Insulinoma-Associated-1: From Neuroendocrine Tumor Marker to Cancer Therapeutics

Insulinoma-associated-1 (IA-1 or INSM1) is a zinc-finger transcription factor that was identified in a human insulinoma subtraction library, from which it was isolated (1). Insm1 is an intronless gene encoding a 510-amino acid protein that contains 5 zinc-finger DNA-binding motifs and several potential functional domains (2, 3). This novel complementary DNA (cDNA) was isolated from a rare human insulinoma due to its unique tissue specificity. However, the restricted specific expression pattern and the biological function of INSM1 were at the time largely unknown, except for its markedly elevated expression levels in human insulinoma tissues (Fig. 1). The first INSM1-specific expression pattern in cells or tissues other than insulinoma was demonstrated in a rat pheochromocytoma cell line (PC-12). Subsequently, normal fetal human brain tissues and primary neuroendocrine tumor samples were observed to be positive for INSM1 expression by Northern blot analyses, including pituitary tumors, carcinoids, small-cell lung carcinoma (SCLC), medullary thyroid carcinoma, neuroblastoma, retinoblastoma, and medulloblastoma (4). An early study on 64 lung cancer cell lines and lung tumor tissues used Northern blot analyses to define IA-1 (later renamed INSM1) as a novel marker for neuroendocrine differentiation in human lung cancer (5). Importantly, additional studies from other laboratories later confirmed that INSM1 is a specific neuroendocrine tumor marker that is critical for neuroendocrine lung differentiation (6–8).

Neuroendocrinology was first described as the recognition of hormones secreted from the pituitary gland, which is closely controlled by the brain, particularly the hypothalamus (9). Endocrine glands contain small islands of secretory cells of epithelial origin that secret hormones. Although the endocrine system consists of numerous different glands distributed throughout the body, the total mass of endocrine cells in each gland with functional hormone production is relatively small. These endocrine cells secrete different hormones that target tissues in response to growth and development for maintaining the body's metabolism. Dysregulation of hormone secretion and abnormal neuroendocrine cell growth could be due to neuroendocrine cell transformation. During endocrine organ development, these endocrine/neuroendocrine cells exhibit a tissue-specific differentiation pathway, which determines various endocrine organs with their specific hormone-producing endocrine cells. Abnormal differentiation or dysregulation of these endocrine/neuroendocrine cells results in the generation of benign or malignant neuroendocrine tumors that could have a profound impact on the body's metabolism. Because INSM1 serves such a critical role in neuroendocrine cell differentiation, it is conceivable that INSM1 could emerge as a novel neuroendocrine-specific tumor marker and a critical regulator of neuroendocrine differentiation (4, 10–12). This review discusses the functional role of INSM1 as a biomarker for neuroendocrine tumors. INSM1 overexpression in neuroendocrine tumors could have negative effects on tumor cell growth. On the basis of our understanding of INSM1 as a neuroendocrine tumor–specific biomarker and its associated signaling pathways, INSM1 could be a prominent target for designing novel cancer therapeutic strategies.

