70°C in an incubation buffered proteinase K solution. The DNA extraction was performed with the Maxwell RSC (Promega GmbH) within 30 min on the next day. It is an automated nucleic acid purification platform which uses prefilled cartridges and preprogrammed methods. The concentration was measured by TECAN Infinite 200 Pro NanoQuant Plate Reader (MSI) and by QuantiFluor® ONE dsDNA on the Quantus (Promega Corporation, Madison, WI, USA) (next-generation sequencing [NGS]).

Real-Time Polymerase Chain Reaction Amplification and MSI Analysis

Purified DNA was amplified using the Biometra Professional Thermocycler (Analytik Jena AG 2019). For MSI determination, the DNA was amplified by PCR targeting five monomorphic mononucleotide markers (see online suppl. Table 3). Tumors showing a shift in at least two markers were classified as MSI. An additional panel was used to identify another allelic shift if there was only one shift. PCR products were analyzed with the Beckmann GenomeLabTM GeXP Genetic Analysis System including determination of the length of the PCR products and the height of the peaks.

Next-Generation Sequencing

PanNETs with either a single loss of one protein or a combined loss of ATRX and DAXX were taken further for NGS. As described above, the DNA concentration was measured by QuantiFluor® ONE dsDNA on the Quantus (Promega Corporation). All further concentration measurements are performed using QubitTM dsDNA HS. A QIAseq Targeted DNA Costume Panel (QIAGEN GmbH, Hilden, Germany) was designed to amplify the coding regions of ATRX and DAXX. According to the manufacturer’s instructions, libraries were prepared with 50 ng of genomic DNA. Sequencing of the final libraries was performed using NextSeq 550 Illumina sequencer (Illumina Inc., San Diego, CA, USA). A QIAseq Targeted DNA Panel Workflow in QIAGEN CLC Genomics Workbench (Version 21) was used for data analysis. Analysis and assessment of detected variants was performed using IGV, MutationTaster, ClinVar, and Cosmic. Only disease-causing mutations with a VAF >5% were reported.

Statistical Analysis

Univariate and multivariate statistical analyses obtained correlation with clinicopathological characteristics. For all multivariate analyses, the data were calculated with the Spearman. For all multivariate analyses, the data were calculated with the Spearman correlation using the JMP statistical software (software from SAS; Statistical Discovery SAS Institute Inc.). The significance level for univariate analysis was also defined by p < 0.05, using the JMP statistical software. Odds ratio was used to calculate the strength of a link between ATRX-negative/-intact tumors or DAXX-negative/-intact tumors compared to the risk of dying.

We analyzed tumor samples of 74 patients with sporadic primary PanNETs and their metastases. 71/74 (96%) were nonfunctional PanNETs and 3/74 (4%) were functional insulinomas without an association to a hereditary disease. Tumor grade and stage were classified according to WHO 2019 and recent TNM classification (see online suppl. Table 1).

Alteration of ATRX in Primary PanNET

IHC determined loss of ATRX and DAXX in all 74 primary tumors (Fig. 3, 3, 4) and in 14 patients with 19 metastases. Those 14 cases showed 12 lymph node metastases and 7 liver metastases, 5 cases had a combination of both. In total, 19 metastases were examined. We observed loss of ATRX in primary tumor tissue in 6/74 (8%), in 3/12 (25%) lymph node metastases, and in 1/7 (14%) liver metastases. ATRX-negative tumors were not statistically significantly associated but showed a trend toward an association with a high tumor stage and grade. The ATRX-negative tumors tend to have an intermediate grade (5/6 cases; 84%), a larger tumor size (>4 cm), extrapancreatic tumor spread (5/6 cases; 84%), and two cases (2/6 cases; 33%) were stage IV at primary diagnosis (Table 1). All PanNETs with loss of ATRX expression were located in the pancreatic tail and body. None was seen in the pancreatic head (Table 1). In addition, there was a statistically balanced gender ratio (Table 1). Concerning the survival status, a tendency toward a low survival was seen with an almost two times higher risk to die compared to patients with intact ATRX tumors and low survival status.

Table 1.

Results of MSI, loss of DAXX and ATRX in correlation to clinical characteristics in primary PanNET and metastases

 Results of MSI, loss of DAXX and ATRX in correlation to clinical characteristics in primary PanNET and metastases
 Results of MSI, loss of DAXX and ATRX in correlation to clinical characteristics in primary PanNET and metastases

Alteration of DAXX in Primary PanNET

8/74 (11%) primary tumors showed a loss of DAXX. Loss of DAXX occurred in 6/12 (50%) lymph node metastases and in 2/7 (28%) liver metastases. Tumors showing DAXX immunoreactivity loss were statistically significant (p value = 0.0041) associated with a high tumor stage and grade. 6/8 (75%) had an intermediate grade, 4/8 (50%) had a higher tumor size (>4 cm) or extrapancreatic tumor spread, 6/8 (75%) had lymph node metastases, and 3/8 (38%) had liver metastases (Table 1). All DAXX-negative PanNETs were located in the pancreas’ tail and body, and a statistically balanced gender ratio was seen (Table 1). Furthermore, DAXX-negative tumors showed a tendency toward a shorter survival with an almost two times higher risk of dying than patients with intact DAXX tumors.

