Aberrant Expression of Long Non Coding RNA HOTAIR and De-Regulation of the Paralogous 13 HOX Genes Are Strongly Associated with Aggressive Behavior of Gastro-Entero-Pancreatic Neuroendocrine Tumors

Gastro-entero-pancreatic neuroendocrine neoplasms (GEP-NENs) are rare diseases occurring in the gastrointestinal tract and pancreas. They are characterized by the loss of epithelial tubular gland elements, and by the increased expression of neuroendocrine markers. GEP-NENs are subdivided into two histo-pathological types, gastro-entero-pancreatic neuroendocrine tumors (GEP-NETs) and gastro-entero-pancreatic neuroendocrine carcinomas (GEP-NECs). According to WHO 2017 and 2019 classification criteria are graded and staged in four categories, NET-G1, NET-G2, NET-G3, and NEC-G3. The molecular characterization of these tumors can be fundamental for the identification of new diagnostic, prognostic and predictive biomarkers. The main purpose of this study was to analyze the expression of the paralogous 13 HOX genes, normally involved in embryogenic development and frequently deregulated in human cancers, and of the HOX regulating lncRNA HOTAIR in GEP-NENs. The expression of HOX genes is gradually lost in the transition from GEP NET G1 to NET/NEC G3 tumors, while HOTAIR expression, inversely correlated with HOX genes expression and weakly expressed in low-grade GEP NENs, becomes aberrant in NET G3 and NEC G3 categories. Our data highlights their potential role in the molecular stratification of GEP-NENs by suggesting new prognostic markers and potential therapeutic targets.


Introduction
Neuroendocrine neoplasms are a heterogeneous group of rare tumors that can arise from different anatomic sites including foregut, midgut, hindgut, lung, bladder, prostate, adrenal gland and sympathetic nervous organs [1].The term "neuro" is associated with the presence of dense core granules that are similar to those present in serotonergic neurons, which store monoamines, while the "endocrine" property refers to the synthesis and secretion of these monoamines [2].Tumors arising from gastrointestinal tract, also defined Gastro-entero-pancreatic neuroendocrine tumors (GEP-NET) account for two-thirds of NENs.
All these categories are specific entities with different prognostic and therapeutic implications, for which a molecular characterization to identify new diagnostic and therapeutic tools, useful in their management, should be required [5].
HOX genes play a main role during the embryonic development, controlling the identity of different regions along the body axis, from the branchial area to the tail [6].In cancer diseases, HOX genes are involved in different oncogenic processes, such as the control of cell differentiation, proliferation, apoptosis, cell invasion and epithelial-mesenchymal transition [7].In particular, the genes belonging to HOX paralogous group 13 (HOXA13, HOXB13, HOXC13, and HOXD13) are strongly associated with cancer development and progression [8][9][10][11][12].Posterior genes of the HOX genes network are involved in the development of the gut [13], and emerging evidences suggest that their de-regulation can be associated with exocrine tumor of pancreas [14] and colon cancers evolution [15].
HOX genes expression can be modulated by different non-coding RNAs (ncRNAs) some of which are localized within HOX loci [16].In particular, the long non-coding RNA (lncRNA) HOTAIR (Hox transcript antisense intergenic RNA) is able to modulate metastatic processes in several human cancers [17].
Since the de-regulation posterior genes of the HOX genes network has been recently described also in some tumors with neuroendocrine differentiation [18], in this study we aimed to investigate the role of the paralogous group 13 HOX genes in a case series of GEP-NEN patients.We also evaluated the role of lncRNA HOTAIR, a regulator of posterior HOX genes expression, recently associated with GEP-NEN tumor progression [19].

