Transgenic mice overexpressing the LH receptor in the female reproductive system spontaneously develop endometrial tumour masses

The receptor for the luteinizing hormone (LH-R) is aberrantly over expressed in cancers of the reproductive system. To uncover whether LH-R over expression has a causative role in cancer, we generated a transgenic (TG) mouse which overexpresses the human LH-R (hLH-R) in the female reproductive tract, under the control of the oviduct-specific glycoprotein (OGP) mouse promoter (mogp-1). The transgene was highly expressed in the uterus, ovary and liver, but only in the uterus morphological and molecular alterations (increased proliferation and trans-differentiation in the endometrial layer) were detected. A transcriptomic analysis on the uteri of young TG mice showed an up regulation of genes involved in cell cycle control and a down regulation of genes related to the immune system and the metabolism of xenobiotics. Aged TG females developed tumor masses in the uteri, which resembled an Endometrial Cancer (EC). Microarray and immunohistochemistry data indicated the deregulation of signaling pathways which are known to be altered in human ECs. The analysis of a cohort of 126 human ECs showed that LH-R overexpression is associated with early-stage tumors. Overall, our data led support to conclude that LH-R overexpression may directly contribute to trigger the neoplastic transformation of the endometrium.

The proper functioning of the transgene construct was tested transfecting it into Hec1A EC cells, where the OGP promoter is known to be active 18 , and evaluating the light emitted by the bioluminescent Luciferine. Hec1A cells transfected with the mogpLuc2AhLH-R vector showed a higher amount of Relative Light Units (RLU) compared to that of cells transfected with the pBluescript SK (+) empty vector (Fig. 1B). The proper expression of hLH-R on the plasma membrane of Hec1A cells transfected with the mogpLuc2AhLH-R construct was confirmed by IF, using a specific antibody against c-myc (Fig. 1C).
The mogpLuc2AhLH-R transgene, devoid of the plasmid backbone, was microinjected into the male pronucleus of mouse zygotes and the resulting puppies were screened by PCR analysis on tail DNA as detailed in "Materials and methods" section. Two out of five puppies turned out to be positive at the screening, one female and one male (Fig. 1D). Both mice turned out to be able to correctly transmit the transgene to their offspring, hence generating two TG lines: FVB-Tg(MOGP-hLH-R)100 and FVB-Tg(MOGP-hLH-R)200, hereafter abbreviated TG-hLH-R-frt-100 and TG-hLH-R-frt-200, respectively. Mice of the two transgenic lines obtained from either founder were maintained in heterozygosity in FVB background. The transgene appeared to be integrated in a head-to-tail tandem array with a higher copy number in the TG-hLH-R-frt-100 line compared to TG-hLH-R-frt-200 (Supplementary Figure S1). Both TG lines were fertile, with a mean number of born puppies similar to those obtained in wild type (WT) mice, with no changes in the number of litters over time ( Fig. 1E; Supplementary Figure S2). Furthermore, the two TG lines had a similar number of follicles in the ovaries (Fig. 3), which is considered an indicator of intact fertility (see below). Anyway, the TG-hLH-R-frt-100 line was lost after 3 years.
The expression of the transgene was quantified by Quantitative Real Time (RQ-) PCR, determining the amount of hLH-R in the RNA extracted from different tissues of 3 months-old female mice. In agreement with what shown by Miyoshi for the mogp promoter 18 and in the website (http:// www. infor matics. jax. org/ marker/ MGI: 106661) for endogenous Ovgp1 expression, the transgene turned out to be highly expressed in the uterus and ovary, as well as ectopically expressed in liver and spleen of TG mice compared to WT animals ( Fig. 2A-D; raw data are in Supplementary Table S1). Nevertheless, no gross phenotypic alterations (such as hepato-splenomegaly, jaundice etc.) which could be related to such ectopic expression emerged. At difference from the other organs, in the ovary we found a significant basal expression of hLH-R (by RQ-PCR), possibly due to the partial overlap of the primers used for RQ-PCR with the mouse LH-R sequence. The expression of the hLH-R protein encoded by the transgene was confirmed both in the ovaries and in the uteri of either TG lines by IHC using anti-c-myc antibodies ( Fig. 2E-J). The analysis of the IHC score (i.e. the product between the intensity and the percentage of positive cells) showed a very high score in both the ovary and uterus of both TG lines (Fig. 2K).