Combined Alteration of ATRX and DAXX in PanNET and Comparison of Primary PanNET with Lymph Node and Liver Metastases

2/74 (3%) cases of primary tumors and 2/12 (17%) lymph node metastases as well as 1/7 (14%) liver metastases showed a combined loss of ATRX and DAXX. Regarding the ATRX and DAXX expression status, 12/14 (80%) cases were equal to the primary PanNET tissue. We observed heterogeneity between primary tumors and lymph node metastases concerning the loss of DAXX and ATRX metastases in 2/12 (17%) cases. These two cases had an additional loss of DAXX. The liver metastases of every case showed the same ATRX and DAXX expression status as the primary tumor. As previously mentioned, primary PanNETs, even with a combined loss of ATRX and DAXX expression, were located in the pancreatic tail and body (Table 1).

NGS of ATRX and DAXX in PanNET

NGS analysis revealed the presence of somatic mutations of DAXX and ATRX in 6/11 (55%) and 2/11 (18%) PanNETs (see online suppl. Table 4). No tumor with a mutation in DAXX had a mutation in ATRX, consistent with their presumptive function within the same pathway. Overall, 8/11 (72%) PanNETs had a mutation in this pathway. One deletion resulting in a disease-causing frameshift mutation and one transversion causing a pathogenic-splicing mutation was detected in ATRX. Four deletions resulting in a frameshift mutation and two nonsense mutations in DAXX, each producing a stop codon, were seen (see online suppl. Table 5).

Comparison of ATRX-/DAXX-IHC and NGS Results

We observed a loss of protein expression of either ATRX or DAXX 14 times. 6/74 (8%) primary PanNET showed a loss of ATRX, and 8/74 (11%) primary PanNET showed a loss of DAXX in IHC. Of those cases, 2/74 (3%) showed a combined loss of ATRX and DAXX resulting in a total of 12/74 (16%) cases with either a single loss of one protein or a combined loss. Those 12/74 (16%) PanNET were taken further for NGS. One case with a combined loss of ATRX and DAXX needed to be excluded from NGS because it did not pass all the quality criteria, resulting in 11/74 (15%) cases to be examined. Of the 22 proteins examined by NGS (11/22 [50%] ATRX and 11/22 [50%] DAXX), 18/22 (82%) showed a result consistent with the IHC observations.

Of the 5/11 (45%) cases with IHC loss for ATRX, 2/5 (40%) revealed a mutation of ATRX. The specificity and sensitivity of ATRX loss for ATRX mutation was 67% and 100%, respectively.

The two-thirds (7/11 [64%]) of the PanNETs with a loss of DAXX immunoreactivity harbored 6/11 (55%) mutated DAXX. The single case with loss of DAXX expression did not reveal any mutation. The DAXX staining on this case was technically not very satisfactory; positive DAXX labeling was rarely observed in internal (lymphocytes) controls and well observed in external controls. The specificity and sensitivity of DAXX loss for DAXX mutations was 80% and 100%. All PanNETs with a retained DAXX or ATRX expression harbored no DAXX or ATRX mutations, respectively (see online suppl. Table 4).

MMR Deficiency and MSI

Possible MMR deficiency has been evaluated in 4/74 (5%) primary PanNETs. A potential MMR deficiency was defined immunohistochemically by a loss of nuclear expression of at least 90%. These specimens were taken further for PCR-MSI analysis using five monomorphic mononucleotide markers (Table 2). We found MSI in 3 out of 74 samples (4%) (Tables 1, 2). The discrepant case, identified as “stable” by PCR, was immunohistochemically negative for MSH6. MSI status was identical in primary tumor and corresponding metastases.

Table 2.

Results of IHC, FLA, and MSI analysis of the four MMR proteins (MLH1, PMS2, MSH6, MSH2) in primary PanNET

 Results of IHC, FLA, and MSI analysis of the four MMR proteins (MLH1, PMS2, MSH6, MSH2) in primary PanNET
 Results of IHC, FLA, and MSI analysis of the four MMR proteins (MLH1, PMS2, MSH6, MSH2) in primary PanNET

Two MSI cases were UICC tumor stage IV, which correlated with worse prognosis and short survival. The remaining case was diagnosed as UICC tumor stage I disease. However, the patient died cancer-related within 5 years. None of the three (3/74) functional PanNETs showed MMR deficiency/MSI or loss of ATRX or DAXX.