IA-1 was later renamed INSM1 in the GenBank DNA database. The isolated cDNA encodes an intronless gene mapped to chromosome 20p11.2 (2). Its 1530 base-pair open-reading frame translates into a protein of 510 amino acids with a predicted molecular weight of 52,923 Da (Fig. 2). The N-terminal portion (∼250 amino acids) of INSM1 contains a Snail/Gfi-1 (SNAG) motif, 2 proline-rich regions (20%–30% proline content), a putative nuclear localization signaling (NLS) sequence (positions 221–246), 4 dibasic amino acid sites (positions 8–9, 11–12, 221–222, and 227–228), and a potential amidation signal sequence (PGKR). The SNAG motif functions as a transcriptional repressor domain involved in cell-cycle arrest, recruitment of histone-modifying factors, and interaction with LSD1 for SCLC cell proliferation (13–16). The presence of dibasic amino acids are characteristics of prohormone convertase–processing sites for peptide hormone such as insulin, glucagon, somatostation, and pancreatic polypeptide, while the amidation signal sequence facilitates the addition of an α-amide group in several bioactive neuroendocrine peptides (17, 18). However, no evidence exists thus far that the IA-1 transcription factor can be processed into a neuroendocrine peptide hormone. The proline-rich region located at the IA-1 N-terminus has been observed in numerous mammalian transcription factors and is essential for protein–protein interactions (19, 20). Studies confirmed that the N-terminal proline-rich region of INSM1 (43–58 amino acids) is capable of binding to cyclin D1 and receptor of activated C kinase 1 (RACK1) during cellular signaling (12, 21, 22). Another unique feature of the encoding sequence of IA-1 is its 5 Cys₂-His₂-type zinc-finger motifs, which suggest a functional role for IA-1 as a DNA-binding transcription factor. The DNA-binding target sequence, T^G/_T^C/_T^C/_T^T/_AGGGG^G/_TC^G/_A recognized by IA-1 was determined using a selected and amplified random oligonucleotide binding assay (3). Several downstream target genes were subsequently identified with this consensus DNA-binding sequence in their promoter region, including neurogenic differentiation factor 1 (NeuroD1), insulin, RE1 silencing transcription factor (REST), ripply transcription repressor 3 (Ripply3), neurogenin 3 (Ngn3), adherens junction belt-specific protein (Plekha7), and Insm1 itself (autoregulation; refs. 21, 23–25). INSM1 encodes a transcription repressor, which is highly conserved among different species. Evolutionary Insm1 homologues in chimpanzee, mouse, rat, Xenopus laevis, zebrafish, Drosophila melanogaster, and Caenorhabditis elegans exhibit 99.4%, 90.6%, 76.7%, 55.7%, 54.8%, 22.6%, and 18.2% identity, respectively (4, 26). A human homolog named INSM2 was identified with a broader tissue expression pattern and similar functional role in pancreatic islets (27, 28). The second zinc-finger motif is the most highly conserved with 96% identity among different species. The second highly conserved zinc-finger motif is particularly important, because only human INSM1 zinc-fingers 2 and 3 are sufficient for target gene binding (3). INSM1 expression is restricted to normal embryonic neuroendocrine tissues, including the developing forebrain, hindbrain, olfactory epithelium, retina, cerebellum, pancreas, thymus, thyroid, adrenal gland, and endocrine cells of the gastrointestinal tract (26, 29–31).

Because the identification of INSM1 as a transcription factor, biochemical analyses of its upstream and downstream target genes revealed that INSM1 binds to its own promoter and autoregulates INSM1 gene expression as a feedback mechanism during embryonic neuroendocrine cell differentiation (3). In addition, the Ngn3, Ripply3, NeuroD1, and insulin target genes were identified to be associated with pancreatic endocrine cell differentiation (21, 23, 24, 32). INSM1 transcriptional repressor activity involved the recruitment of cyclin D1 and histone deaceylatase-3 (HDAC-3) that led to the modification of the acetylation state of histone H3/4. INSM1 serves an important role in pancreatic cell differentiation. Two of the critical endocrine islet transcription factors, Ngn3 and NeuroD1, activate the E-box elements present in the INSM1 promoter, whereas INSM1 downregulates Ngn3 and NeuroD1 expression during pancreatic development (21, 29, 31, 33). In combination, the INSM1 autoregulatory mechanism and the Ngn3/NeuroD1/INSM1 feedback regulation strongly suggest that INSM1 is critical for the modulation of the transient expression of INSM1 during early pancreatic endocrine cell differentiation (33–36). In vivo analysis of the developing pancreas supports that Insm1 is the immediate downstream target of ngn3 and is upstream of neuroD1, pair-box 4 (pax4), aristaless-related homeobox (arx), homeodomain transcription factor 6.1 (nkx6.1), and pair-box 6 (pax6) gene expression in endocrine pancreas differentiation (29). The Insm1-mutant mouse is embryonic lethal (37). In Insm1 mutants, endocrine pancreatic β-cell development was severely impaired. Additional evidence supports that Insm1 is a transcription factor that acts in a pan-endocrine manner, which is essential for the differentiation of endocrine cells into pancreas, intestine, adrenal gland, and anterior pituitary gland. A recent study suggested that haplo-insufficiency of insm1 impaired postnatal baseline β-cell mass (38). The SNAG domain at the N-terminus of INSM1 recruits histone-modifying factors, lysine (K)-specific demethylase 1A (Kdm1a), histone deacetylase ^1/2 (Hdac1/2), and REST corepressor 1–3 (Rcor1–3) and other proteins implicated in transcriptional regulation high mobility group 20a/b (Hmg20a/b) and genetic suppressor element 1 (Gse1) in the anterior pituitary gland (15).