Paralogous 13 HOX Proteins Expression in GEP-NEN Patients in GEP-NEN Categories
The immunohistochemical analysis mainly revealed a nuclear localization of HOXA13, HOXB13, HOXC13 and HOXD13 proteins, whereas a cytoplasmic localization was observed only in some areas (Figure 1).The performed statistical analysis reveals for both HOXA13, HOXC13, HOXD13 a strong association with tumor grading classification (Table 2).The performed statistical analysis reveals for both HOXA13, HOXC13, HOXD13 a strong association with tumor grading classification (Table 2).The general trend of all paralogous 13 HOX genes is a downregulation in high grade tumors.The statistical analysis showed that for HOXA13 the expression is significantly downregulated in both NETG3 and NEC G3 (Figure 1e-h; Figure 2a).HOXB13 is expressed heterogeneously in the 4 categories even if with a faint decrease in the G3 categories (Figure 1i-n; Figure 2b).HOXC13 is overexpressed in NET G1 and NET G2, downregulated in NEC G3, strongly downregulated in NET G3 (Figure 1o-r; Figure 2c).Similarly, HOXD13 is overexpressed in NET G1 and NET G2 tumors, but downregulated in NET G3, and strongly down regulated in NEC G3 categories (Figure 1s-v; Figure 2d).The general trend of all paralogous 13 HOX genes is a downregulation in high grade tumors.The statistical analysis showed that for HOXA13 the expression is significantly downregulated in both NETG3 and NEC G3 (Figure 1e-h; Figure 2a).HOX B13 is expressed heterogeneously in the 4 categories even if with a faint decrease in the G3 categories (Figure 1i-n; Figure 2b).HOXC13 is overexpressed in NET G1 and NET G2, downregulated in NEC G3, strongly downregulated in NET G3 (Figure 1o-r; Figure 2c).Similarly, HOX D13 is overexpressed in NET G1 and NET G2 tumors, but downregulated in NET G3, and strongly down regulated in NEC G3 categories (Figure 1s-v; Figure 2d).

HOTAIR Relationship with Tumor Location and GEP-NEN Categories
In situ hybridization (ISH) HOTAIR expression showed a tissue-specific distribution in our cases and no signals in non tumor and stromal cells were detected (Supplementary Figure S1).HOTAIR staining can be detected in both nucleus and cytoplasm of tumor cells (Figure 3).

HOTAIR Relationship with Tumor Location and GEP-NEN Categories
In situ hybridization (ISH) HOTAIR expression showed a tissue-specific distribution in our cases and no signals in non tumor and stromal cells were detected (Supplementary Figure S1).HOTAIR staining can be detected in both nucleus and cytoplasm of tumor cells (Figure 3).
Statistical elaboration showed that, while there is not direct relation between HOTAIR expression and GEP-NEN tumor location, a positive relationship between HOTAIR expression and grade has been detected.The HOTAIR expression significantly increased with the grade (p value = <0.001)(Table 3) (Figure 4).

HOTAIR Relationship with Tumor Location and GEP-NEN Categories
In situ hybridization (ISH) HOTAIR expression showed a tissue-specific distribution in our cases and no signals in non tumor and stromal cells were detected (Supplementary Figure S1).HOTAIR staining can be detected in both nucleus and cytoplasm of tumor cells (Figure 3).Statistical elaboration showed that, while there is not direct relation between HO-TAIR expression and GEP-NEN tumor location, a positive relationship between HOTAIR expression and grade has been detected.The HOTAIR expression significantly increased with the grade (p value = <0.001)(Table 3) (Figure 4).

Relationship Between Paralogous 13 HOX Genes and HOTAIR
HOXA13 is overexpressed in cases with low and intermediate HOTAIR scores, while it is almost completely absent in cases with high HOTAIR scores (p value = 0.004) (Figure 5a).No relationship between HOXB13 and HOTAIR has been detected (p value = 0.118) (Figure 5b).
HOXC13 is overexpressed in cases with low HOTAIR scores, and de-regulated in cases with high HOTAIR scores (p value = 0.001) (Figure 5c).HOXD13 is upregulated in cases with low HOTAIR scores, while it is downregulated in cases with intermediate/high HOTAIR scores (p value = 0.001) (Figure 5d).