The scientific interest has moved to the genetic properties of PanNETs because recent population-based studies have shown a 3- to 5-fold increase of the appearance of PanNENs [2‒5, 38]. Some studies identified mutations of ATRX and DAXX in PanNETs as new prognostic markers [2, 24, 25, 39]. The IHC can be used to visualize an inactivating mutation of ATRX or DAXX in PanNETs [16]. Hechtman et al. [40] recently published sensitivity and specificity of IHC for DAXX mutation of 85% and 95%, respectively, in PanNETs. Based on our IHC staining, we examined 12/74 (16%) with a loss of at least ATRX (6/74; 8%) or DAXX (8/74; 11%). Two cases showed a combined loss. Jiao et al. [18] proclaimed a higher mutation rate of 17.6% for ATRX and 25% for DAXX. No tumor showed a mutation in both [18]. Marinoni et al. [41] showed similar percentages but identified four cases with a mutation in both. Our results have more in common with the studies of Singhi et al. [42], who identified a loss of nuclear expression for DAXX, ATRX, or both in 12%, 9%, and 5%. The common hypothesis is that a loss of ATRX and DAXX is involved in tumor development and progression, leading to a more aggressive subtype of PanNETs. However, Jiao et al. [18] proclaimed that a loss of ATRX and DAXX was significantly associated with longer survival after recurrences and longer overall survival of metastatic PanNETs [43]. On the contrary, Marinoni et al. [41] showed a correlation of ATRX- and DAXX-intact PanNETs with longer disease-free surveillance [43]. Furthermore, Jiao et al. [18] did not mention which type of tumor characteristics predominate in his study. We therefore cannot thoroughly compare the results. More recent studies examined the characteristics of ATRX- or DAXX-negative PanNETs and showed a correlation with a larger tumor size of >2–3 cm, an intermediate grade, extrapancreatic tumor spread, metastases, a reduced time of relapse-free survival, and a decreased time of tumor-associated survival as well as an alternative length of telomeres [16, 20, 21, 41, 42, 44, 45]. This is consistent with our results showing that ATRX- and DAXX-negative tumors are associated with an intermediate grade (ATRX: 84%, DAXX: 75%), a larger tumor size (>4 cm), and extrapancreatic tumor spread (ATRX: 84%, DAXX: 50%).

In conclusion, we showed that our cases with loss of DAXX were statistically significantly (p value = 0.0041) associated with a high tumor stage and grade. ATRX-negative tumors were not significantly associated but appear to correlate with it. Further, DAXX- and ATRX-negative tumors correlated with low survival, which is consistent with former studies that proclaimed a reduced survival of patients with ATRX- and/or DAXX-negative tumors [20, 39, 41, 42, 46]. In addition, we observed an almost two times higher risk to die with ATRX-/DAXX-negative tumors than patients with intact ATRX/DAXX tumors. These results underline the general hypothesis that loss of ATRX and DAXX is involved in tumor development and progression and thus leads to a more aggressive subtype of PanNETs. Our findings indicate a genetic association with prognosis and establish ATRX and DAXX as putative tumor suppressor genes in PanNETs. Therefore, these tumors need a more aggressive therapeutic strategy.

For the first time, we compared primary tumor tissue of PanNETs and their metastases. 2/74 (3%) primary PanNETs and 3/19 (16%) metastases divided into 2/12 (17%) lymph node metastases and 1/7 (14%) liver metastases showed a combined loss of ATRX and DAXX, underlining the correlation of a mutation with a more aggressive subtype of PanNETs. One reason, as recently published, might be the association of a mutation of ATRX and DAXX with an altered lengthening of the telomeres [16, 17, 42, 47].

Heterogeneity between primary tumors and lymph node metastases was found in 17%. These cases showed a different status for DAXX in lymph node metastases compared to the primary PanNET. These findings may result from tumor cells with a high mutation rate, a loss of DNA repair pathways, and therefore an intratumoral heterogeneity [48‒50]. Our results are underlined by the work of Cives et al. [51], who showed an accumulation of mutations concerning an increased size of primary PanNET. Interestingly, one of those cases had liver metastases, too. The liver metastases did not show any heterogeneity compared to the primary tumor tissue. In the future, this intratumoral heterogeneity may need to be considered in the diagnosis of punches. Hechtman et al. [40] showed very high sensitivity and specificity of IHC for DAXX mutation of 85% and 95%, respectively, in punches and fine needle aspirations of PanNETs. This underlines that IHC of DAXX and ATRX can be an accurate surrogate marker for mutations [40, 44]. As punches and needle aspirations are regularly performed preoperatively, the obtained specimens could deliver additional value for therapy decisions by performing additional IHC of ATRX and DAXX.