Morphological and immunohistochemical characterization of TG-hLH-R-frt mice. Based on the above expression data, we performed a morphological characterization of both the ovaries and the uteri of TG (from either transgenic lines) and WT mice at different ages: young (3-12 months) and old (> 12 months) aged mice. We did not detect any gross morphological or histological alteration in the ovaries of mice from either TG lines at any ages (representative pictures of > 12 months mice are in Fig. 3A; representative pictures of < 12 months mice are in Supplementary Figure S3). In particular, the two TG lines had a similar number of follicles in the ovaries (Fig. 3a), which indicates the preservation of ovulation capacity, hence suggesting the maintenance of proper fertility. We then analyzed the uteri from mice of either TG lines, quantifying uterine morphometry taking into account: the longitudinal and transversal uterus lengths ("Y" and "X" values in Fig. 3b, respectively), the uterine radius (UR), the inner circular muscle (ICM) and the height of the luminal epithelium (LEH) (Fig. 3B, b) as in Wood et al. 20 . While no differences in longitudinal and transversal uterus lengths, or in LEH emerged between TG and WT controls at young ages, the uteri of young TG mice (of either lines) showed an increased (although not statistically significant) mean UR and thickness of the ICM, compared to WT mice (Fig. 3B). TG and WT mice older than 12 months had similar UR and LEAH values, while the mean ICM thickness was higher in TG mice, although not statistically significant ( Fig. 3B and Supplementary Figure S4). Two mice older than 12 months belonging to the TG line with the greater expression of the transgene (i.e. TG-hLH-R-frt-100) showed an increased (although not statistically significant) size of the uteri internal cavity, compared to age matched WT mice (Supplementary Figure S5). Because of the similarity of uterine features in either TG lines, we used these lines interchangeably hereinafter in our studies.
The endometrial layer in the uteri of TG mice was then better characterized by IHC analysis, evaluating Ki67 staining, to determine the extent of cell proliferation, cytokeratin 8 (CK-8, an epithelial cell-specific marker) and α-smooth-muscle actin (α-sma, a marker of stromal cells), to assess the state of cell differentiation 21 . Compared to WT mice, 33% (2 out of 6) of TG-hLH-R-frt mice were positive to CK-8 (Fig. 3C, c). The glandular epithelial and stromal cells of 6 months-old TG mice showed a statistically significant higher percentage of Ki67-positive cells, compared to WT animals (Fig. 3D, d). An increased Ki67 staining was observed in the luminal epithelial cells of TG mice older than 12 months, while it was no more evident in the stroma and glandular epithelium. Furthermore, the uteri of all (8 in total, randomly chosen) TG mice of both age groups (3-12 and > 12 months) showed a positive α-sma staining muscle and glandular epithelial cells (Fig. 3E, panels on the right). The glandular nature of α-sma positive structures was confirmed by their positivity to FOXA2 and negativity to CD31 staining (Fig. 3E′,E″). On the contrary, WT mice showed a significant staining in smooth muscle cells and a scanty signal in blood vessels (Fig. 3E, panel on the left).
Overall, IHC data corroborated morphological results, indicating the occurrence of epithelial hyperplasia and stromal trans-differentiation of epithelial cells, in the uteri of transgenic mice.
Transcriptomic characterization of the uteri of TG-hLH-R-frt mice. Since the uterus was apparently the main organ affected by LH-R over expression, we studied in detail this organ, performing a whole transcriptomic analysis of the uteri of two young (6-months old) TG (belonging to the TG-hLH-R-frt-200 mouse The Functional annotation analysis (FAA) performed using the NIH-DAVID database, revealed that the majority of the upregulated genes belong to four functional categories, e.g. "terms": ATP binding, microtubule, protease inhibitor and Ubl conjugation. We also analyzed the DE genes according to their possible role in known pathways (Fig. 4A). We found upregulated genes ( Fig. 4A and Supplementary Table S2) involved in the p53 pathway, as well as in cell cycle pathway, such as cyclins (Ccna2, Ccnb1, Ccnb2) and members of the cyclin-dependent kinases family (such as Cdk1). Other upregulated genes belong to biological processes such as: progesterone-mediated oocyte maturation, steroid biosynthesis and focal adhesion. The FAA of down regulated genes ( Fig. 4A and Supplementary Table S3) showed that 82 (30%) genes were related to either the AMPK (i.e. protein phosphatase 2A gene, PPA2) and the PPAR signaling pathways. Interestingly, some of the downregulated genes codify proteins involved in protective mechanisms against damage or in toxic compounds metabolism: the metabolism of xenobiotics by P450, drug metabolism-P450 cytochrome, and chemical carcinogenesis. In addition, some downregulated genes belong to the chemokine signaling, such as Cxcl1, Cxcl10, Cxcl15, Cxcr2, Ccl17, Ccl24 and Ccl4. Eight of the genes belonging to the most DE pathways (cell cycle, FoxO signaling pathways, focal adhesion, p53 signaling pathways, chemokine signaling pathways, metabolism of xenobiotics by cytochrome P450 and damage protection from chemical carcinogenesis) of TG-hLH-R-frt mice were then validated by RQ-PCR (Fig. 4B). A statistically significant correlation was obtained between Log2 fold change values from microarray data and Log2 fold change values from RQ-PCR (p = 0.0032, p value by two-tailed t test), with a R square = 0.85 (R = Pearson correlation coefficient). In the same samples a high expression of the h-LH-R transgene, evaluated by RQ-PCR with primers that recognize both human and mouse orthologs, also emerged, further confirming dataset validity (Fig. 4B).
