Gynaecological cancers and their cell lines

Abstract Cell lines are widely used for various research purposes including cancer and drug research. Recently, there have been studies that pointed to discrepancies in the literature and usage of cell lines. That is why we have prepared a comprehensive overview of the most common gynaecological cancer cell lines, their literature, a list of currently available cell lines, and new findings compared with the original studies. A literature review was conducted via MEDLINE, PubMed and ScienceDirect for reviews in the last 5 years to identify research and other studies related to gynaecological cancer cell lines. We present an overview of the current literature with reference to the original studies and pointed to certain inconsistencies in the literature. The adherence to culturing rulesets and the international guidelines helps in minimizing replication failure between institutions. Evidence from the latest research suggests that despite certain drawbacks, variations of cancer cell lines can also be useful in regard to a more diverse genomic landscape.


| INTRODUC TI ON
There is a variety of methods to study the properties of malignancies. However, many of those methods depend on clinical subjects and/or patients and are in certain regards restricted/limited. A possible to avenue to form a better understanding of the carcinogenesis and behaviour of specific malignancies are cell cultures and cell lines (CLs). [1][2][3] Areas that commonly utilize CLs are pathology and oncology as well as pre-clinical areas such as pharmacology. Research on cancer cell lines (CCLs) in these areas holds the potential to lead to translational, clinically applicable results. Several papers have discussed and compared different cell models. This includes CLs, in vivo models, or cell models that are derived from individual patients.
It comes as no surprise that all have their drawbacks and advantages.
One of the oldest and still most recurring problems in CL culturing remains to be the potential of genetic or epigenetic changes can potentially arise during their growth and use. The latter limits the correlation potential of results based on these with the properties of the primary tissue. 2,4,5 By this, also their clinical relevance might be questioned. Nevertheless, there are several obvious benefits of CCLs. Figure 1 shows some of these potential applications of CCLs in medicine and translational research.
The limitations of CCLs, considerations about their relevance for clinical outcomes, and the discrepancies in their characteristics through the decades, were discussed before. For example, a recent paper by Ben-David et al presented evidence that points to a variety of mutational processes that affect the homogeneity of commonly used CCLs. On a practical level, this translates to the phenotypical characteristics of these CLs and their behaviour (therapy response). 6,7 Consequently, when considering the replicability of experiments, one understands that a number of factors can alter the final outcome. Some of these are as follows: (a) the number of population doublings that affect the cell geno-and phenotypes; (b) the origin of CCLs could be from the primary source tissue or a metastasis; (c) culture conditions could promote differentiation or even dedifferentiation. Despite this seemingly 'negative' prognosis to CL use, multiple mechanisms and rules have been proposed that help either prevent or at least help resolve these and other occurrences.
Some of these recommendations and guidelines will be discussed in the later sections of the review.
In comparison with complex culturing techniques (eg transgenic mice, transfection-based models and xenografts), CLs still provide a combination of a stable culturing setting, control over the experiment, relatively quick results and moderate expenses. 8,9 Furthermore, thanks to recent advances in genomic studies and the digitalization of data, specific CL properties can be simply checked via various databanks, including their possible uses and genetic heritage (eg Cellosaurus, ECLA). [10][11][12][13] We present an overview of three types of gynaecological CCLs.
Firstly, we present our methodology of search. The next chapter will discuss breast cancer (BC) and its CLs, the following endometrial cancer (EC) and its CLs, the subsequent cervical cancer (CC), and its CLs. Finally, the last chapter will be dedicated to new emerging and potential methods of use for CL research and our own related experiences.

| ME THODS
A literature review was conducted via the biggest medical literature databases (MEDLINE, PubMed, ScienceDirect) to obtain studies related to gynaecological CCLs. The employed search terms in the form of keywords were "breast cancer cell lines", "endometrial cell lines", "cervical cancer cell lines", "Gynaecological cancer cell lines".
Used MeSH identifiers were "Breast Neoplasms", "cell line", "cell lines, tumor", "cell line, transformed", "uterine cervical neoplasms" and "endometrial neoplasms". With the help of this search algorithm and specific filters (5 years, human, review), we were able to find relevant new impactful studies on gynaecological CLs (Table 1). We specifically searched for the corresponding originators' study which was crosschecked via the Cellosaurus database. The search inquiries and presentation of the final number of included studies after exclusion and filtering is presented in Figure 2 and has been prepared in accordance with the PRISMA guidelines for review articles.