To examine if IHC is a surrogate for sequencing, we used NGS to verify mutations of DAXX and ATRX in ATRX- and DAXX-negative PanNETs. Of the 12/74 (16%) ATRX- and DAXX-negative PanNETs, 11/12 (92%) could be taken for further analysis. Somatic mutations of DAXX and ATRX were detected in 6/11 (55%) and 2/11 (18%) PanNETs. A loss of immunolabeling of ATRX (5/11, 45%) only corresponded with a mutation in 2/5 (40%) PanNETs with a specificity of 67%. Our results are consistent with Yamamichi et al. [52] and Purkait et al. [53], who showed a correspondence between ATRX loss for ATRX mutations in 77.8% (14/18) and 50% (3/6) on gliomas. The discrepancy between IHC ATRX results and NGS might be explained by intratumoral heterogeneity as described by Purkait et al. [53] on gliomas. They showed various statuses of mutation and expression when sampling was performed from different regions within the same tumor. These two studies on glioma and our study indicate that ATRX IHC might not be an optimal surrogate for sequencing. As the relation between the loss of immunolabeling of ATRX and ATRX mutation in PanNET is currently not well examined, further studies are needed.

Of the 7/11 (64%) DAXX-negative PanNETs, 6/11 (55%) harbored a mutation, revealing a specificity of 80% of DAXX loss for DAXX mutation. This finding indicates that also, contrary to ATRX, the immunolabeling of DAXX is a strong surrogate for sequencing. This assumption is further supported by Hechtman et al. [40], who showed a sensitivity and specificity of IHC for DAXX mutation of 85% and 95%. In contrast to Hechtman et al. [40], we identified a sensitivity of 100% as we did not examine a retained DAXX expression with DAXX mutation. Our examination might result from a small cohort as recent studies on ALT show that PanNETs positive for ALT had retained DAXX and ATRX expression on IHC in 6% up to 21% [41, 42, 44]. Further, Hechtman et al. [40] identified 4/27 (15%) PanNET with retained IHC DAXX expression and ALT due to a mutation in the last exon. These results suggest that the sensitivity of DAXX immunolabeling might be less than 100%. Our results further strengthen the evidence that DAXX immunolabeling can identify a more aggressive subgroup of PanNETs. This is shown by a statistical association of DAXX loss with a bad prognosis and a specificity of 80% of DAXX loss for DAXX mutation.

All through NGS examined mutations of ATRX and DAXX were heterozygote and correlated with immunolabeling loss (see online suppl. Table 4). We hypothesize that both copies of DAXX are generally inactivated, one allele by mutation and the other either by loss of the nonmutated allele or epigenetic silencing. In particular, we hypothesize in female individuals that both copies of ATRX are generally inactivated, too, one by mutation and the other allele by inactivation of the X chromosome.

The role of MMR gene deficiency contributing to MSI and tumorigenesis in PanNET has not yet been well investigated. The current opinion of the role and the percentage of MSI in PanNET varies from 0% to 36% [32‒35, 37, 54]. In line with these findings, our analysis of the four MMR proteins MSH2, MSH6, MLH1, and PMS2 using IHC revealed a loss in 4/74 (5%) PanNETs. All cases showed a loss of MSH6, two with an additional loss of MSH2. This finding is consistent with the current understanding of the rate of MSH6 mutations, which are most common in tumors, such as endometrial cancer or colorectal adenocarcinoma [55, 56]. MSI was performed to evaluate the meaning of an IHC loss of MMR proteins. Through FLA, MSI was obtained in 3/74 (4%). Due to the small number of MSI, no statistically significant association of MSI and a bad prognosis could be proven. Our results show an association of MSI with a higher UICC tumor stage and an unfavorable outcome for the patient. This is consistent with the results of Mei et al. [34], who found that MSI was significantly associated with tumor malignancy and immediately incurable disease. Further investigation is required for an additional treatment strategy of this aggressive subgroup of tumors.

This study protocol was reviewed and approved by the Institutional Review Board (REC reference number 15-5431). The written informed consent was obtained from each patient, whose tissue samples were used.

The authors have no conflicts of interest to declare.

This study was funded by Ipsen Pharma GmbH (Germany).