Overall, transcriptomic data further support that LH-R overexpression in the uterus dysregulates the expression of genes involved in cell proliferation, differentiation as well as damage protection and inflammation. This was also confirmed by comparing the transcriptomic data of our TG mice with FOXO1 uterine KO 22 and progesterone (Wnt7acre/+mPgrALsL/+) overexpressing models 23 (see the complete analysis in Supplementary  Table S4), Some DE pathways (e.g. cell cycle, inflammation and cell-cell contact) turned out to be shared by the three TG mice (Fig. 4A, highlighted in red). The "small molecules biochemistry related pathway" turned out to be deregulated both in our TG model and in the progesterone overexpressing model (Fig. 4A, highlighted in blue). Pathways related to the occlusion of vessels and arteries were downregulated in the FoxO KO model, which is comparable with the downregulation of the complement and coagulation cascades pathways observed in our LH-R over-expressing mice (Fig. 4A, highlighted in green).
Aged female TG-hLH-R-frt mice develop uterine tumor masses.. In two aged (17-months-old) female TG-hLH-R-frt mice from either TG lines (TG-hLH-R-frt-105 and TG-hLH-R-frt-200) we observed signs of suffering such as slow motility and rough fur. Animals were then euthanized and the abdominal cavity examined, showing the presence of large masses (3 cm 3 in the TG-hLH-R-frt-105 mouse and 2.8 cm 3 in the TG-hLH-R-frt-200 mouse) in the lower abdomen, at the uterine level. Moreover, one 17-months-old mouse, belonging to the line LH-R-100 (TG-hLH-R-frt-123) showed a smaller-size mass (45 mm 3 ) in the lower-third of the left uterine horn when randomly sacrificed for the routine characterization analysis (Supplementary Figure S6). The histology of the large-sized masses of TG-hLH-R-frt-105 and TG-hLH-R-frt-200 mice showed a poorly differentiated tissue and the loss of the typical uterine architecture (Supplementary Figure S6A, B). Both masses were composed of cells with big and highly basophilic nuclei at the H&E staining. The small mass of TG-hLH-Rfrt-123 showed a loss of the normal uterine architecture, although the inner circular muscle layer was still visible (Supplementary Figure S6C). Overall, 3 out of 9 (33%) female TG mice who reached an age > 17 months showed masses at the uterine level. All the masses positively stained with CK-8, indicating their origin from endometrial epithelial cells (Fig. 5A). The mass derived from the TG-hLH-R-frt-105 showed a slight diffuse staining, while  www.nature.com/scientificreports/ that of the TG-hLH-R-frt-200 showed a more intense and homogeneous staining (Fig. 5A). In the mass derived from the mouse TG-hLH-R-frt-123 the staining was localized in areas resembling glandular structures, while was absent in surrounding tissues (Fig. 5A). Applying the scoring system described in "Materials and methods" section, all the tumor masses showed a higher mean score compared to WT (Fig. 5a). As expected, all the tumor masses positively stained for c-myc ( Fig. 5B) with high scores (Fig. 5b). This indicates that hLH-R is indeed expressed in the tumor masses.
Overall, the histology of the masses arising in aged transgenic females is suggestive of a malignant endometrial tumor mass, i.e. an EC.