Relationship Between Paralogous 13 HOX Genes and HOTAIR
HOXA13 is overexpressed in cases with low and intermediate HOTAIR scores, while it is almost completely absent in cases with high HOTAIR scores (p value = 0.004) (Figure 5a).No relationship between HOXB13 and HOTAIR has been detected (p value = 0.118) (Figure 5b).
HOXC13 is overexpressed in cases with low HOTAIR scores, and de-regulated in cases with high HOTAIR scores (p value = 0.001) (Figure 5c).HOXD13 is upregulated in cases with low HOTAIR scores, while it is downregulated in cases with intermediate/high HOTAIR scores (p value = 0.001) (Figure 5d).

Discussion
The incidence of GEP-NENs has increased over the last few decades.About 2/3 of patients can present distant metastases and the five-year survival rate exceeds 60%.For this reason, the incidence of GEP-NEN is higher than that of pancreatic, gastric and oesophageal adenocarcinomas, making it the second most prevalent type of cancer of the gastrointestinal tract [1].Moreover, the recent classification criteria based on tumor grading and cell differentiation, has allowed the identification of categories for GEP-NENs, with very different prognostic implications [3,4].Consequently, the possibility of early detection and the development of appropriate therapeutic strategies promoted the research of new diagnostic and prognostic biomarkers.
In this study we focused the attention on the role of HOX13 paralogous genes in GEP-NEN tumors and investigated the potential contribute of the lncRNA HOTAIR in the definition of their prognostic categories.
The alteration of expression of the HOX genes seems to be strongly associated with the different development sites of GEP-NEN tumors but also with the 4 grading related categories.In detail, HOXA13 always appears to be downregulated in the prognostically most unfavorable category, NECG3, compared to the other tumor types.Similarly, HOXC13 and HOXD13 are downregulated in NECG3.
During development, paralogous 13 HOX genes are involved in mediating the transition from the early to the late-distal limb program, controlling the spatial-temporal expression patterns of target genes [20] and in mediating gut and urogenital system formation [21][22][23].However, many of them are still active in adult human organs and tissues and frequently deregulated in human cancers [8,11,23].We have recently shown their deregulation in colon cancer, describing the absence of expression of HOXA13, HOXB13, HOXC13 and HOXD13 in the normal colon mucosa, a slight increase in expression in the transitional mucosa, up to in many cases over-expressed in the tumor.In particular, HOX B13 and HOXC13 expression strongly correlated with lymph nodes metastasis and showed a prevalent expression in CRC samples with a very poor prognosis.On the contrary, in pancreatic cancer, HOXD13 displayed on opposite trend being strongly downregulated in cancer cells.For this reason, for these genes a dual role has been suggested during tumor evolution, as oncogenes and tumor suppressor genes [9].

