Comparison of INSM1 immunostaining with established neuroendocrine markers synaptophysin and chromogranin A in over 14,000 neuroendocrine and non-neuroendocrine tumors

neoplasms (NEN) and in 3.5% of 11,815 non-neuroendocrine neoplasms that were successfully analyzed. At least an occasional weak INSM1 positivity was observed in 59 different non-neuroendocrine tumor entities, of which 15 entities contained at least one case with strong INSM1 staining. A comparison with synaptophysin and chromogranin A staining revealed that in NEN, synaptophysin showed the highest sensitivity (93.3%), followed by INSM1 (89.2%) and chromogranin A (87.5%). In neuroendocrine carcinomas (NEC), sensitivity was highest for INSM1 (88.0%), followed by synaptophysin (86.5%) and chromogranin A (66.4%). If INSM1 was used as an additional marker, the sensitivity for detecting neuroendocrine differentiation in NEN increased from 96.6% (synaptophysin and chromogranin A) to 97.2% (synaptophysin, chromogranin A and INSM1). Our study shows that INSM1 is a useful additional marker for neuroendocrine differentiation with high sensitivity, particularly in NEC.


Introduction
The identification of neuroendocrine differentiation in tumors is important for the correct classification of tumors and to guide further therapy.Immunohistochemical detection of the well-established neuroendocrine markers synaptophysin and chromogranin A is commonly used for assuring neuroendocrine differentiation.However, both proteins can be partially or completely absent in some neuroendocrine neoplasms (Strojan et al., 2021, Derks et al., 2019) and are known to be positive in some non-neuroendocrine tumors, such as basal cell carcinoma (Houcine et al., 2017, Terada, 2013) or adrenal cortical carcinoma (Weissferdt et al., 2014, Sangoi and McKenney, 2010, Komminoth et al., 1995).Furthermore, both markers are expressed in the cytoplasm, a cell compartment where false interpretation of positive immunostaining is particularly common.
To comprehensively assess the diagnostic utility of INSM1 immunohistochemistry, a highly standardized study analyzing a large number of unselected neuroendocrine and non-neuroendocrine tumors is needed.Therefore, INSM1 expression was analyzed in 14,908 tumor tissue samples from 117 different tumor types and subtypes as well as 76 different non-neoplastic tissue types by immunohistochemistry in a tissue microarray (TMA) format in this study.

Tissue microarrays (TMAs)
Normal and tumor tissue TMAs were used for this study.The normal tissue TMA was composed of 8 samples from 8 different donors for each of 76 different normal tissue types (608 samples on one slide).The tumor TMAs contained a total of 14,908 primary tumors from 117 tumor types and subtypes.Detailed histopathological data on grade, pT and pN status, (as well as HER2, ER, and PR status for breast cancer) were available from 2,139 breast cancers, 1,158 urothelial carcinomas, 327 gastric adenocarcinomas and 192 neuroendocrine neoplasms.Clinical follow up data were available from 877 patients with breast cancer (median follow-up time 43 months; range 1-88 months) and 254 patients with urothelial carcinomas (median follow time 39 months; range 1-250).In these patients, the median follow-up time was 43 (range 1-88) months.The composition of both normal and tumor TMAs is described in detail in the results section.All samples were from the archives of the Institutes of Pathology, University Hospital of Hamburg, Germany, the Institute of Pathology, Clinical Center Osnabrueck, Germany, and Department of Pathology, Academic Hospital Fuerth, Germany.Tissues were fixed in 4% buffered formalin and then embedded in paraffin.The TMA manufacturing process was described earlier in detail (Dancau et al., 2016, Kononen et al., 1998).In brief, one tissue spot (diameter: 0.6 mm) was transmitted from a tumor containing donor block (≥70% tumor cells) in an empty recipient paraffin block.The use of archived remnants of diagnostic tissues for manufacturing of TMAs and their analysis for research purposes as well as patient data analysis has been approved by local laws (HmbKHG, §12) and by the local ethics committee (Ethics commission Hamburg, WF-049/09).All work has been carried out in compliance with the Helsinki Declaration.

