Genetic Characterization of Rat Hepatic Stellate Cell Line PAV-1

The rat hepatic stellate cell line PAV-1 was established two decades ago and proposed as a cellular model to study aspects of hepatic retinoic acid metabolism. This cell line exhibits a myofibroblast-like phenotype but also has the ability to store retinyl esters and synthesize retinoic acid from its precursor retinol. Importantly, when cultured with palmitic acid alone or in combination with retinol, the cells switch to a deactivated phenotype in which the proliferation and expression of profibrogenic marker genes are suppressed. Despite these interesting characteristics, the cell line has somehow fallen into oblivion. However, based on the fact that working with in vivo models is becoming increasingly complicated, genetically characterized established cell lines that mimic aspects of hepatic stellate cell biology are of fundamental value for biomedical research. To genetically characterize PAV-1 cells, we performed karyotype analysis using conventional chromosome analysis and multicolor spectral karyotyping (SKY), which allowed us to identify numerical and specific chromosomal alteration in PAV-1 cells. In addition, we used a panel of 31 species-specific allelic variant sites to define a unique short tandem repeat (STR) profile for this cell line and performed bulk mRNA-sequencing, showing that PAV-1 cells express an abundance of genes specific for the proposed myofibroblastic phenotype. Finally, we used Rhodamine-Phalloidin staining and electron microscopy analysis, which showed that PAV-1 cells contain a robust intracellular network of filamentous actin and process typical ultrastructural features of hepatic stellate cells.


Introduction
Immortalized hepatic stellate cell (HSC) lines are useful experimental tools for the study of cellular aspects of hepatic fibrogenesis [1,2]. Like primary HSC, most of these continuously growing cell lines produce a well-developed α-smooth muscle actin (α-SMA) network, express a multitude of characteristic connective tissue markers, and have the capacity to take up and esterify retinol [1,2]. Compared to their primary counterparts that have a limited lifespan, established HSC lines grow faster and can be continuously passaged, indicating that these cells have developed features that allow them to escape from 1 cells by using conventional chromosome analysis of metaphase spreads. Furthermore, several cryptic chromosome rearrangements were detected with spectral karyotyping.
Moreover, we used a panel of 31 species-specific allelic variant sites to define a unique STR profile for PAV-1 cells. PAV-1 cells cultured under basal conditions were further subjected to bulk mRNA-sequencing, showing that PAV-1 cells express an abundance of genes specific for HSC. By employing Rhodamine-Phalloidin staining, we further demonstrate that PAV-1 cells form a robust intracellular network of filamentous actin. PAV-1 cells further possess ultrastructural features of HSC, including large, electron-dense nuclei with excessive bulges of the nuclear envelope, elongated mitochondria, distinct rough endoplasmic reticulum, pronounced lysosomes, and an abundance of intracellular lipid droplets, as assessed by electron microscopy analysis.
In summary, this study defines unique authentication standards for PAV-1 cells that will form the basis for avoiding misidentification when using this cell line in biomedical research.

Culturing of PAV-1 Cells
PAV-1 cells were initially established and characterized two decades ago by Patrick Sauvant and colleagues [23,24]. The cells were spontaneously immortalized from cultured primary rat HSCs that were purified from male Wistar rats by a pronase-collagenase perfusion protocol [23]. In our laboratory, PAV-1 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) (#D6171), supplemented with 10% fetal bovine serum (FBS, #F7524), 2 mM L-Glutamine (#G7513), 1 mM sodium pyruvate (#S8636), and 1× Penicillin/Streptomycin (#P0781). All reagents were obtained from Sigma-Aldrich (Taufkirchen, Germany). For routine culturing, the medium was replaced every third day and cells were sub-cultured using Accutase ® solution (#A6964, Sigma-Aldrich). Detection of potential Mycoplasma spp. contaminations in cell culture supernatants was completed using the Venor ® GeM OneStep kit (#11-8050, Minerva biolabs GmbH, Berlin, Germany) according to the manufacturer's instructions. The PCR products were separated on a 2% standard agarose gel including ethidium bromide and amplicons visualized using a standard gel imager. The clearance of PAV-1 cells from Mycoplasma contamination was completed using the Plasmocin ® -Mycoplasma elimination reagent and by essentially following the instructions provided by the manufacturer (abt-mpt-1, InvivoGen, Toulouse, France). The rat HSC line HSC-T6 and murine hepatocytic AML-12 cells were cultured as previously described [19,21].

