Frequent PIK3CA mutation in normal endometrial gland drives spheroid formation and may be involved in stem cell propagation

Abstract Recent studies reported the presence of oncogenic mutations in normal endometrial glands, but the biological significance remains unclear. The present study investigated the status of KRAS/PIK3CA driver mutations in normal endometrial glands as well as spheroids derived from single glands. The normal endometria of surgically removed uteri (n = 3) were divided into nine regions, and 40 endometrial single glands were isolated from each region. The DNAs of 10 glands in each region were extracted and subjected to Sanger sequencing for KRAS or PIK3CA driver mutations, while the remaining 30 glands were conferred to a long‐term spheroid culture, followed by Sanger sequencing. Immunohistochemical analyses of stem cell (Axin2, ALDH1A1, SOX9) markers were undertaken for spheroids. Sanger sequencing successfully detected oncogenic mutations of KRAS or PIK3CA in a single gland. Twenty‐five of the 270 glands (9.3%) had mutations in either KRAS or PIK3CA, and the mutation frequency in each endometrial region varied from 0% to 50%. The droplet digital PCR showed high mutation allele frequency (MAF) of PIK3CA mutation, suggestive of clonal expansion of mutated cells within a gland. Over 60% of the collected spheroids had PIK3CA mutations, but no KRAS mutations were detected. Immunohistochemically, spheroids were mainly composed of cells with stem cell marker expressions. High MAF of PIK3CA mutation in a single gland as well as frequent PIK3CA mutation in stem cell‐rich spheroids that originated from a single gland suggest the role of PIK3CA mutation in stem cell propagation. This information could improve our understanding of endometrial physiology as well as stem cell‐oriented endometrial regeneration and carcinogenesis.


| INTRODUC TI ON
Normal endometrium periodically exfoliates and regenerates in reproductive women. Its functional or genetic abnormalities are tightly associated with many gynecological disorders, such as endometrial cancer and endometriosis. [1][2][3] Recently, it has been revealed that pathogenic mutations of oncogenes occur not only in these disorders but in the normal endometrium. [4][5][6][7][8][9][10][11][12] However, their biological significance for the etiology of gynecologic diseases or abnormal conditions in menstrual women is uncertain. We previously reported that each endometrial gland exhibits a monoclonal growth pattern with regional diversity. 13 To explain the monoclonal growth of each endometrial gland, we have proposed a hypothesis in which stem-like cells are present probably in the basal layers in each gland, and the daughter cells occupy the whole gland with monoclonality, which is linked to the vigorous proliferation of endometrial glands in each cycle or the development of endometrial cancer when oncogenic mutations are added. 13,14 To test this hypothesis, in the present study we focused on the frequency and regional distribution in the entire endometrium of PIK3CA and KRAS mutations, which are representative driver genes for endometrial cancer that are also present in normal endometrial glands. 4 Furthermore, we observed the rates of spheroid formation in vitro from a single gland in each endometrial region and examined the diversity in the region as well as the types of oncogenic mutation. Finally, the immunohistochemical characteristics of these spheroids obtained by long-term stem cell culture were evaluated.

| Clinical specimens
The endometria in the proliferative phase of the menstrual cycle were examined from three perimenopausal women who underwent total hysterectomy with a diagnosis of uterine fibroids (Cases 1 and 2) or adenomyosis (Case 3). The protocol for acquiring and using Immediately after hysterectomy, a gynecologist macroscopically divided the endometrium into nine regions with a scalpel or scissors. A total of 40 endometrial glands were randomly picked up and collected with microscopic manipulation 13 from each region after collagenase treatment. Ten of these glands in each region were DNAextracted and subjected to mutational analysis for KRAS and PIK3CA using the Sanger method, and the remaining 30 were subjected to long-term spheroid cultures. To examine the clonality of the single glands sampled in our experiment, we used ddPCR for analyzing the MAFs of KRAS G12V and PIK3CA E542Q detected by the Sanger methods. Spheroids grown to a diameter of 2 mm or more were collected and subjected to mutational analysis in the same manner as for single glands. Immunohistochemistry was carried out to characterize the spheroid constituents and origin of the expression of endometrial epithelial marker PAX8, endometrial stem cell markers Axin2, ALDH1A1, and SOX9, and indicator of cell proliferation Ki-67.

| Spheroid culture
As described above, the normal endometrium was divided into nine sections. Thirty single glands collected under a microscope from each region were isolated. All single glands obtained from one section were cultured together in one well of a 6-well nonadhesive dish (#3471;

| DNA extraction
DNA was extracted by the alkaline method. Single glands or spheroids were collected from a nonadhesive dish with a sterile micropipette tip and suspended in 50 μL sterile water, then washed twice with 1 mL PBS. After adding 15 μL of 100 mM NaOH to the samples, they were heated at 95°C for 10 min, followed by adding 3 μL of 1 M Tris-HCl (pH 7.0). After centrifugation at 8000-10,000 g for 1 min, the supernatants were collected to prepare DNA samples.