Immunohistochemistry (IHC)
Freshly prepared TMA sections were immunostained on one day in one experiment.Slides were deparaffinized with xylol, rehydrated through a graded alcohol series, and exposed to heat-induced antigen retrieval for 5 min in an autoclave at 121 • C in pH 7.8 buffer.Endogenous peroxidase activity was blocked with Dako Peroxidase Blocking Solution™ (Agilent, CA, USA; #52023) for 10 min.Primary monoclonal rabbit recombinant antibody INSM1 ( MSVA-456R, #3218-456R, MS Validated Antibodies GmbH, Hamburg, Germany) was applied at 37 • C for 60 min at a dilution of 1:100.For the purpose of antibody validation, the normal tissue TMA was also analyzed by the monoclonal rabbit recombinant INSM1 antibody MRQ-70 (Cell Marque, Darmstadt, Germany; # 475R-94) at a dilution of 1:50 and an otherwise identical protocol.Bound antibody was then visualized using the EnVision Kit™ (Agilent, CA, USA; #K5007) according to the manufacturer's directions.The sections were counterstained with haemalaun.All tissue spots were analyzed by one experienced pathologist.Several studies have demonstrated that involving multiple pathologists or non-pathologists in reading the same TMA slides would not significantly impact study outcome (Callau et al., 2015;Parker et al., 2002;Tennstedt and Sauter, 2010;Thomson et al., 2009).For tumor tissues, the percentage of INSM1 positive tumor cells was estimated and the staining intensity was semi-quantitatively recorded (0, 1+, 2+, 3+).For statistical analyses, the staining results were categorized into four groups as follows: Negative: no staining at all, weak staining: staining intensity of 1+ in ≤70% or staining intensity of 2+ in ≤30% of tumor cells, moderate staining: staining intensity of 1+ in >70%, staining intensity of 2+ in >30% but in ≤70% or staining intensity of 3+ in ≤30% of tumor cells, strong staining: staining intensity of 2+ in >70% or staining intensity of 3+ in >30% of tumor cells.

Statistics
Statistical calculations were performed with JMP 14 software (SAS Institute Inc., NC, USA).Contingency tables and the chi 2 -test were performed to search for associations between INSM1 and tumor phenotype.Survival curves were calculated according to Kaplan-Meier.The Log-Rank test was applied to detect significant differences between groups.A p-value of ≤0.05 was defined as significant.Sensitivity and specificity were calculated using the formulas TP/(TP + FN) and TN/ (TN + FP), respectively, where TP = number of true positive, TN = number of true negative, FP = number of false positive, FN = number of false negative.

Technical issues
A total of 12,286 (82.4%) samples were successfully analyzed.Noninterpretable samples were due to lack of unequivocal tumor cells or loss of the tissue spot during technical procedures for one or both of the markers.A sufficient number of samples of each normal tissue type was evaluable.

INSM1 immunostaining in normal tissues
In normal tissues, a nuclear INSM1 immunostaining was observed in scattered neuroendocrine cells in the tubular gut, such as the mucosa of the stomach, duodenum, ileum, rectum and appendix, in islets of Langerhans of the pancreas, in scattered cells in the thymic medulla, in medullar but not cortical cells of the adrenal gland, and in a fraction of epithelial cells in the adenohypophysis.Representative images of INSM1 staining in normal tissues are shown in Fig. 1.All these stainings were obtained by using both rabbit recombinant antibodies MSVA-456R and MRQ-70 and therefore considered to be specific (Supplementary Fig. 1).By using both antibodies, nuclear INSM1 immunostaining was not seen in any other normal tissues including squamous epithelium, sebaceous glands, Brunner glands, gall bladder, salivary glands, breast, endocervical gland, endometrium, fallopian tube, ovary, placenta, chorion cells, amnion cells, respiratory epithelium, lung, kidney, urothelium, prostate, seminal vesicles, testis, epididymis, thyroid, parathyroid, neurohypophysis, and the brain.An occasional weak staining of the cytoplasm and apical membranes of some renal tubuli was considered a tolerable antibody specific cross-reactivity of MSVA-456R.