A. Tannapfel, A. Reinacher-Schick, and W. Uhl conceived the presented idea. W. Uhl performed the surgical resection of the PanNET. A. Reinacher-Schick, S. Nöpel-Dünnebacke, and O. Overheu collected the clinical data. A. Tannapfel and I. Tischoff verified the methods. I. Tischoff and D.M. Gisder evaluated the histologic sections. J. Keller and D.M. Gisder performed library preparation, sequencing, and data analysis for NGS. D.M. Gisder and O. Overheu developed the theory and performed the calculations and statistical analyses. A. Tannapfel, I. Tischoff, and D.M. Gisder wrote the manuscript with the cooperation of all authors. All authors discussed the results and contributed to the final manuscript.

All data generated or analyzed during this study are included in this article and its online supplementary material files. Further inquiries can be directed to the corresponding author.

1.
Matsukuma
KE
.
Neuroendocrine tumors of the pancreas
.
Pathobiology of human disease. A dynamic encyclopedia of disease mechanisms
.
Elsevier Science
.
2014
:
2284
92
.
2.
Fraenkel
M
,
Faggiano
A
,
Valk
GD
.
Epidemiology of neuroendocrine tumors
.
Front Horm Res
.
2015
;
44
:
1
23
.
3.
Halfdanarson
TR
,
Rabe
KG
,
Rubin
J
,
Petersen
GM
.
Pancreatic neuroendocrine tumors (PNETs): incidence, prognosis and recent trend toward improved survival
.
Ann Oncol
.
2008
;
19
:
1727
33
.
4.
Genus
T
,
Bouvier
C
,
Wong
K
,
Srirajaskanthan
R
,
Rous
B
,
Talbot
D
,
.
Incidence and prevalence of neuroendocrine tumours in England
.
Endocr Abstr
.
2017
;
52
.
5.
Dasari
A
,
Shen
C
,
Halperin
D
,
Zhao
B
,
Zhou
S
,
Xu
Y
,
.
Trends in the incidence, prevalence, and survival outcomes in patients with neuroendocrine tumors in the United States
.
JAMA Oncol
.
2017
;
3
(
10
):
1335
42
.
6.
Pape
U
,
Berndt
U
,
Müller-Nordhorn
J
,
Böhmig
M
,
Roll
S
,
Koch
M
,
.
Prognostic factors of long-term outcome in gastroenteropancreatic neuroendocrine tumours
.
Endocr Relat Cancer
.
2008
;
15
(
4
):
1083
97
.
7.
Hochwald
S
,
Zee
S
,
Conlon
KC
,
Colleoni
R
,
Louie
O
,
Brennan
M
,
.
Prognostic factors in pancreatic endocrine neoplasms: an analysis of 136 cases with a proposal for low-grade and intermediate-grade groups
.
J Clin Oncol
.
2002
;
20
:
2633
42
.
8.
Souche
R
,
Coignac
A
,
Dupuy
M
,
Bertrand
M
,
Raingeart
I
,
Guiu
B
,
.
Outcome after pancreatectomy for neuroendocrine neoplams according to the WHO 2017 grading system: a retrospective multicentric analysis of 138 consecutive patients
.
Clin Res Hepatol Gastroenterol
.
2020
;
44
(
3
):
286
94
.
9.
Basturk
O
,
Yang
Z
,
Tang
L
,
Hruban
R
,
Adsay
V
,
McCall
C
,
.
The high-grade (WHO G3) pancreatic neuroendocrine tumor category is morphologically and biologically heterogenous and includes both well differentiated and poorly differentiated neoplasms
.
Am J Surg Pathol
.
2015
;
39
(
5
):
683
90
.
10.
Tang
LH
,
Basturk
O
,
Sue
JJ
,
Klimstra
DS
.
A practical approach to the classification of WHO grade 3 (G3) well-differentiated neuroendocrine tumor (WD-NET) and poorly differentiated neuroendocrine carcinoma (PD-NEC) of the pancreas
.
Am J Surg Pathol
.
2016
;
40
(
9
):
1192
202
.
11.
Schindl
M
,
Kaczirek
K
,
Kaserer
K
,
Niederle
B
.
Is the new classification of neuroendocrine pancreatic tumors of clinical help
.
World J Surg
.
2000
;
24
(
11
):
1312
8
.
12.
Konukiewitz
B
,