We then performed a transcriptomic analysis of the tumor mass originated in the TG-hLH-R-frt-200 mouse, comparing the gene expression profile (GEP) of the tumor mass with that of the uterus of an aged-matched TG-hLH-R-frt mouse (belonging to the same TG-hLH-R-frt-200 line), but with no evidence of tumor mass. The modified p value less than 0.01 was used as threshold. Than a gene was assumed to be DE when the fold change (log2 fold change) was ≥ 2. Applying these thresholds, a list of 1701 DE genes was obtained, 718 upregulated and 983 downregulated. The FAA led to identify the most altered biological processes: DE genes were significantly associated to 20 terms with a false discovery rate (FDR) p value < 0.05 (Supplementary Table S5 and S6). The most significant terms associated to upregulated genes were related to the membrane term, as well as to terms related to inflammatory response, positive regulation of ERK1 and ERK2 cascade, signal transduction, chemokine-mediated signaling pathway and positive regulation of angiogenesis (Table 1a). On the other hand, downregulated genes were associated to terms related to epithelial cell differentiation, positive regulation of cell migration, positive regulation of epithelial cell proliferation, focal adhesion, positive regulation of angiogenesis, apical plasma membrane and calcium ion binding (Table 1b). Notably, we detected the downregulation of two genes encoding proteins essential for the establishment of cell-cell contacts: CTNNAL1 gene (α-catenin) and CDH2 gene (N-cadherin) (Supplementary Table S6), and of the genes encoding the transforming growth factor beta receptor III (TGFBR3), LAMC2 (laminin-γ2) and MSX1 (Msh homeobox) (see the genes highlighted in bold in Supplementary Table S6), which are known to be downregulated in human EC 23,24 . The most relevant DE genes are indicated in the "most relevant genes" column in Table 1. Furthermore, we validated the identified signature of the tumor mass of TG-hLH-R-frt-200 mouse comparing it with the gene expression profile of other ECs obtained from publicly available datasets deposited into the GEO databases. We chose the datasets E-MTAB-2532, E-GEOD-56026, GSE32507, GSE24537, E-GEOD-23518, E-GEOD-17025, E-GEOD-2109, analysed in O'Mara 2016 and the datasets GSE17025, GSE63678 and GSE35794, analysed in Liu 2018 26,27 . Two out of the 1701 DE genes found in the tumor mass of the TG-hLH-R-frt-200 mouse matched with the 10 more downregulated genes identified by Liu 2018 in EC, including Tgfbr3. When compared with the 145 more deregulated genes from O'Mara 2016, 11 commonly deregulated genes emerged, including Sox17 and Esr1 (the whole analysis is shown in Supplementary Table S9, and the most relevant genes in common with the above datasets are highlighted in red in Table 1). Overall, the transcriptomic analysis of the masses arise in aged TG females supports the conclusion that they are endometrial tumor masses.
We further validated the upregulation of signaling pathways controlling cell proliferation and angiogenesis, through an IHC analysis of pAkt, ERK1, VEGF, Ki67 and p53. All the tumor masses overexpressed the above markers, and a statistically significant positive correlation emerged between c-myc and VEGF, ERK, pAKT, Ki67 www.nature.com/scientificreports/ and p53 score (Fig. 5C). Interestingly, the ERK and VEGF expression profile better matched the c-myc expression profile (i.e. the hLH-R overexpression) in the same tumor mass (Fig. 5C). Among the DE genes, the "membrane" term contained the vast majority of the upregulated genes. Among them, we observed several genes encoding for ion channels (e.g. KCNK13, CACNA1F, TRPV2, P2RX4, P2RX7) and transporters (e.g. SLC2B, SLC 7A7, SLC 11A1, SLC 15A, ABCA1, ABCA9, ATP1A3, ATP13A2). Moreover, we performed an IHC analysis on the tumor endometrial masses arising in TG-LH-R-frt mice, using a hERG1 specific antibody, which also recognizes the mouse ERG1 27 . It emerged that while the uteri of either WT or TG did not express the potassium channel, all three tumor masses arising in TG showed a high level of expression of hERG1 (Fig. 5D, d). Interestingly, a statistically significant positive correlation emerged between c-myc Tag (which indicates the expression of the transgene) and hERG1 score ( Patients with endometrial cancer (EC) express LH-R: clinico-pathological correlations. Based on the above data and on the previously described expression of LH-R 10-12 and Kcnh2 28 in human primary ECs, we performed a transcriptomic analysis (applying RQ-PCR analysis) on a cohort of 126 patients with EC of various stages and grades (Raw data are in Supplementary Table S10). All the EC samples turned out to express the LH-R at high levels with a median value of 73.78 (folds). LH-R expression levels significantly associated with risk (high LH-R expression and low risk; p = 0.025) and myometrial invasion (high LH-R expression and myometrial infiltration less than 50% of myometrial depth; p = 0.014) ( Table 2). In addition, most of the EC samples turned out to express the Kcnh2 at high levels with a median fold value of 79.6. Statistically significant correlations were found between Kcnh2 high expression and low FIGO stages (p = 0.01), low risk endometrial cancer (p = 0.019) and infiltration less than 50% myometrial depth (p = 0.036) ( Table 2). Finally, LH-R expression significantly associated with the expression of Kcnh2 (p = 0.006, Fisher's exact test; Spearman Index = 0.245).