INSM1 immunostaining in neoplastic tissues
A positive immunostaining was seen for INSM1 in 89.2% of 471 neuroendocrine neoplasms including 16.6% with weak, 22.3% with moderate, and 50.3% with strong staining (Table 1).In nonneuroendocrine neoplasms, INSM1 was detected in 3.5% of 11,815 cases including 2.2% with weak, 0.9% with moderate and 0.3% with strong positivity.Among 102 analyzed non-neuroendocrine tumor entities, INSM1 positivity was seen in at least one case in 59 different tumor entities, of which 15 entities contained at least one case with strong INSM1 staining.A ranking order of INSM1 immunostaining is given in Table 2.These data illustrate, that INSM1 staining predominates in various types of neuroendocrine neoplasms (>75% positive), followed by pancreatic acinar cell carcinoma (5 of 15 (33.3%) with weak positivity).All other non-neuroendocrine tumors showeda mostly weak to moderatepositivity in <20% of tumors.Representative images of INSM1 positive neuroendocrine and non-neuroendocrine neoplasms are shown in Figs. 2 and 3.The relationship of INSM1 expression with histopathological, molecular and clinical features is shown in Supplementary Table 1.In 1,140 analyzable invasive breast carcinomas of no special type detectable INSM1 immunostaining was linked to low tumor stage (p = 0.0026), positive estrogen receptor (p = 0.0004) and progesterone receptor status (p = 0.0025), but unrelated to overall survival (data not shown).INSM1 immunostaining was unrelated to histopathological features in neuroendocrine neoplasms, urothelial carcinoma and gastric adenocarcinomas.

Comparison of INSM1 immunostaining with synaptophysin and chromogranin A
Data on synaptophysin and chromogranin A immunostaining obtained on the same TMAs was available from a previous study for 8,444 tumors (Uhlig et al., 2022).A comparison of INSM1, synaptophysin and chromogranin A results is given for all tumor entities with at least one positive case for one of these antibodies in Fig. 4. The major differences between these three antibodies, that are considered the most suitable existing neuroendocrine markers, was that chromogranin A stained basal cell carcinomas of the skin and synaptophysin stained adrenocortical neoplasms while INSM1 did not stain any particular non-neuroendocrine tumor entity.
In the group of neuroendocrine neoplasms, synaptophysin showed the highest sensitivity (93.3%; n = 511), followed by INSM1 (89.2%; n = 471) and chromogranin A (87.5%; n = 473).In the subgroup of neuroendocrine carcinomas, INSM1 showed the highest sensitivity (88.0%; n = 117), followed by synaptophysin (86.5%; n = 126) and chromogranin A (66.4%; n = 122).As expected from the results above, the combination synaptophysin/chromogranin A (96.6%) was comparable with the combination synaptophysin/INSM1 (96.3%) in detecting neuroendocrine differentiation in neuroendocrine neoplasms.In the subgroup of neuroendocrine carcinomas, the combination of INSM1/  Evaluating the specificity of INSM1 in the group of nonneuroendocrine tumors is difficult, since many non-neuroendocrine tumors can show neuroendocrine differentiation to some extent.On the basis of the hypothesis that true neuroendocrine differentiation would be detected by more than one marker, we compared how many tumors were positive for one, two or three markers.The percentage of "morphologically non-neuroendocrine" tumors showing positivity for at least 2 neuroendocrine markers was 1.2% for synaptophysin/chromogranin A, 1.2% for synaptophysin/INSM1, 1.1% for INSM1/chromogranin A, and 0.8% for synaptophysin/chromogranin A/INSM1.Positivity for at least one additional neuroendocrine marker was seen in non-neuroendocrine tumors for synaptophysin positive tumors in 46.1%, for INSM1 positive tumors in 38.9%, and for chromogranin A positive tumors in 34.4% (Fig. 5).