| Droplet digital PCR
To confirm the clonality of the endometrial glands, MAFs were analyzed by ddPCR using each of the four glands containing KRAS

| Immunohistochemistry
The expressions of PAX8, Axin2, ALDH1A1, and SOX9 were evaluated by IHC analysis. The FFPE spheroid sections (3 µm thick) were dewaxed in xylene and hydrated in graded alcohol.
After antigen retrieval in sodium citrate buffer, the slides were incubated in overnight at 4°C with Abs at the following dilutions: For immunofluorescence staining, FFPE spheroids or endometrium slides were loaded into a glass slide holder and dewaxed in xylene and then hydrated with alcohol. Sodium citrate buffer pH 6.0 was used for antigen retrieval in an autoclave for 10 min followed by incubation with primary Abs at a dilution of Axin2 1:200 and Ki-67 1:100 overnight at 4°C followed by incubation with fluorescencelabeled secondary Abs for 1 h at room temperature. Slides were then counterstained with Fluoroshield Mounting Medium with DAPI (Sigma-Aldrich), and the immunofluorescence was detected using a Nikon Eclipse 50i fluorescence microscope with the appropriate filter.

| Sanger sequencing identified frequent KRAS/PIK3CA mutations in normal endometrium with regional diversity
We detected mutations of KRAS or PIK3CA in single endometrial glands using the Sanger method rather than NGS in consideration of cost benefits and utility in clinical practice. The present study is the first to use the Sanger method to detect oncogenic mutations in a single gland. Ten microscopically isolated single glands from each of the nine regions of the entire endometrium were subjected to sequencing ( Figure 1). KRAS or PIK3CA mutations were detected at different rates in each region (Figure 2A), some of which showed identical mutations in multiple glands, whereas others had no mutations detected, exhibiting region specificity. Figure 2B

| Spheroids derived from a single gland had frequent PIK3CA mutations
We next determined the source of oncogenic mutations in the glands. Considering the monoclonal composition of cells in a gland, it is possible that such mutations could be present in stem or stemlike cells. The existence of stem or stem-like cells in endometrial glands has been postulated, especially in the basal layers of the endometrium. 15 However, pure isolation of a single stem cell has not been accomplished, despite multiple potential markers having been reported. To reconstitute stem-rich populations, spheroid cultures have been proposed. 16 We thus grew spheroid cultures from a single gland isolated from each region of the endometria and attempted to identify the oncogenic mutations. The regional specificity of oncogenic mutations in the endometrium led us to expect that the efficiency of spheroid formation is higher in mutation-prone regions.
Thirty glands were isolated from each of the nine endometrial regions per patient, and long-term spheroid culturing was carried out for a total of 27 regions. Single endometrial glands were cultured with stem cell media in nonadhesive dishes. They first showed morphologically round shapes, after which most cells gradually became isolated, scattered, and eventually dropped to the bottom of the dish, while some cells remained floating, forming spheroid-like structures ( Figure 3A). Some spheroids gradually increased in size ( Figure 3B,C) and eventually became macroscopically visible ( Figure 3D). Each spheroid with a diameter of 2 mm or more after 3 months was isolated and subjected to Sanger sequencing. Table 1

| Cells in spheroids show nuclear Axin2 expression, a pattern similar to a rare population of basal cells in normal endometrium
Next, IHC was carried out to characterize the properties of the spheroids. First, the expression of ALDH1A1, Axin2, and SOX9, previously reported stem cell marker of endometrial epithelium, 17 was examined in spheroids after grown in stem-cell media for over 3 months. Figure 4A, the predominant expressions of cytoplasmic ALDH1A1, nuclear Axin2, and SOX9 were confirmed in all the spheroids examined, suggesting that spheroids were mainly composed of cells with stem-like characteristics through long-term culture. Next, to identify the origin of the components in the spheroids, PAX8 18,19 expression was examined for spheroids obtained from region F of