Discussion
In the present study we generated TG mice over-expressing the human form of the LH-R encoding gene in the female reproductive tract (TG-hLH-R-frt mice), to study the role of LH-R mis-and over-expression in the initiation and progression of cancers of the female reproductive system, in particular of EC.
To drive the expression of the hLH-R in the female reproductive tract, as well as to better mimic the expression levels reached by the LH-R encoding gene occurring in the human setting 18 , we exploited the mogp-1 promoter. Furthermore, the 2A peptide 19 was included into the construct to drive an equimolar expression of the hLH-R cDNA and of the luciferase gene, a device to monitor the proper expression of LH-R, which we exploited during the initial check of the appropriateness of the construct once transfected into Hec1A cells (see Fig. 1).
TG-hLH-R-frt mice were vital and displayed a normal fertility, at least for four generations. One of the two TG lines, the TG-hLH-R-frt-100 expressing the hLH-R mRNA at higher levels, however, experienced a decline in fertility after 3 years, and was lost afterwards.
A high expression of the transgene (witnessed by the expression of the hLH-R mRNA and by the immunohistochemical detection of the c-myc tag) occurred in the ovaries and the uteri of transgenic females, as expected on the bases of the reported expression of the mogp-1 promoter 18 . An ectopic expression in the liver and spleen was also observed, which however did not produce any pathological phenotype. It is worth noting that ectopic expression of transgenes can often occur as a result of the position effect from random integration into the genome 29 . In addition, despite the very high hLH-R expression in the ovaries, we did not observe gross morphological or functional alteration in this organ, in particular no difference in the number of follicles, which could explain the lack of effects on fertility.
On the contrary, TG-hLH-R-frt females showed alterations in the proliferation and differentiation state of the endometrial layer, which were more evident in aged mice. Indeed, both epithelial, either glandular or luminal, and stromal cells of the uterine mucosa had a high Ki67 staining, and an aberrant expression of α-sma, which is normally expressed in myofibroblasts, but absent in the glandular epithelium 21 . Overall, signs of hyperplastic growth and trans-differentiation were evident in the uteri of TG-hLH-R-frt mice, starting from 3 months and were maintained afterwards (Supplementary Table S7).
A whole transcriptomic analysis, performed on the uteri of young (6 months old) mice showed the deregulation of several genes involved in cell proliferation, differentiation as well as damage protection and inflammation. In particular, a high expression of genes involved in the p53 pathway, as well as in the regulation of the cell cycle and proliferation emerged. The upregulation of the p53 pathway related genes merits attention, since mutations of p53 are common in Type II EC, while are late events and hence apparently only contribute to progression of Type I ECs 30 . The upregulation of cyclins together with cyclin-dependent kinases family are suggestive of a possible dysregulation in cell cycle, control of cell death and DNA damage repair, hence predisposing cells to malignant transformation 31 . Between the downregulated pathways, AMPK signaling pathways is noteworthy. This pathway is involved in regulating growth and reprogramming metabolism and includes the downregulated protein phosphatase 2A gene (PPA2), which is one of the four major Ser/Thr phosphatases, and it is implicated in the negative control of cell growth and division 32 . Another downregulated pathway is the PPAR signaling pathway. PPAR pathway plays essential roles in the regulation of cellular differentiation, development, and metabolism 33 . Moreover, it was observed a lower expression of genes involved in protective mechanisms against toxic compounds (such as cytochrome P450 family genes) and genes belonging to chemokine signaling, suggesting a reduced immune response. Both the inability of metabolizing toxic compounds and the impairment of the native and adaptive immune system are signs of vulnerability and predisposition to cancer. The comparison between TG-hLH-R-frt mice with FOXO1 uterine KO and progesterone overexpressing mice, evidenced commonly altered processes, such as the deregulation of pathways related to the inflammatory response process and cell survival, thus reinforcing the suggestion of a role of LH-R in the FoxO and progesterone signaling pathways and thus in the development of ECs.
Considered together, all the dysregulated pathways emerging from the transcriptomic analysis are suggestive of an increased cell proliferation, of an alteration of epithelial differentiation accompanied by a lower ability of the cells to respond to damage, including cancerogenic insults.