Schlitter
A
,
Jesinghaus
M
,
Pfister
D
,
Steiger
K
,
Segler
A
,
.
Somatostatin receptor expression related to TP53 and RB1 alterations in pancreatic and extrapancreatic neuroendocrine neoplasms with a Ki67-index above 20
.
Mod Pathol
.
2017
;
30
(
4
):
587
98
.
13.
Milione
M
,
Maisonneuve
P
,
Spada
F
,
Pellegrinelli
A
,
Spaggiari
P
,
Albarello
L
,
.
The clinicopathologic heterogeneity of grade 3 gastroenteropancreatic neuroendocrine neoplasms: morphological differentiation and proliferation identify different prognostic categories
.
Neuroendocrinology
.
2016
;
104
(
1
):
85
93
.
14.
Rindi
G
,
Klersy
C
,
Albarello
L
,
Baudin
E
,
Bianchi
A
,
Buchler
M
,
.
Competitive testing of the WHO 2010 versus the WHO 2017 grading of pancreatic neuroendocrine neoplasms: data from a large international cohort study
.
Neuroendocrinology
.
2019
;
107
(
4
):
375
86
.
15.
Karnik
S
,
Hughes
C
,
Gu
X
,
Rozenblatt-Rosen
O
,
McLean
G
,
Xiong
Y
,
.
Menin regulates pancreatic islet growth by promoting histone methylation and expression of genes encoding p27Kip1 and p18INK4c
.
Proc Natl Acad Sci U S A
.
2005
;
102
(
41
):
14659
64
.
16.
Heaphy
C
,
De Wilde
R
,
Jiao
Y
,
Klein
A
,
Edil
B
,
Shi
C
,
.
Altered telomeres in tumors with ATRX and DAXX mutations
.
Science
.
2011
;
333
(
6041
):
425
.
17.
De Wilde
R
,
Heaphy
C
,
Maitra
A
,
Meeker
A
,
Edil
B
,
Wolfgang
C
,
.
Loss of ATRX or DAXX expression and concomitant acquisition of the alternative lengthening of telomeres phenotype are late events in a small subset of MEN-1 syndrome pancreatic neuroendocrine tumors
.
Mod Pathol
.
2012
;
25
(
7
):
1033
9
.
18.
Jiao
Y
,
Shi
C
,
Edil
B
,
De Wilde
R
,
Klimstra
D
,
Maitra
A
,
.
DAXX/ATRX, MEN1, and mTOR pathway genes are frequently altered in pancreatic neuroendocrine tumors
.
Science
.
2011
;
331
(
6021
):
1199
203
.
19.
Fishbein
L
,
Nathanson
KL
.
Inherited and somatic genetics of pancreatic neuroendocrine tumors
.
Management of pancreatic neuroendocrine tumors
.
New York
:
Springer
;
2014
. p.
9
32
.
20.
Watson
L
,
Solomon
L
,
Li
J
,
Jiang
Y
,
Edwards
M
,
Shin-Ya
K
,
.
Atrx deficiency induces telomere dysfunction, endocrine defects, and reduced life span
.
J Clin Invest
.
2013
;
123
(
5
):
2049
63
.
21.
Pea
A
,
Yu
J
,
Marchionni
L
,
Noe
M
,
Luchini
C
,
Pulvirenti
A
,
.
Genetic analysis of small well-differentiated pancreatic neuroendocrine tumors identifies subgroups with differing risks of liver metastases
.
Ann Surg
.
2020
;
271
(
3
):
566
73
.
22.
Marinoni
I
,
Schmitt
AM
,
Di Domenico
A
,
Perren
A
.
Genomic landscape in gastroenteropancreatic neuroendocrine neoplasms and its usefulness in improving the prognostic evaluation
.
Diagn Histopathology
.
2018
;
24
(
3
):
111
9
.
23.
Lewis
PW
,
Elsaesser
SJ
,
Noh
KM
,
Stadler
SC
,
Allis
CD
.
Daxx is an H3.3-specific histone chaperone and cooperates with ATRX in replication-independent chromatin assembly at telomeres
.
Proc Natl Acad Sci U S A
.
2010
;
107
(
32
):
14075
80
.
24.
Scarpa
A
,
Chang
D
,
Nones
K
,
Corbo
V
,
Patch
A
,
Bailey
P
,
.
Whole-genome landscape of pancreatic neuroendocrine tumours
.
Nature
.
2017
;
543
(
7643
):
65
71
.
25.
Chan
C
,
Laddha
S
,
Lewis
P
,
Koletsky
M
,
Robzyk
K
,
Da Silva
E
,
.
ATRX, DAXX or MEN1 mutant pancreatic neuroendocrine tumors are a distinct alpha-cell signature subgroup
.
Nat Commun
.
2018
;
9
(
1
):
1
10
.
26.
Loeb