On the basis of the embryonic lethality and deficient catecholamine synthesis observed in Insm1-mutant mice, Insm1 expression appears to be crucial for sympatho-adrenal (SA) lineage differentiation (39). SA lineage cells are derived from the neural crest, which lead to sympathetic neurons and adrenal chromaffin cells. Cell differentiation and catecholamine synthesis involve multiple transcription factors such as achaete-scute complex homolog-like 1 (Mash1), paired-like homeobox 2a/b (Phox2a/b), Hand2, and GATA binding protein 2/3 (Gata2/3), which are critical for mouse fetal development and survival (40, 41). Castro and colleagues reported (42) that Insm1 could be a direct target of Mash1 regulation because Insm1 is not correctly initiated from the SA lineage in Phox2b- and Mash1-mutant mice (39). Therefore, Insm1 serves a key role in modulating SA lineage development and catecholamine synthesis.

In the developing brain, Insm1 expression serves a critical role in areas where neurogenesis occurs. Insm1-mutant mice failed to develop properly in the granule cell layer of the cerebellum, the dentate gyrus of the postnatal hippocampus, the ventricular zone, and the subventricular zone of the neocortex (43). The REST corepressors and the repressor INSM1 regulate the developing brain proliferation–differentiation balance (44). Insm1 downregulates adherens junction belt-specific protein Plekha7 through developing neocortex delamination (25). High levels of INSM1 expression were also detected in neuronal progenitors and developing sensory neurons (30, 31). A recent study demonstrated that the homeotic cell transformation of outer hair cells without INSM1 into inner hair cells reveals a mechanism by which these neighboring mechanosensory cells begin to differ (45). Consistently, IA-1 constitutes a novel molecule downstream of Xenopus homolog of achaete-scute (Xash1) that is involved in the formation of Xenopus noradrenergic primary neuron populations (46).