Electron Microscopy Analysis
Electron microscopy analysis of PAV-1 cells was completed as described previously for the rat HSC line HSC-T6 [19]. In brief, cells were fixed, dehydrated, and subsequently embedded in Epon resin. Ultrathin sections (70-100 nm) were prepared and packed upon Cu/Rh grids and stained with 0.5% uranyl acetate and 1% lead citrate for better contrast. The samples were analyzed at an acceleration voltage of 60 kV using a Zeiss Leo 906 (Carl Zeiss AG, Oberkochen, Germany) transmission electron microscope. The magnification range varied from 4646× to 27,800×.

Preparation of PAV-1 Metaphase Chromosomes and Karyotyping
Chromosomes of PAV-1 cells were prepared according to standard protocols [19]. Briefly, cultures of semi-confluent PAV-1 cells were exposed to colcemid solution (Gibco, ThermoFisher Scientific, Dreieich, Germany), detached through brief trypsinization, harvested by centrifugation, treated with 0.56% (w/v) hypotonic potassium chloride solution, and fixed with a mixture of cold methanol and acetic acid (3:1) at 37 • C. For G-banding, air-dried chromosome spreads were prepared, and slides were treated with 0.025% (w/v) trypsin solution followed by staining with Giemsa solution. The heterochromatin in metaphase chromosomes was stained using a standardized protocol [25]. At least 10 GTG- banded metaphases from different culture passages were photographed and karyotyped according the method outlined in [26].

In Situ Hybrdization, Spectral Imaging, and Nucleolar Organizer Region Staining
Detection of chromosome rearrangements in PAV-1 cells was accomplished using a commercially available rat SKY probe (Applied Spectral Imaging Inc., Carlsbad, CA, USA) and published protocols [19,20,27]. For silver (Ag)-staining of the nucleolus organizer regions (NORs), slides with metaphase chromosomes were stained with AgNO 3 according to the method of Goodpasture and Bloom [28].

Short Tandem Repeat (STR) Profiling
STR profiling and interspecies contamination testing for PAV-1 cells was performed using the commercially offered cell line authentication service from IDEXX (Kornwestheim, Germany). The profiling was completed using the CellCheck TM Rat system that includes 31 species-specific STR markers that are distributed on the 20 different rat autosomes. The mouse cell line AML12 was subjected for STR profiling using the CellCheck TM Mouse system that contains the 19 species-specific consensus STR markers proposed by the Consortium for Mouse Cell line Authentication for STR profiling in mouse [29,30]. An STR similarity search with the obtained STR profile of AML12 was performed using the Cellosaurus STR Similarity Search Tool CLASTR 1.4.4 (release 41.0) and the Cellosaurus mouse STR database [31,32]. In the search, the settings were set to the following: Scoring algorithm: Tanabe, Mode: Non-empty markers, Score filter: 40%, and Min. Markers: 8.

Next Generation Sequencing and Data Analysis
RNA from cultured PAV-1 grown to a density of 80% was isolated by a standard CsCl density gradient centrifugation as described before [19]. The resulting RNA pellet was resuspended in sterile water, purified by ethanol precipitation, re-suspended in sterile water, and quantified by using UV spectroscopy. For RNA bulk sequencing, the quality of the RNA was determined on the Agilent 4200 TapeStation platform (Agilent Technologies Inc., Waldbronn, Germany). Depletion of ribosomal RNAs, library preparation, sequencing, and bioinformatics analysis were essentially performed as described before [19]. The abundance of found individual gene transcripts is given in 'Transcripts Per Million (TPM)'.