Discussion
The incidence of GEP-NENs has increased over the last few decades.About 2/3 of patients can present distant metastases and the five-year survival rate exceeds 60%.For this reason, the incidence of GEP-NEN is higher than that of pancreatic, gastric and oesophageal adenocarcinomas, making it the second most prevalent type of cancer of the gastrointestinal tract [1].Moreover, the recent classification criteria based on tumor grading and cell differentiation, has allowed the identification of different categories for GEP-NENs, with very different prognostic implications [3,4].Consequently, the possibility of early detection and the development of appropriate therapeutic strategies promoted the research of new diagnostic and prognostic biomarkers.
In this study we focused the attention on the role of HOX13 paralogous genes in GEP-NEN tumors and investigated the potential contribute of the lncRNA HOTAIR in the definition of their prognostic categories.
The alteration of expression of the HOX genes seems to be strongly associated with the different development sites of GEP-NEN tumors but also with the 4 grading related categories.In detail, HOXA13 always appears to be downregulated in the prognostically most unfavorable category, NECG3, compared to the other tumor types.Similarly, HOXC13 and HOXD13 are downregulated in NECG3.
During development, paralogous 13 HOX genes are involved in mediating the transition from the early to the late-distal limb program, controlling the spatial-temporal expression patterns of target genes [20] and in mediating gut and urogenital system formation [21][22][23].However, many of them are still active in adult human organs and tissues and frequently deregulated in human cancers [8,11,23].We have recently shown their de-regulation in colon cancer, describing the absence of expression of HOXA13, HOXB13, HOXC13 and HOXD13 in the normal colon mucosa, a slight increase in expression in the transitional mucosa, up to in many cases over-expressed in the tumor.In particular, HOX B13 and HOXC13 expression strongly correlated with lymph nodes metastasis and showed a prevalent expression in CRC samples with a very poor prognosis.On the contrary, in pancreatic cancer, HOXD13 displayed on opposite trend being strongly down-regulated in cancer cells.For this reason, for these genes a dual role has been suggested during tumor evolution, as oncogenes and tumor suppressor genes [9].
The ability of the HOX genes to molecularly characterize and differentiate the categories of GEP-NEN at risk of progression could add an important step to the understanding of the molecular mechanisms associated with the evolution of this tumor.
Although the entire HOX network plays a central role in cancer development and progression, the most posterior genes of the network are crucial in modulating these processes, in cooperation with co-localizing lncRNAs [24].
In particular, the lncRNA HOTAIR is able to act as a regulator of chromatin states by binding PRC2, with its 5 end and at 3 end, with LSD1 (lysine-specific demethylase 1) promoting the epigenetic activation/repression of gene expression [25].
In this study we have analyzed the expression of HOTAIR on our GEP-NEN cases, highlighting its gradual upregulation from the NETG1 to the more prognostically unfavorable NECG3 tumor type.
Currently, HOTAIR is considered as an important biomarker associated with the pathogenesis and progression of many tumors [26,27].In recent years, its prognostic role has appeared even more evident being involved also in the main mechanisms of therapeutic resistance [28].In different solid tumors, its role as a circulating marker has been shown, suggesting a main role in monitoring tumor evolution and response to specific therapies [29].
Our data support the idea that HOTAIR, as well as being an important prognostic marker, may represent also a potential therapeutic target in GEP-NEN tumors.In fact, different studies showed that the direct or indirect block/inhibition of HOTAIR may represent a new and effective cancer therapeutic strategy.The majority of the functional studies on HOTAIR performed its direct inhibition by siRNA methods, but the translation of these methods in clinical practice is complicated.Instead, more recently the design of small molecules able to specifically interfere with conserved RNA structures and to block HOTAIR protein complexes have proved more useful [30,31].
In the present study we also tried to establish a relationship between the expression of the paralogous 13 HOX genes and HOTAIR in GEP/NEN categories.HOXA13, HOXC13 and HOXD13 expression appears to be inversely related with HOTAIR expression, highlighting an inverse relationship with the prognostically unfavourable grading category.
The data appears quite in line with what is described in the literature, in particular for HOXD13.In fact, it is known that mechanistically HOTAIR is responsible for gene-silencing of an entire area of chromosome 12 including the HOXD locus and a series of metastasis suppressor genes, promoting metastatic switch [32].The only major limitation of this study is the low availability of recruitable patients, being very rare diseases and often treated in different Institutions.
In conclusion, although an accurate molecular characterization of GEP-NENs is necessary, especially to predict their evolution, our data could suggest paralogous 13 HOX genes expression combined with lncRNA HOTAIR as a new useful tool for the prognostic definition and management of GEP-NENs patients.