LA
.
A mutator phenotype in cancer
.
Cancer Res
.
2001
;
61
(
8
):
3230
9
.
27.
Kunkel
TA
,
Erie
DA
.
DNA mismatch repair
.
Annu Rev Biochem
.
2005
;
74
:
681
710
.
28.
Meyers
M
,
Hwang
A
,
Wagner
M
,
Boothman
D
.
Role of DNA mismatch repair in apoptotic responses to therapeutic agents
.
Environ Mol Mutagen
.
2004
;
44
(
4
):
249
64
.
29.
Arnold
C
,
Nagasaka
T
,
Goel
A
,
Scharf
I
,
Grabowski
P
,
Sosnowski
A
,
.
Molecular characteristics and predictors of survival in patients with malignant neuroendocrine tumors
.
Int J Cancer
.
2008
;
123
(
7
):
1556
64
.
30.
Sameer
AS
,
Nissar
S
,
Fatima
K
.
Mismatch repair pathway: molecules, functions, and role in colorectal carcinogenesis
.
Eur J Cancer Prev
.
2014
;
23
(
4
):
246
57
.
31.
Müller
C
,
Schulmann
K
,
Reinacher-Schick
A
,
Andre
N
,
Arnold
D
,
Tannapfel
A
,
.
Predictive and prognostic value of microsatellite instability in patients with advanced colorectal cancer treated with a fluoropyrimidine and oxaliplatin containing first-line chemotherapy. A report of the AIO Colorectal Study Group
.
Int J Colorectal Dis
.
2008
;
23
(
11
):
1033
9
.
32.
House
M
,
Herman
J
,
Guo
M
,
Hooker
C
,
Schulick
R
,
Cameron
J
,
.
Prognostic value of hMLH1 methylation and microsatellite instability in pancreatic endocrine neoplasms
.
Surgery
.
2003
;
134
(
6
):
902
8
.
33.
Arnason
T
,
Sapp
H
,
Rayson
D
,
Barnes
P
,
Drewniak
M
,
Nassar
B
,
.
Loss of expression of DNA mismatch repair proteins is rare in pancreatic and small intestinal neuroendocrine tumors
.
Arch Pathol Lab Med
.
2011
;
135
(
12
):
1539
44
.
34.
Mei
M
,
Deng
D
,
Liu
T
,
Sang
X
,
Lu
X
,
Xiang
H
,
.
Clinical implications of microsatellite instability and MLH1 gene inactivation in sporadic insulinomas
.
J Clin Endocrinol Metab
.
2009
;
94
(
9
):
3448
57
.
35.
Arnold
CN
,
Sosnowski
A
,
Schmitt-Gräff
A
,
Arnold
R
,
Blum
HE
.
Analysis of molecular pathways in sporadic neuroendocrine tumors of the gastro-entero-pancreatic system
.
Int J Cancer
.
2007
;
120
(
10
):
2157
64
.
36.
Liu
IH
,
Ford
JM
,
Kunz
PL
.
DNA-repair defects in pancreatic neuroendocrine tumors and potential clinical applications
.
Cancer Treat Rev
.
2016
;
44
:
1
9
.
37.
Nakata
B
,
Wang
Y
,
Yashiro
M
,
Nishioka
N
,
Tanaka
H
,
Ohira
M
,
.
Prognostic value of microsatellite instability in resectable pancreatic cancer
.
Clin Cancer Res
.
2002
;
8
(
8
):
2536
40
.
38.
Yao
J
,
Hassan
M
,
Phan
A
,
Dagohoy
C
,
Leary
C
,
Mares
J
,
.
One hundred years after “carcinoid”: epidemiology of and prognostic factors for neuroendocrine tumors in 35,825 cases in the United States
.
J Clin Oncol
.
2008
;
26
(
18
):
3063
72
.
39.
Chou
A
,
Itchins
M
,
de Reuver
P
,
Arena
J
,
Clarkson
A
,
Sheen
A
,
.
ATRX loss is an independent predictor of poor survival in pancreatic neuroendocrine tumors
.
Hum Pathol
.
2018
;
82
:
249
57
.
40.
Hechtman
JF
,
Klimstra
DS
,
Nanjangud
G
,
Frosina
D
,
Shia
J
,
Jungbluth
AA
.
Performance of DAXX immunohistochemistry as a screen for DAXX mutations in pancreatic neuroendocrine tumors
.
Pancreas
.
2019
;
48
(
3
):
396
9
.
41.
Marinoni
I
,
Kurrer
A
,
Vassella
E
,
Dettmer
M
,
Rudolph
T
,
Banz
V
,
.
Loss of DAXX and ATRX are associated with chromosome instability and reduced survival of patients with pancreatic neuroendocrine tumors
.
Gastroenterology
.
2014
;
146
(
2
):
453
60
.
42.
Singhi
A
,
Liu
T
,
Roncaioli