Consistently, 33% of old (> 17 months) TG-hLH-R-frt female mice spontaneously developed masses at the uterine level, which resemble human ECs (see Supplementary Table S5). Indeed, these masses showed a poorly differentiated tissue and the loss of the typical uterine architecture, and were then interpreted by us as ECs. This was confirmed by the high expression of CK-8 as well as by the results of transcriptomic analysis. EC is characterized by many genetic and molecular alterations, in particular those concerning proteins responsible for signal transduction and cell adhesion 34 . Consistently, we found the downregulation of N-cadherin (CDH2) and α-catenin (CTNNA1) genes in the tumor mass of TG-hLH-R-frt-200 mouse, according to what occurs in human ECs 35 . In particular, both cadherins and catenins are closely involved in cell adhesion, and α-catenin is involved in the epithelial-mesenchymal transition (EMT). A decreased expression of epithelial markers (such as E-cadherin and α-catenin) has been detected in human EC tissue 36 . Hence, the dysregulation of both cadherins and catenins represents a central aspect in the aggressive behavior of EC. Furthermore, in the tumor masses arising in TG mice, we detected the downregulation of LAMC2 and MSX1 (laminin subunit gamma 2-and MSH homeobox 1-encoding genes, respectively), which are also downregulated in human poorly differentiated EC 24 . Finally, the transcriptomic analysis performed in the tumor masses arising in the uteri of old female TG mice, showed the dysregulation of genes encoding angiogenic factors, such as the vascular endothelial growth factor (VEGF), frequently dysregulated in ECs 37 . We also validated the identified signature of the tumor mass of TG-hLH-R-frt-200 mouse comparing it with the GEP of other endometrial cancer obtained from publicly available datasets deposited into the GEO database. Interestingly, some common deregulated genes emerged. For example, the downregulation Tgfbr3, whose reduced expression has been demonstrated in multiple types of human cancer www.nature.com/scientificreports/ including breast, prostate, ovarian, pancreatic, non-small cell lung cancer and renal cell carcinoma 38 . Among other deregulated genes, the downregulation of Sox17, identified as tumor suppressor in endometrial cancer 39 and the deregulation of Esr1, involved in epithelial-mesenchymal transition (EMT) and considered as prognostic marker for endometrial cancer 40 , which confirm our results merit further considerations. This comparison with other endometrial cancer gene signatures validates the clinical relevance of gene expression profile here identified. We confirmed some of the deregulated pathways emerging from the transcriptomic analysis by IHC. In particular, a statistically significant positive correlation emerged between the expression of hLH-R (witnessed by c-myc expression) and VEGF, ERK, pAKT, Ki67 and p53 staining. The upregulation of pAkt is particularly interesting, since it confirms what described in human Type I ECs, where the most common genetic mutations are detected in the Pten (phosphatase and tensin homolog) gene 41 , and what shown in Pten -/-mice 21 . Finally, an interesting correlation also emerged between hLH-R expression and the overexpression of the hERG1 potassium channel in the tumor masses, confirming data obtained in several human cancers 28 , including EC 25 . The clinical relevance of LH-R over-expression in EC and the correlation between LH-R and Kcnh2 (hERG1) mRNA expression in primary human ECs was strengthen evaluating LH-R and Kcnh2 mRNA expression by RQ-PCR in a cohort of 126 human EC specimens. A significant positive correlation between the expression level of these two genes emerged. Moreover, significant correlations were found between the high expression of LH-R and Kcnh2 and low-risk EC subgroup and between the high expression of the two genes and invasion less than 50% myometrial depth.
Taken together, these data corroborate the main finding obtained in the present work, i.e. the causative role of LH-R over expression in the development and malignancy of EC. In particular, LH-R overexpression apparently leads to an increased probability to develop spontaneous endometrial tumor masses. Whether LH-R overexpression could act making mice more susceptible to external carcinogens, including hormones, or could trigger per se the cancerogenic process, by activating pro-neoplastic signaling pathways, remains a matter of fact, to be further deciphered. In this scenario, taking into account the association between LH-R and hERG1 expression, we could hypothesize that plasma membrane receptors (in our case LH-R) and hERG1 cooperate in triggering intracellular signaling pathways in EC as previously shown in several cancer types 42 . Besides speculations, the targeting the of the LH-LH-R axis may represent a valuable therapeutic strategy in EC.

Materials and methods
Production of the transgene construct. The mogpLuc2AhLH-R construct was assembled starting from the following fragments:   Collection, Manassas, VA), and retro-transcribed into cDNA, using the procedure described below, and the primers reported in Supplementary Table S8. For PCR amplification, the Phusion High-Fidelity DNA polymerase (Finnzymes, New England Biolabs) was used, and the touchdown protocol starting from 70 °C with a 2 °C decrease each 4 cycles, till 52 °C was applied, as above.