| DISCUSS ION
Our data show that oncogenic mutations of KRAS/PIK3CA are common in the normal endometrium of perimenopausal women and can be detected by the Sanger method (Figure 1). Previous studies using Sanger sequencing have reported that such oncogenic mutations are absent in normal endometrium, unlike endometrial cancer or its precursors. 20,21 The reason for this discrepancy might be due to our unique methodology using microscopic isolation of single glands, which can increase the purity of the glandular components while minimizing contamination with stromal components. As a single gland is known to be mainly composed of monoclonal cells, mutant cells are likely to account for a considerable portion of a whole gland, potentially enabling detection by Sanger sequencing.
In fact, the ddPCR suggested clonal expansion of mutated cells, especially for PIK3CA mutations with 16% of MAFs.
Of note, pathogenic mutations occurred territorially ( Figure 2). In support of our results, a somatic mutation analysis of cancer-related genes using NGS for 1311 endometrial glands from 37 women has also shown that multiple glands with the same somatic mutation occupy a substantial area of the endometrium. 22 Interestingly, recent studies focusing on the 3D structure of the endometrium by in vivo were PIK3CA mutations, not KRAS, and the detection rate of mutations in spheroids was much higher than that of a single gland. Taken together, these findings indicate that cells with the PIK3CA mutation have a prominent ability to form spheroids in a nonadhesive dish environment, unlike with KRAS mutations.
Many theories have been reported so far regarding the character of stem cells in the endometrial epithelium. [25][26][27][28] However, the roles of epithelial stem cells in endometrial regeneration and uterine tumorigenesis are controversial. Therefore, the clinical application of epithelial stem cells remains unresolved. 25 Here, we focused on Axin2 as a protein that negatively regulates the β-catenin pathway.
Axin2 has long been known to be expressed in the basal layer of the endometrium 15 and has been reported to be a marker for endometrial epithelial stem cell regeneration and endometrial cancer. 29 We found that spheroids produced by long-term stem cell culture express Axin2 in the nucleus, not cytoplasm. They also expressed nuclear SOX9 and cytoplasmic ALDH1A1, known epithelial stem cell markers ( Figure 4).  mutations, probably localized in the basal layers, continue to provide daughter cells with the same mutations even after shedding so that they could account for a considerable portion of a whole gland or extend to neighboring glands. 1 Supporting this theory, a recent NGS study clearly showed that a mutant gland in the functional layer originated vertically from several stem cells located in the basal layer. [22][23][24] However, to the best of our knowledge, it is still unclear whether oncogenic mutations detected in normal endometrium contribute to carcinogenesis. Although NGS has enabled the detection of cancerrelated gene mutations with low MAF, their functional significance remains to be elucidated. One of the major barriers for analyzing the functional significance of these mutations is a lack of in vitro or in vivo systems available for long-term culture of mutated cells, for which our spheroid culture might be useful.
In contrast to the stem cell hit theory, another hypothesis is that genetic hits can occur in nonstem cells under the special circumstance of a lack of cyclic shedding. Assuming that some areas of the functional layer remain during menstruation that escape shedding, it is thought that pathogenic mutations can accumulate in nonstem daughter cells and could partly contribute to carcinogenesis.
Supporting this, a study using hysteroscopy and electron microscopy to capture morphological changes in the endometrium during menstruation showed that menstrual shedding occurs in a piecemeal manner. 34 They revealed that areas of unshed, shed, and healing endometrium coexist, and that a single gland remains like a clay tube after shedding. Notably, 1 day after the start of menstruation, stromal cells quickly cover the surface of the endometrium, and the uterine lumen becomes flat when observed with a hysteroscope. 34 Therefore, we may not have clinically captured the area of normal endometrium that does not shed. Recently, age-related accumulation of mutations in an endometrial cancer driver gene was shown in endometrial polyps that are not affected by regular shedding. 35 Abnormal endometrial repair mechanisms during the menstrual cycle may be involved in the accumulation of gene mutations in the endometrium. However, this needs further consideration.
Our research has the following limitations. First, mutations other than KRAS/PIK3CA hotspots were not evaluated. Second, it is possible that mutations with low mutation allele frequency were missed because NGS was not used. Third, this is a prospective experiment using clinical specimens with a limited number of patients in the proliferative phase.
However, even considering these limitations, this study clearly identified that major oncogenic mutations of KRAS and PIK3CA in normal endometrium can be detected by the simple and inexpensive Sanger method.
Currently, no early detection method for endometrial cancer in asymptomatic women has been established other than conventional cytological screening. Attempts have been made to detect genetic mutations with cytological samples, 36 but the results are inadequate. 37 Analysis using the Sanger method with a single gland is practical and clinically useful to identify oncogenic mutations in normal endometrium, and this could be developed as a novel tool to predict endometrial disorders or malignancies in perimenopausal or postmenopausal women with high risk factors.
In conclusion, Sanger sequencing detected KRAS or PIK3CA mutations in normal endometria with regional diversity. Spheroids grown from a single gland frequently had PIK3CA mutations expressing epithelial stem cell markers, supporting the stem cell hit theory. This information improves our understanding of endometrial physiology as well as stem cell-oriented endometrial regeneration and carcinogenesis.

ACK N OWLED G M ENTS
This work was supported by JSPS KAKENHI grant numbers JP22K09572 and JP21H03077.

CO N FLI C T O F I NTER E S T S TATEM ENT
The authors have no conflicts of interest to declare.

E TH I C A L A PPROVA L
Approval of the research protocol by an institutional review board: The protocol for acquiring and using tissue specimens was approved