Regarding the structure and biological function of INSM1, this protein is a specific neuroendocrine tumor marker. INSM1-associated transformation processes are mainly involved in neuroendocrine tumorigenesis. Elevated IA-1 messenger RNA (mRNA) levels were first detected in a panel of 64 human lung cancer cell lines and multiple human lung tumors (5). Northern blot analysis revealed 97% (30/31) positivity in SCLC cell lines. In contrast, IA-1 mRNA was detected in only 13% (4/30) of non-SCLC cell lines, including 9 with a neuroendocrine phenotype (4/9 neuroendocrine positive non-SCLC cells) and 3 carcinoids. Recently, a mAb that facilitates the diagnosis of neuroendocrine tumors, namely clone A8 against INSM1 (sc-271408; Santa Cruz Biotechnology, Inc.) has become available commercially. IHC analyses of INSM1 in neuroendocrine and neuroepithelial neoplasms (NEN) revealed that INSM1 was detectable by IHC in 88.3% of 129 NEN specimens (47), with only 1 false positive detected in 27 neoplasms without a neuroepithelial or neuroendocrine component. Further IHC analysis for INSM1 was conducted with a tissue array containing 35 SCLC samples at different clinical stages and 5 normal lung tissues. All the SCLC tissues were strongly positive for INSM1 expression (48). Additional studies demonstrated that INSM1 is a superior biomarker for diagnosing neuroendocrine tumors in lung, thoracic cavity, carcinoid tumorlets of the lung, neuroendocrine of the uterine cervix, pancreatic neuroendocrine tumors, neuroendocrine carcinomas of the head and neck, Merkel cell carcinoma, and primary central nervous system neoplasms (49–58). Figure 3 shows examples of INSM1 IHC staining pattern using multiple neuroendocrine tumor tissue arrays (US Biomax, Inc.). INSM1 protein staining displays a clear nuclear pattern in SCLC, esophagus neuroendocrine, lung carcinoid, rectum carcinoid, colon neuroendocrine, cardiac neuroendocrine, retinoblastoma, medulloblastoma, and neuroblastoma derived from a TH-N-myc transgenic mouse (N-myc activated neuroblastoma tumor model; ref. 59).

Because INSM1 is only expressed in normal fetal tissues and neuroendocrine tumors, it was further investigated whether high INSM1 expression levels contribute to neuroendocrine tumor cell growth. In addition to the transcriptional regulation of target genes by INSM1, the extranuclear activity of INSM1 in neuroendocrine tumor cells was further explored. It was postulated that insulin and insulin receptor (InR) signaling could serve a key role in fueling tumors, thus connecting obesity with diabetes and cancer (60). Previous studies using an insulin-lowering drug known as metformin demonstrated that it was associated with the modulation of InR signaling and led to a significant decrease in cancer incidence (61). A study demonstrated that INSM1 induces pancreatic cell trans-differentiation by enhancing the InR-mediated phosphatidylinositol-4,5-bisphosphate 3-kinase/RAC-alpha serine/threonine-protein kinase (PI3K/AKT) signaling pathway (22). INSM1 serves an important role, because it is closely associated with pancreatic endocrine cell differentiation (24, 29, 37). The biochemical studies revealed that INSM1 binds to cyclin D1 and interrupts cyclin D1 and CDK4 binding, which induces non-neuroendocrine cell-cycle arrest (12). In neuroblastoma, Sonic Hedgehog (Shh) signaling positively correlates with N-myc and INSM1 expression, which contributes to neuroblastoma cell viability. N-myc binds and activates the E2-box of the INSM1 promoter, whereas INSM1 increases N-myc levels by activating the PI3K/AKT/glucogen synthase kinase 3 beta (GSK3β) signaling pathways and enhances N-myc stability (62). The positive feedback loop of N-myc and INSM1 stimulates neuroblastoma cell proliferation. In addition, INSM1 expression enhances cellular invasiveness and oncogenesis in neuroblastoma cells. A recent study revealed that INSM1 signals through the same Shh signaling pathway, with N-myc/Ascl1 activation of INSM1 via extracellular signal-regulated protein kinase/MAPK^1/2 (ERK/MEK^1/2) phosphorylation signaling occurring in neuroendocrine lung cancer (63). Shh facilitates N-myc stability through PI3K/AKT and INSM1 activation in neuroendocrine lung cancer cells. Both MEK/ERK^1/2 and PI3K/AKT phosphorylation signaling are required for Ser-54 (C-myc, Ser-62) phosphorylation and blockage of GSK3β-mediated phosphorylation of N-myc (Thr-50, C-myc/Thr-58), which stabilizes N-myc protein and prevents it from entering the proteasomal degradation pathway (Fig. 4). The negative effects of Shh inhibitor and/or knockdown of INSM1 in neuroendocrine lung cancer cells were further investigated. Although common treatment of human cancers using an inhibitor of Shh signaling pathway is marginally effective, the combination of a Smoothened inhibitor (cyclopamine or vismodegib) and a Gli-mediated inhibitor (GANT-61) greatly enhanced the inhibition of SCLC in the context of INSM1 expression. Therefore, targeting multiple signaling pathways associated with the activation of the INSM1/N-myc transcription factors could be a prominent approach to delay or interrupt neuroendocrine tumor growth.