| Introduction
The second most common cancer worldwide is breast cancer. This type of cancer is ranked first in incidence in women living either in developed or developing countries (Table 2). It is ranked as the fifth common cause of cancer death (627 000 deaths/per year). 14 A short overview of the most important histological subtypes is shown in Table 3.

| Invasive breast cancer
We distinguish multiple different types of invasive breast cancer (IBC). The distinctions are based on the molecular as well as histopathological properties. 15 In a routine diagnostic procedure, multiple different criteria will be gathered and analysed; tumour size and grade (Elston and Ellis); presence of lymphovascular invasion; DCIS and LCIS; assessment of surgical margins; the immunohistochemical profile of hormonal receptors (eg oestrogen -ER (most commonly the alpha receptor), progesterone -PR), the HER2 receptor status and analysis of proliferation index, determined with the Ki-67 (Mib 1) percentage of nuclear expression, is an additional independent prognostic parameter for DFS and OS in BC patients. 16 Furthermore, proteases and genetic profiling can be of great help in determining the need for additional treatment. These features are not only of great importance for the clinician but also of great importance for the researcher during CL culturing. Namely, for successful culturing, the scientist must know the specific properties to provide and use an adequate environment, culturing methods, characterization methods, etc, to preserve the characteristics of the source tissue.
Primary BC can be classified into two categories. The first are the non-invasive BCs. These are ductal carcinoma in situ (DCIS) and lobular carcinoma in situ (LCIS). These types of cancer are confined within the milk ducts and lobules of the breast and surrounded by an intact basement membrane and myoepithelial cells (Table 3). 16 More common are invasive BC types. The term invasive breast carcinoma (IBC) refers to a large heterogeneous group of malignant epithelial neoplasms of the breast.
Due to treatment purposes, distinct outcomes and responses to therapy, all IBCs are grouped into the following subtypes: (a) ER-positive, HER2-negative; (b) ER-positive, HER2-positive; (c) ERnegative, HER2-positive; (d) ER-negative, HER2-negative. 16 The majority of BCs are unifocal and can occur in any quadrant of the breast, with a higher frequency in the upper outer quadrant. A synchronous contralateral tumour is found in approximately 2% of patients. The macroscopically most common features in advanced stages of IBC are skin retraction, nipple inversion, nipple discharge, change of texture or colour of the skin. Otherwise, lesions that are discovered early are mostly asymptomatic and may clinically show themselves on palpation as a lump. About 5.15% of all palpable BCs are not seen on a mammogram but can be identified with ultrasound.
MRI is the most sensitive method. 16 The most common is the invasive carcinoma of no special type (NST), which accounts for 40%-75% of all invasive BC types and is commonly present alongside DCIS. The special morphological patterns include a variety of different patterns. These BCs can be mixed IBC-NST and special subtypes such as the pleomorphic carcinoma, BC with osteoclast-like giant cells, carcinoma with choriocarcinomatous features and IBC with melanocytic differentiation, oncocytic pattern, lipid-rich pattern, glycogen-rich clear cell pattern and sebaceous pattern. 16 To continue, based on epidemiological data, invasive lobular carcinoma (ILC) comes in second place. 16