J
,
Cao
D
,
Zeh
H
,
Zureikat
A
,
.
Alternative lengthening of telomeres and loss of DAXX/ATRX expression predicts metastatic disease and poor survival in patients with pancreatic neuroendocrine tumors
.
Clin Cancer Res
.
2017
;
23
(
2
):
600
9
.
43.
Park
JK
,
Paik
WH
,
Lee
K
,
Ryu
JK
,
Lee
SH
,
Kim
YT
.
DAXX/ATRX and MEN1 genes are strong prognostic markers in pancreatic neuroendocrine tumors
.
Oncotarget
.
2017
;
8
(
30
):
49796
806
.
44.
VandenBussche
C
,
Allison
D
,
Graham
M
,
Charu
V
,
Lennon
A
,
Wolfgang
C
,
.
Alternative lengthening of telomeres and ATRX/DAXX loss can be reliably detected in FNAs of pancreatic neuroendocrine tumors
.
Cancer Cytopathol
.
2017
;
125
(
7
):
544
51
.
45.
Pipinikas
C
,
Dibra
H
,
Karpathakis
A
,
Feber
A
,
Novelli
M
,
Oukrif
D
,
.
Epigenetic dysregulation and poorer prognosis in DAXX-deficient pancreatic neuroendocrine tumours
.
Endocr Relat Cancer
.
2015
;
22
(
3
):
13
8
.
46.
Hackeng
W
,
Brosens
L
,
Kim
J
,
O’Sullivan
R
,
Sung
Y
,
Liu
T
,
.
Non-functional pancreatic neuroendocrine tumours: ATRX/DAXX and alternative lengthening of telomeres (ALT) are prognostically independent from ARX/PDX1 expression and tumour size
.
Gut
.
2022 May
;
71
(
5
):
961
73
.
47.
Kim
J
,
Brosnan-Cashman
J
,
An
S
,
Kim
S
,
Song
K
,
Kim
M
,
.
Alternative lengthening of telomeres in primary pancreatic neuroendocrine tumors is associated with aggressive clinical behavior and poor survival
.
Clin Cancer Res
.
2017
;
23
(
6
):
1598
606
.
48.
Manley
BJ
,
Hakimi
AA
.
Molecular profiling of renal cell carcinoma: building a bridge toward clinical impact
.
Curr Opin Urol
.
2016
;
26
(
5
):
383
7
.
49.
Li
Z
,
Hao
P
,
Wu
Q
,
Li
F
,
Zhao
J
,
Wu
K
,
.
Genetic mutations associated with metastatic clear cell renal cell carcinoma
.
Oncotarget
.
2016
;
7
(
13
):
16172
9
.
50.
Baldus
SE
,
Schaefer
KL
,
Engers
R
,
Hartleb
D
,
D
,
Gabbert
HE
,
.
Prevalence and heterogeneity of KRAS, BRAF, and PIK3CA mutations in primary colorectal adenocarcinomas and their corresponding metastases
.
Clin Cancer Res
.
2010
;
16
(
3
):
790
9
.
51.
Cives
M
,
Partelli
S
,
Palmirotta
R
,
Palmirotta
R
,
Lovero
D
,
Mandriani
B
,
.
DAXX mutations as potential genomic markers of malignant evolution in small nonfunctioning pancreatic neuroendocrine tumors
.
Sci Rep
.
2019
;
9
(
1
):
1
10
.
52.
Yamamichi
A
,
Ohka
F
,
Aoki
K
,
Suzuki
H
,
Kato
A
,
Hirano
M
,
.
Immunohistochemical ATRX expression is not a surrogate for 1p19q codeletion
.
Brain Tumor Pathol
.
2018
;
35
(
2
):
106
13
.
53.
Purkait
S
,
Miller
C
,
Kumar
A
,
Sharma
V
,
Pathak
P
,
Jha
P
,
.
ATRX in diffuse gliomas with its mosaic/heterogeneous expression in a subset
.
Brain Pathol
.
2017
;
27
(
2
):
138
45
.
54.
Kunstmann
E
,
Vieland
J
,
Brasch
F
,
Hahn
S
,
Epplen
J
,
Schulmann
K
,
.
HNPCC: six new pathogenic mutations
.
BMC Med Genet
.
2004
;
5
:
16
.
55.
Stefansson
I
,
Akslen
L
,
MacDonald
N
,
Ryan
A
,
Das
S
,
Jacobs
I
,
.
Is frequent in sporadic endometrial carcinomas with microsatellite loss of hMSH2 and hMSH6 expression instability
.
Clin Cancer Res
.
2002
;
8
(
1
):
138
43
.
56.
Wu
Y
,
Berends
M
,
Mensink
R
,
Kempinga
C
,
Sijmons
R
,
Van Der Zee
A
,
.
Association of hereditary nonpolyposis colorectal cancer: related tumors displaying low microsatellite instability with MSH6 germline mutations
.
Am J Hum Genet
.
1999
;
65
(
5
):
1291
8
.