At the 3′ end of hLH-R cDNA, the sequence encoding the c-myc epitope was placed by PCR, inserting the myc sequence into the primer downstream to the hLH-R. The c-myc epitope was useful for easier identification of LH-R expressed by the transgene.
The list of the primers used for all the PCRs are listed in Supplementary Table S8. The resulting construct (mogpLuc2AhLH-R) was assembled in the pBluescript SK (+) vector.

Microinjection into oocytes and generation of transgenic mice.
The 8000-bp mogpLuc2AhLH-R construct was excised with NotI restriction enzyme from the pBluescript SK(+) vector and microinjected into the male pronucleus of fertilized zygotes from FVB mice. Fertilized eggs were re-implanted into the oviduct of pseudo pregnant mice according to standard procedures 43     www.nature.com/scientificreports/ exon boundaries, but are not able to distinguish human from mouse LH-R orthologs. The primer sequences for hLH-R are: 5′-TGC CTA CCT CCC TGT CAA AG-3′ forward primer; 5′-TTG AGG AGG TTG TCA AAG G-3′ reverse primer. β-actin gene was used as housekeeping gene. The primer sequences for β-actin are: 5′-GGG GTG  TTG AAG GTC TAA A-3′ forward primer; 5′-GAT CTG GCA CCA CAC CTT CT-3′ reverse primer 27 . For RQ-PCR on human samples the expression levels of hLH-R and Kcnh2 were normalized to the levels of GAPDH housekeeping gene. For hLH-R primers we used the same described above. The sequences of Kcnh2 and GAPDH primers are the following: GAPDH sense 5′-GCT CTC GCT CCT CCT GTT -3′; GAPDH antisense 5′-ACG ACC AAA TCC GTT GAC TC-3′; Kcnh2 sense 5′-ACG TCT CTC CCA ACA CCA AC-3′; Kcnh2 antisense 5′-GAG TAC AGC CGC TGG ATG AT-3′ (used at final concentration of 300 nM). Primers used for Kcnh2 are previously described in Pillozzi et al., 2018 44 .
PCR program was programmed with an incubation at 95 °C for 10 min, followed by 40 cycles of amplification: denaturation at 95 °C for 15 s and annealing-extension step at 60 °C for 1 min as described in Noci et al. 10 . The relative quantification of hLH-R expression levels was performed by the ΔΔCt method. Melting curve analysis were performed to exclude the amplification of a specific products or primer-dimer artefacts. Each reaction was performed in triplicate. Relative expression values of hLH-R were normalized for the expression value of Hec1A cells (low expression cell line).
Gene expression analysis. Gene expression analysis was performed using Agilent one color Gene Expression slides with 44 k resolution (mouse GE 4X44K V2 microarray kit, Agilent Technologies). Purified RNA samples were prepared as described in Masselli et al. 45 Raw data files obtained from the scanned slides were processed using Agilent Feature Extraction Software. Bioinformatic analysis were performed using the open source Rstudio software applying the Linear Models for Microarray Data ("LIMMA" package, Bioconductor) by Genomix4Life srl (Baronissi, Salerno, Italy). Each spot was first corrected for background intensity using the "normexp" approach of the background correct function. The between-array normalization and the quantile approach were performed according to the LIMMA package's functions. To identify DE genes an average cut-off of twofold changes was considered. The identified DE genes were addressed as specific for the tumor when they were not found to be deregulated neither in the uteri from healthy TG mice or in the uteri from WT animals. Gene expression data discussed in this paper have been deposited in NCBI's Gene Expression Omnibus and are accessible through GEO Series accession number GSE163488. The specimen characteristics of corresponding datasets in GSE163488 are reported in Supplementary Table S11. Bioluminescence assay. Two days after cell transfection we evaluated the light emitted by bioluminescent Luciferine. Luminescent data were obtained from cells either transfected with pBluescript SK(+) vector containing the mogpLuc2AhLH-R transgenic construct and transfected with the empty vector. Luciferine undergoes a luciferase-catalyzed oxidation resulting in an excited state that emits upon decaying to its ground state. The resulting sample light output was measured by using a current-measuring luminometer that read in arbitrary light units, usually referred to as "Relative Light Units" (RLU).

Immunofluorescence (IF).