| Breast cancer cell lines
The field of BC CLs has been established with the first BC CL from Lasfargues and Ozzello in 1958 called BT-20. 18 Following an increased interest in this area, researchers have, in the following years, isolated and presented an increasing number of CCLs. We face a growing assortment of different CCLs, whereas the integrity and uniqueness of certain CCLs are doubtful. The most commonly used BC CL is still MCF7, T47D and MDAMB231. 4,19 Their properties can be seen in Table 4. We also published some more information on the less known BC CLs in one of our previous studies. 4 The first to successfully culture a BC CL was Lasfargues and Ozzello in 1958. The CL was named BT-20. 18 The tissue was obtained from a 74-year-old patient, who had a mammary duct-cell carcinoma of no special type. The cell collection was performed via sampling of the spillage that occurred during tumour preparation (slicing). 18  The study from Holiday et al 2011 was, at that time, one of the rare and most comprehensive studies on this specific subject at that time.
Nine years after their study, we can see that although there is now a reasonable number of BC CLs, these still lack some of the rarer histo- What is surprising is that their results differ from previously reported CL expression profiles for BT20 CL. This CL has been in a wide variety of studies described as a TNBC BC CL. 12,30,31,33 However, the authors report results of it being HER2 overexpressed. 32 Other BC CLs that have inconsistent expression profiles in various studies are HCC1007, HCC1419, HCC1500, HCC2185, SUM52PE, SUM44PE, EVSA-T and EVSA-T. 20,31,[34][35][36][37] The reasons for these inconsistencies can be several.
We will give an overview of this topic in the last segment of the paper.

| Endometrial cancer
EC presents a very common gynaecological malignancy. It has a yearly incidence rate of 60 000 (United States) new cases and more than 10 000 deaths. [38][39][40][41] On a global scale, there are 382 069 new cases yearly and 89 929 deaths. [38][39][40][41] The incidence of this cancer is very tightly linked to certain epidemiologic factors. According to literature, obesity being one of the most important. [38][39][40]42 Others include the exposure to unopposed oestrogens or tamoxifen, diabetes, nulliparity, early-onset menarche and late-onset menopause. In one of our previous publications, we provided an in-depth overview of EC CLs. 2 We can, based on the latest WHO classification (2020, 5th), divide EC into multiple histological subtypes. The most common is the endometrioid type (up to 80%) followed by mixed cell type (up to 10%), serous (up to 10%); carcinosarcoma (<10%), clear cell (<10%), undif- The standard treatment for endometrial cancer consists of primary hysterectomy and bilateral salpingo-oophorectomy (eg laparoscopic or robotic). The 5-year overall survival ranges from 74% to 91% in patients without metastatic disease. 49

| Endometrial cancer cell lines
Endometrial CCL that is widely used is Ishikawa, HEC-1-  Table 5.

| Cervical cancer
The last cancer type to be discussed in this review is uterine cervix cancer (CC). CC is a common type of cancer and has a relatively high mortality among gynaecological cancers with stark regional differences. The incidence and death toll are on a global scale estimated at approximately 570 000 and 311 000 per year. 54 Interregional differences in mortality can be attributed to countries lacking cervical cancer screening and prevention programmes, since this type of cancer remains the second most common type (17.8 per 100 000 women) as well as the cause of cancer deaths (9.8 per 100 000) among all types of cancer in women that live in lower income countries (Table 2). 14 Furthermore, collectively, most of CC cases (80%-90%) occur in these countries (eg parts of Africa and Asia). 55 The most important directly linked factor for CC is the human papillomavirus (HPV) infection, 56 which is also the most common sexually transmitted disease. Through research, more than 200 different types have been found, identified and systematically classified into 5 genera (α, β, γ, μ and ν). 57 Based on their oncogenic potential, these are labelled as high risk and low risk. The most studied types are of the α genera, since these have been directly linked to almost all squamous intraepithelial lesions and cancers of the cervix and anus as well as to a subset of penile, vulvar and vaginal cancers. 57 HPV types and their biologic potential can be seen in Table 6.
It has been estimated that HPV is responsible for almost a tenth of human malignancies (7%-8%). It is associated with almost all cases of CCs (96%) and anal cancers (93%). Furthermore, almost two thirds of all vaginal cancers (64%) and oropharyngeal carcinomas (63%) arise due to its oncogenic potential. And lastly, HPV also presents an important factor in the development vulvar cancer (51%) and penile cancers (36%). 57 The most common histologic types of CC are squamous cell carcinoma (64.5%), adenocarcinoma (28.9%) and 6.6% other histology's. 58,59 Based on a recent study from 2019, HPV infection may also be implicated in developing some types of breast cancer. 60 The presence of screening programmes for cervical cancer and public health programmes for HPV vaccination and education of the general public led, according to sources, to a significant decrease in the incidence and mortality of cervical cancer over the past 50 years in developed countries (75%). [61][62][63] It is worth noting that viral load seems to be in an inverse correla-