Discussion
The successful analysis of 12,286 tumor samples revealed INSM1 positivity in 6.8% of cases.As expected, more than half of the INSM1 positive tumor samples belonged to the small group of neuroendocrine neoplasms.In the group of non-neuroendocrine tumors 3.5% of cases showed INSM1 positivity, as compared to 89.2% in neuroendocrine neoplasms.Since INSM1 is a relatively new immunohistochemical marker for neuroendocrine differentiation, INSM1 immunostaining data were compared with the results from a previous study analyzing the most commonly used neuroendocrine markers synaptophysin and chromogranin A (Uhlig et al., 2022).In contrast to INSM1, which is a transcription factor not involved in the vesicle transportation system,    (Leblebici et al., 2019, Lilo et al., 2018).However, if all neuroendocrine neoplasms including neuroendocrine tumors of different sites, neuroendocrine carcinomas of different sites, Merkel cell carcinomas, medullary thyroid carcinomas, paragangliomas and pheochromocytomas were jointly analyzed, synaptophysin (93.3%) had the highest sensitivity in detecting neuroendocrine differentiation, followed by INSM1(89.2%)and chromogranin A (87.5%).Although INSM1 is recommended in the WHO classification of endocrine and neuroendocrine tumors and has been successfully implemented in some laboratories (Rindi et al., 2022, Juhlin, 2021, Mete et al., 2022), many others still use the immunohistochemical markers synaptophysin and chromogranin A in combination to gain higher specificity and sensitivity for neuroendocrine differentiation in daily practice.However, when combining two different immunohistochemical markers it might be advantageous if the individual markers stain different cellular compartments.In contrast to synaptophysin and chromogranin A, which both show a cytoplasmatic staining, INSM1 is a transcription factor and hence localized in the nucleus and could therefore represent an ideal marker in combination with synaptophysin or chromogranin A. We therefore compared the combination of    (90.7%).It is of note, that the nuclear location of the INSM1 protein also has the advantage that non-specific nuclear staining occurs less commonly than non-specific cytoplasmic immunostaining.The biological and clinical significance of the expression of neuroendocrine markers in non-neuroendocrine tumors is not clear, especially if they show well differentiated tumor specific morphology.For example, studies on adenocarcinoma or squamous cell carcinoma of the lung with positivity for neuroendocrine markers but lack of neuroendocrine morphology showed different results.While some studies showed that expression of neuroendocrine markers had an impact on survival (Feng et al., 2016, Petrovic et al., 2011) other authors reported no effect (Kriegsmann et al., 2021, Sterlacci et al., 2009, Ionescu et al., 2007).Defining "sensitivity" and "specificity" of neuroendocrine markers for detecting true "neuroendocrine" features is therefore arbitrary and potentially controversial in these tumors.Our assessment of the positivity rate for at least two different neuroendocrine markers was based on the assumption that tumors with expression of more than one of these proteins would be most likely to have undergone true "neuroendocrine differentiation".Considering that at least one additional neuroendocrine marker occurred in 46.1% of synaptophysin positive, 38.9% of INSM1 positive, and in 34.4% of chromogranin A positive tumors, we consider these figures as potentially reflective of marker specificity for neuroendocrine differentiation of non-neuroendocrine tumors.
The evaluation of large cohorts of breast and urinary bladder carcinomas as well as neuroendocrine neoplasms did not suggest a major clinical and prognostic impact of INSM1 expression in these entities.Significant associations between INSM1 expression and parameters of tumor aggressiveness were only seen in breast cancer.The significant relationship of INSM1 positivity with high pT stage -a parameter for aggressive disease -and with PR and ER positivity -two parameters associated with favorable disease outcomedid not provide unambiguous evidence for a prognostic role of INSM1 expression.One study on 1,523 patients had previously suggested a favorable prognostic impact of INSM1 expression in luminal B breast carcinoma (Razvi et al., 2021).Other breast cancer studies on 43 and 37 patients had failed to show any correlation between INSM1 staining and prognosis (Chen et al., 2021, Kuji et al., 2017).It is possible that INSM1 expression has more significance in other tumor entities.Two studies described a link between high INSM1 expression and a worse prognosis in small cell lung carcinomas (Sakakibara et al., 2020, Xu et al., 2022).Minami et al. found a significant correlation between positive INSM1 immunostaining and an unfavorable prognosis in patients with pulmonary high grade neuroendocrine carcinoma (Minami et al., 2020).Kim et al. showed that INSM1 staining decreased with tumor grade in pancreatic neuroendocrine tumors (Kim et al., 2020).
Considering the large scale of our study, emphasis was placed on a thorough validation of our assay.The International Working Group for Antibody Validation (IWGAV) has proposed that antibody validation for immunohistochemistry on formalin fixed tissues should include either a comparison of the findings obtained by two different independent antibodies or a comparison with expression data obtained by another independent method (Uhlen et al., 2016).RNA data obtained in three independent RNA screening studies, including the Human Protein Atlas (HPA) RNA-seq tissue dataset (Thul et al., 2017), the FANTOM5 project (Lizio et al., 2019, Lizio et al., 2015), and the Genotype-Tissue Expression (GTEx) project (Consortium, 2013)  ssue).That the immunohistochemical analysis of 76 different normal tissue categories by using MSVA-456R resulted in complete restriction of nuclear INSM1 staining to these organs and that the same cell types as detected by MSVA-456R were also positive by MRQ-70 strongly validates our assay.MSVA-456R was selected for the tumor cohort analysis because of the lower costs (about 9 € per slide) as compared to MRQ-70 (about 40 € per slide).The fact that INSM1 was not detected by both antibodies in the brain is consistent with data from Ames et al. who described INSM1 expression in a considerable fraction of brain tumors but also failed to detect INSM1 positivity in normal brain tissue (Ames et al., 2018).The occasional cytoplasmic and apical membrane staining of renal tubuli by MSVA-456R was considered a tolerable antibody specific cross-reactivity.It is of note that our use of a very broad range of different tissues for antibody validation increases the likelihood for detecting undesired cross-reactivities because virtually all proteins occurring in normal cells of adult humans are subjected to the validation experiment.
In summary, our large study shows that INSM1 is a highly useful marker for detecting neuroendocrine differentiation with high sensitivity particularly in neuroendocrine carcinomas.INSM1 can be used as a stand-alone first-line marker or can be used as an additional nuclear marker to complement the well-established cytoplasmatic markers synaptophysin and chromogranin A.