IF was performed on Hec1A cells previously transfected with the vector containing the mogpLuc2AhLH-R sequence. The anti c-myc antibody (4 μg/ml, Santa Cruz Biotechnology) was used as primary antibody, and the anti-mouse Alexa488 (1 μg/ml, Invitrogen, Thermo Fisher) as secondary antibody. Images were acquired with Nikon D-Eclipse C1 (Nikon) confocal microscope.
Analysis of uterine morphometry. Mice were euthanized by inhalation of 100% CO 2 , the whole reproductive tract of female mice was excised and immediately fixed in 4% formaldehyde for 4 h, processed and embedded in paraffin following standard procedures. Lengthwise sections of 6 µm thick were prepared and put on positive-charged slides. Samples were stained with Hematoxylin and Eosin (H&E) following a standard protocol. The uterine radius (UR) was measured from the outer longitudinal smooth muscle layer (myometrium) to the apical surface of the luminal epithelium. The muscle layer was considered the inner circular layer. The luminal epithelial height (LEH) was measured from the basement membrane to the apical surface as described in Wood et al. 20 All measurements were performed using a light microscope (Leica DMR, Germany) equipped with Leica DC Viewer and Leica Qwin software. The evaluation and measures were performed on the sample displaying the whole uterine cavity and at least three measurements per area were determined.

Immunohistochemistry (IHC).
To block sample's endogenous peroxidase, 1% H 2 O 2 solution in phosphate-buffered saline was usen on dewaxed and dehydrated tissue slides. Antigen retrieval was performed by using different procedures: (1) dipping the samples in citrate buffer pH 6.0 and heated by microwave oven at www.nature.com/scientificreports/ 600 W for 12 min (for c-myc, Ki67, CK8 and α-sma staining) and (2) treating with proteinase K (5 μg/ml) in PBS at 37 °C for 5 min (for hERG1 staining). Samples were permeabilized with a 0.1% Triton X100 in UltraVBlock solution (LabVision) (for c-myc, Ki67 and α-sma staining) and incubated overnight at 4 °C with the following primary antibodies: anti-c-myc (monoclonal antibody, Santa Cruz Biotechnology, Santa Cruz, CA, 1:100), anti-cytokeratin-8 (Developmental Studies Hybridoma Bank, Iowa City, IA 1:100), anti-KI67 antigen (Dako, 1:50), anti α-sma (Dako, 1:100) and anti hERG1 (monoclonal antibody, MCK Therapeutics, 0.005 μg/μl). IHC was performed with commercially available kit (PicTure-Max polymer Detection kit, Invitrogen) according to manufacturer's instruction. Hematoxylin was used for nuclear counterstaining. Immunohistochemistry slides were scored by three independent operators (T.L., J.I. and E.L.). For each protein, a specific scoring system was applied. α-sma expression was evaluated estimating the percentage of immunoreactive cells; samples were classified as positive when the percentage of labeled cells was equal to or greater than one. Ki67 staining was evaluated taking into account the percentage of labeled nuclei in three different areas: the luminal epithelium, stromal cells and the endometrial glandular cells. Stained sections were analyzed at a total magnification of 40 × field by field, from top left to bottom right. Previously published scoring system was applied for c-myc and hERG1 50 . Samples were scored combining the estimate of the staining intensity with the percentage of positive cells. Staining intensity was rated on a scale of 0-3, with 0 = negative; 1 = weak; 2 = moderate, and 3 = strong. The final scoring data were obtained by multiplying the percentage of labelled cells by the staining intensity values, obtaining a value between 0 and 300 for each sample. The c-myc and hERG1 scoring system was applied also to cytokeratin-8 for all samples.
Statistical analysis. Statistical analyses were performed with OriginPro 8 (Origin Lab, Northampton, Massachusetts) on at least three independent experiments. The normality distribution of data the variance homogeneity were performed as previously described in Petroni et al. 51 For comparisons between two groups, Student's t test was used. In case of multiple comparisons, one-way ANOVA followed by Bonferroni's post-hoc test was performed (as in 51 ). All the data reported satisfy the assumptions of the tests. Test of normality distribution and variance homogeneity assumptions have been proper performed and used to choose the right test for compare groups.
For the human EC samples the statistical analysis the clinical-pathological parameters were categorized as fallows: histology = endometrioid vs non endometrioid; differentiation grade = low, medium, high; FIGO = I, II, III and IV; risk = low vs high; myometrial invasion = less than 50% versus greater than 50%. The associations between the values of expression LHR, Kcnh2 and clinical-pathological parameters were assessed applying the Spearman correlation test. p < 0.05 was considered statistically significant.