| Cervical cancer cell lines
The most famous and first CCL in the world is HeLa (Figure 3). The HeLa cells are named after the 30-year-old patient Henrietta Lacks who died in 1951 due to an aggressive adenocarcinoma of the cervix. 66 Some of the tissue obtained from a cervical biopsy was supplied to the Tissue Culture Laboratory in the Department of Surgery at The Johns Hopkins Hospital for research purposes. Contrary to their previous results, the cells grew robustly and became the first human CCL immortalized in tissue culture. As is widely known, the naming process consisted of utilizing the initial 2 letters of Henrietta Lacks' first and last names. 66 The cells have been shown to contain human papillomavirus (HPV) 18 DNA, and HPV18-positive HeLa cells have been linked to changes in microRNA expression. These results were obtained by Gey and colleagues, which published their study in 1952. 67 The most common cervical CCLs are shown in Table 7. What is important to note is the fact that these cells have been in circulation for more than 60 years. Subsequently, they are almost ubiquitous. 68 This has, through the years, also led to many contaminations and has become a crucial problem in CL culturing.  Note: Summarized after. [61][62][63] database for all articles stating the names of one of the 451 listed CLs. 68 Figure 4 shows  protein ig-h3 and NRF2. 70 Their analysis pointed out that the expression of NRF2 indicates that aberrant NRF2-mediated oxidative stress response (OSR) is a prominent feature of cervical carcinogenesis. 70 Another fascinating study showed that cervical CCLs express markers associated with immunosurveillance. These were as follows: Some measures for better consistency and reproducibility in cell culturing can be seen in Table 8.

| Author experience and recommendations
Culturing procedures have drastically changed over the years, especially with the development of new reagents and new materials.
If comparing our current procedures with the past, we can now appreciate a much more streamlined process. 2 The core principle, which is the usage of a culturing medium that supports the growth of cells as well as additions in the form of antibiotics that stop the growth of bacteria, is of course universally the same. However, the current procedures are much more uniform and perhaps easier to use as those described by the CL 'pioneers' (eg Dawe et al, AN3-CA).
Nevertheless, the protocols can still differ between laboratories.
To illustrate, we present a simple workflow protocol ( Figure 6) that we used for the isolation of our cell cultures and our characterization. 3

CL identification
To avoid misidentification, acquired CLs should come from a reliable source and must be authenticated, bought from a reliable source and banked for future use. Additional STR profiling is also important. 94 Mycoplasma testing To avoid contamination good tissue culture practice and frequent testing should be performed to ensure that CLs are clear of contamination. 95 Use of validated reagents To avoid a variety of errors only reagents of certified laboratory purity should be used. Decontamination should not be avoided but rather the experiment repeated. 96 Statistical standards To avoid misinterpretations and promote transparency, mandatory reporting checklist that catalogued details of statistical information, experimental design and reagents, should be included.
. 7, [97][98][99] Profiling To avoid contaminations with other CLs and possible erroneous results, laboratory's own CLs should be compared to reference CL genomes. 74 Cryopreservation To avoid loss of data and ensure replicability, preservation of the primary cultures and early passages with subsequent final comparison and validation of key findings before publication is of grave importance. 74 Reporting DTs To promote transparency and replicability, accurate and diligent monitoring as well as reporting of DTs as well as a finite usage number of passages should be standard practice. 74 Standardized conditions To ensure interlaboratory replicability, international standardized culture conditions for individual CLs and documentation of the heterogeneity metrics in datasheets should be standard practice. 74 Naming To ensure coherent scientific reporting, the international naming guidelines should be used. 100 Abbreviations: CLs, cell lines; DT, doubling time; STR, short tandem repeat.