Fig. 1 .
Fig. 1.INSM1 immunostaining in normal tissues.(A) Moderate to strong INSM1 staining in scattered neuroendocrine cells in the mucosa of duodenum.(B) Weak to moderate INSM1 staining in islets of Langerhans in the pancreas.(C) Moderate to strong INSM1 staining in the medulla of the adrenal gland (D) Moderate to strong INSM1 staining in scattered epithelial cells in the thymus.(E) Moderate INSM1 staining in a fraction of cells in the anterior lobe of the pituitary gland.(F) Moderate INSM1 staining in scattered cells in the mucosa of the bronchus.

Fig. 2 .
Fig. 2. INSM1 immunostaining in neuroendocrine neoplasms.Strong staining in (A) a neuroendocrine tumor of the lung, (B) a neuroendocrine tumor of the pancreas, (C) a neuroendocrine carcinoma of the pancreas, (D) a neuroendocrine tumor of the illeum, (E) a neuroendocrine colorectal tumor, (F) a neuroendocrine colorectal carcinoma, (G) a neuroendocrine tumor of the appendix, and (H) a neuroendocrine carcinoma of the gall bladder.

Fig. 3 .
Fig. 3. INSM1 immunostaining in non-neuroendocrine tumors.(A) Strong INSM1 immunostaining in a mucinous carcinoma of the breast.(B) Moderate to strong INSM1 staining in a urothelial carcinoma.(C) Moderate to strong INSM1 staining in a gastric adenocarcinoma, diffuse type.(D) Moderate to strong staining INSM1 staining in a colorectal adenocarcinoma.(E) Strong INSM1 immunostaining in a carcinosarcoma of the ovary.(F) Moderate to strong INSM1 staining in an adenocarcinoma of the prostate.

Fig. 4 .
Fig. 4. Comparison of INSM1, synaptophysin and chromogranin A results in all tumor entities with at least one positive case for one of these antibodies.For each tumor the bar shows the percentage of (weak/moderate/strong) staining of each neuroendocrine marker (INSM1: blue; chromogranin A: red; synaptophysin: green).

Fig. 5 .
Fig. 5. Comparison of INSM1, chromogranin A and synaptophysin staining in non-neuroendocrine neoplasms.For each of the three neuroendocrine markers the bar shows the percentage of tumors which were positive for at least one additional marker.

Table 1
INSM1 in tumors.(Neuroendocrine neoplasms are highlighted in grey).

Table 2
Ranking of INSM1 immunostaining in tumors.(Only tumor entities with ≥ 3 evaluable tumors were included in the ranking.Neuroendocrine neoplasms are highlighted in grey).

Table 2
synaptophysin/chromogranin A, synaptophysin/INSM1 and chromogranin/INSM1.As expected from the results of the single markers, the combination synaptophysin/INSM1 (96.3%) was comparable to the combination synaptophysin/chromogranin A (96.6%) in its capability of detecting neuroendocrine differentiation in neuroendocrine neoplasms.However, in the subgroup of neuroendocrine carcinomas, the combination of INSM1/synaptophysin (94.4%) was more sensitive than synaptophysin/chromogranin A (91.6%) and INSM1/chromogranin A