INSM1 can be considered as an oncofetal differentiation factor according to its exclusive expression pattern in the fetal stage of neuroendocrine precursors and tumors, while being silenced in adult normal tissues. Upon transformation, the INSM1 gene is reexpressed strictly in neuroendocrine tumors, which resembles its fetal neuroendocrine tissue expression pattern. The reexpression of the INSM1 gene in neuroblastoma tumors contributes to the aggressive phenotype of a certain subtype of neuroblastomas with amplified N-myc oncoprotein expression (62). INSM1 and N-myc counter-regulation promotes neuroblastoma tumor growth in vivo. Therefore, it is logical to use an INSM1 promoter–driven luciferase screening-platform to investigate the role of INSM1 in the suppression of neuroendocrine tumors. A novel adenosine kinase (ADK) inhibitor, 5-iodotubercidin (5′-IT), was identified to exert potent inhibition of neuroblastoma tumor cell growth in parallel with INSM1 suppression (64). The inhibition effect of 5′-IT mediated through the adenosine receptor-3 signaling pathway that suppresses the cyclic AMP (cAMP), Wnt, and pERK^1/2 signaling pathways, which are involved in apoptosis. INSM1 expression and ADK signaling could contribute to cell growth and tumorigenicity in human neuroblastoma tumors. Because elevated INSM1 expression in NB and neuroendocrine lung cancer is closely associated with Shh signaling and N-myc stability, using inhibitors targeting both the Shh and ADK signaling pathways could be a novel approach to downregulate INSM1 and suppress neuroendocrine cancer growth. The development of a preclinical test for Shh and ADK signaling inhibitors would be useful. The blockage of upstream or downstream molecules in the Shh signaling pathway downregulates INSM1 expression and restricts SCLC proliferation (63). Because both Shh signaling and increased expression of INSM1/N-myc serve critical roles in the development and maintenance of SCLC (65, 66), these molecules could be potential direct therapeutic targets for cancer therapy. A recent study also demonstrated that the lysine-specific demethylase 1A (LSD1) inhibitor T-3775440 inhibits SCLC cell proliferation. T-3775440 disrupts LSD1 interactions with the SNAG domain containing INSM1 and growth factor–independent 1B (GFI1B) proteins (16).