Patients and Specimens
Thirty-four GEP-NEN patients admitted to the National Cancer Institute "Giovanni Pascale" of Naples, between 2015 and 2019, were recruited in this study.All patients had provided written informed consent for the use of tumor samples according to the institutional regulations and the study was approved by the ethics committee of the National Cancer Institute "G.Pascale"(Project name: "Fighting Cancer resistance: Multidisciplinary integrated Platform for a technological Innovative Approach to Oncotherapies (Campania Oncotherapies)"-Ethical approval n.40/19 OSS-Date 21 November 2019).All cases have been reviewed by two pathologists (FT, MDB) and graded and staged according to WHO 2017 and 2019 classification criteria (NET-G1, NET-G2, NET-G3, NEC-G3) on tissue sections.The 4 main categories are distinguished on the basis of the proliferative activity, measured through the mitotic count and the Ki67 expression.Medical records have been reviewed for clinical information, including histologic parameters, assessed on standard H&E-stained slides combined with immunohistochemical staining with neuroendocrine markers (Cromogranin, Synaptophisin, CD56), and tumor location.

Evaluation of Immunostaining
Antigen expression was independently evaluated by two experienced pathologists (FT/MDB) using light microscopy.For paralogous 13 HOX genes nuclear and cytoplasmic localization were considered.All values of immunostaining were expressed only in percentage terms of positive cells.The percentage of positive cancer cells was evaluated in each sample by counting the number of positive cells over the total cancer cells in 10 non-overlapping fields using ×400 magnification as previously described [15].

RNA In Situ Hybridization Assay (RNA ISH)
In situ detection of HOTAIR was performed using the RNAscope (RNAscope ® 2.5 HD Detection Reagent-BROWN User Manual) according to the manufacturer's instructions.The tissue sections were boiled at 95 • C for 30 min in Target Retrieval solution.Protease treatment was then applied at 40 • C for 40 min.Then, we hybridized the Hotair probe for 2 h at 40 • C. The detection kit (BROWN) was used to amplify and reveal the signal, according to the manufacturer's instructions.The assay was performed using two controls: peptidylprolyl isomerase B (cyclophilin B) (PPIB) mRNA, a positive control and duplex negative control probe (dapB), a negative control.The slides were independently evaluated by two separate observers (FT/MDB).Positive staining was indicated by signals as brown dots present in the nucleus and/or cytoplasm.The number of signal staining was counted in 60 tumor cells.HOTAIR has an expression level varying between 0 to >10 copies per cell.We used a semi-quantitative scoring utilizing the estimated number of dots present within each cell boundary.We have categorized staining into 3 scores: High Score (>6 dots/cellclusters), Intermediate Score (>3 <6 dots/cell), Low Score (no staining or <3 dot/cell) (Figure 1).

Statistical Analysis
Non-parametric tests were used to compare independent groups of numerical data.Differences in the expression of HOXA13, HOXB13, HOXC13, HOXD13 according to age, gender, location and GEP-NEN categories were analyzed using Mann-Whitney U-test and Kruscall-Wallis tests The Pearson χ2 test was used in order to determine relation between Hotair expression and the variables included in the study.p < 0.05 was considered statistically significant.All tests used were two-tailed.All statistical analyses were carried out using SPSS version 20.0 software (SPSS, Chicago, IL, USA).Results are illustrated by boxplot graphs, a standardized way of displaying the distribution of data based on a five number summary ("minimum", first quartile (Q1), median, third quartile (Q3), and "maximum").

Figure 5 .
Figure 5. Box plot representation of the relation between HOTAIR expression scores and paralogous 13 HOX proteins expression: (a) HOXA13; (b) HOXB13; (c) HOXC13; (d) HOXD13.Each value that deviates from the central trend of the distribution is represented in the graph with the symbol °/*.

Figure 5 .
Figure 5. Box plot representation of the relation between HOTAIR expression scores and paralogous 13 HOX proteins expression: (a) HOXA13; (b) HOXB13; (c) HOXC13; (d) HOXD13.Each value that deviates from the central trend of the distribution is represented in the graph with the symbol • /*.

Table 3 .
Relation between lncRNA HOTAIR and tumor grading classification.

Table 3 .
Relation between lncRNA HOTAIR and tumor grading classification.