The specific expression of INSM1 in neuroendocrine tumors indicates that INSM1 could be a prominent target for neuroendocrine tumor therapy. Instead of directly targeting the INSM1 transcription factor or INSM1-associated signaling pathways, it is also logical to target INSM1-positive tumor cells using INSM1 promoter–specific suicide gene therapy. The expression pattern of INSM1 is tightly controlled via specific INSM1 promoter expression, which is restricted to embryonic neuroendocrine precursors and tumors (31, 67). INSM1 promoter–linked suicide gene therapy led to specific death of neuroendocrine tumor cells in vivo and in vitro (68–70). To maintain INSM1 tissue specificity and to avoid the over powering effects of adenoviral regulatory element, the chicken β-globin HS4 insulator and two tandem copies of the neuronal restrictive silencer element (2xNRSE) were successfully constructed, which prevented interference from strong regulatory elements present in the adenoviral genome (71, 72). An unexpected benefit of the modification of the Ad-INSM1-promoter construct was 10-fold increase in promoter activity in intratumoral injected xenograft tumors. A conditionally replicating adenovirus (CRAds) was further constructed, which contained the 1.7-kb human INSM1 promoter (which drives the expression of Δ24E1A and HSV-tk) separated by an internal ribosomal entry site (IRES) element. Conditional expression of Δ24E1A in neuroendocrine tumor cells enhanced safe CRAds replication (oncolytic), which increased the cell killing efficiency (70). In a previous study, the INSM1 promoter was modified with an upstream chicken β-globin HS4 insulator element and 2xNRSE to maintain the promoter specificity and increase the sensitivity of the adenoviral vector. The results of cell viability MTS [3-(4,5-dimethyl-2yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetraazolium] assay of CRAd (HS4-INSM1-2xNRSE-Δ24E1A-IRES-HSV-tk)–mediated neuroendocrine lung tumor cell death are shown in Fig. 5. The INSM1 promoter–regulated replicating HSV-tk armed oncolytic adenovirus exhibited effective cell killing capacity and sensitivity, similar to the effects observed when the INSM1 promoter expression pattern is restricted to neuroendocrine tumors. INSM1-negative adeno-squamous cells (NCI-H596) and large cell lung carcinoma cells (NCI-H460) survived due to the lack of INSM1 promoter activity, in contrast to INSM1-positive cells (UMC-11, NCI-H1155, and NCI-H69), which were destroyed effectively. On the basis of the fact that the INSM1 promoter regulates specific gene expression in neuroendocrine tumors, a neuroendocrine-specific reporter assay using a modified INSM1 promoter linked to a Gaussia reporter gene was developed for the diagnostic evaluation of neuroendocrine tumor growth in vivo (73). Therefore, in addition to the therapeutic approach described above, it is possible to monitor neuroendocrine tumor growth during therapy using an INSM1 promoter–driven Gaussia reporter adenoviral vector.

INSM1 was identified in a human insulinoma subtraction library >25 years ago (1). INSM1 represents a common zinc-finger transcription factor with a unique tissue-specific expression pattern. The encoded protein recognizes a specific DNA target sequence and downregulates the expression of multiple target genes associated with neuroendocrine cell differentiation. Functional analyses revealed that INSM1 serves a critical role in pancreatic endocrine, SA lineage, and pan-neurogenic precursor development, which is consistent with the notion that INSM1 functions as a neuroendocrine differentiation factor (37, 39, 43). Another important feature of INSM1 is its high expression level in neuroendocrine tumors. INSM1 has emerged as a sensitive and specific biomarker for neuroendocrine tumors, particularly SCLC (6–8). However, it is not clear whether INSM1 solely is a biomarker or whether its presence in neuroendocrine tumors also contributes to neuroendocrine tumorigenicity. A previous study on the molecular mechanisms of INSM1 in neuroendocrine tumors revealed that INSM1 exerts extranuclear activities that potentiate the tumorigenicity by activating Shh, PI3K/AKT, MEK/ERK^1/2, ADK, p53, cyclin D1, β-catenin, LSD1, and N-myc. Although it is generally perceived that direct targeting of transcription factors is relatively difficult when trying to achieve effective tumor suppression, recent studies have shown that targeting INSM1-associated signaling pathways in the context of INSM expression is effective (63, 64). Furthermore, fusing the specific INSM1 promoter to guide suicide or reporter genes targeting neuroendocrine tumors could be a novel approach for INSM1 promoter–mediated direct diagnosis or inhibition of neuroendocrine tumors. This review reports a unique neuroendocrine transcription factor (INSM1) that may serve as an neuroendocrine biomarker for the development of novel cancer therapeutics against neuroendocrine tumors.

No potential conflicts of interest were disclosed.

The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

The authors acknowledge the support received from the Departments of Pediatrics and Genetics, Louisiana State University Health Sciences Center and Children's Hospital (New Orleans, LA). The research reported in this publication was supported by the NCI of the NIH under the award no. CA218764 (to M.S. Lan). This work was supported in part by the LSU Research Enhancement Program and LSU Leveraging Innovation for Technology Transfer (LIFT²; grant no. HSCNO-2017-LIFT-008, to M.S. Lan).