Western Blot Analysis
Preparation of protein extracts, protein quantification, and Western blot analysis were performed using established protocols [34]. In brief, equal protein amounts (40 µg/lane) were heated at 80 • C for 10 min and separated in 4-12% Bis-Tris gels (Invitrogen, Darmstadt, Germany) under decreasing conditions using MES running buffer. The separated proteins were then electro-blotted on nitrocellulose membranes (Schleicher & Schuell, Dassel, Germany), and successful protein transfer and equal protein loading controlled by Ponceau S stain. After blocking unspecific binding sites in 5% milk powder in Tris-buffered saline with Tween 20 (20 mM Tris, 150 mM NaCl, 0.1% (w/v) Tween 20 detergent), the membranes were successively probed with the primary antibodies depicted in Table 1. Primary antibodies were detected with horseradish peroxidase (HRP)-conjugated secondary antibodies and the Supersignal™ chemiluminescent substrate (Perbio Science, Bonn, Germany).

Phenotypic Characterization of PAV-1 Cells
PAV-1 cells were established over two decades ago [23]. After receiving the cells from the original laboratory that generated this cell line, the cells were first stored in our laboratory for about 10 years in a liquid nitrogen tank and later transferred into a −150 • C freezer.
Because we have not worked with this cell line before, we first tested for mycoplasma infection that might be introduced during establishment, handling, parceling, storing, or by contaminated cell culture reagents such as newborn bovine serum frequently contaminated with mycoplasma [36,37]. According to the information provided by the manufacturer, the conventional PCR-based assay used for this analysis is designed to specifically target and amplify the highly conserved 16S rRNA coding region of the mycoplasma genome, presumably allowing the detection of 85 different Mycoplasma species, 7 Acholeplasma, and 1 Ureaplasma including M. orale, M. hyorhinis, M. arginini, M. fermentans, M. salivarium, M. hominis, usually encountered as contaminants in cell cultures. Unfortunately, the original stock tested positive for mycoplasma infection; however, this could be alleviated via the usage of a Mycoplasma elimination reagent before starting with the phenotypic and genotypic characterization described in the following. Because we have not worked with this cell line before, we first tested for mycoplasma infection that might be introduced during establishment, handling, parceling, storing, or by contaminated cell culture reagents such as newborn bovine serum (frequently contaminated with mycoplasma; see Section 2.1) ( Figure S1).
As already mentioned above, the PAV-1 cell line is a spontaneously immortalized HSC line derived from cultured primary rat HSCs originating from male Wistar rats [23]. These cells resemble activated HSC and share many characteristics with primary HSC/myofibroblasts. In particular, they acquire a distinct fibroblast-like morphology in culture, supporting the notion that they originate from connective tissue (Figure 1). At low density, PAV-1 cells appear in fusiform, elongated spindle-shaped structures, whereas at higher density, the cells form a uniform, dense cell layer.

Phenotypic Characterization of PAV-1 Cells
PAV-1 cells were established over two decades ago [23]. After receiving the from the original laboratory that generated this cell line, the cells were first stored i laboratory for about 10 years in a liquid nitrogen tank and later transferred into a −1 freezer. Because we have not worked with this cell line before, we first tested for m plasma infection that might be introduced during establishment, handling, parc storing, or by contaminated cell culture reagents such as newborn bovine serum quently contaminated with mycoplasma [36,37]. According to the information prov by the manufacturer, the conventional PCR-based assay used for this analysis is desi to specifically target and amplify the highly conserved 16S rRNA coding region o mycoplasma genome, presumably allowing the detection of 85 different Mycop species, 7 Acholeplasma, and 1 Ureaplasma including M. orale, M. hyorhinis, M. arginin fermentans, M. salivarium, M. hominis, usually encountered as contaminants in cel tures. Unfortunately, the original stock tested positive for mycoplasma infection; ever, this could be alleviated via the usage of a Mycoplasma elimination reagent b starting with the phenotypic and genotypic characterization described in the follow Because we have not worked with this cell line before, we first tested for mycopl infection that might be introduced during establishment, handling, parceling, storin by contaminated cell culture reagents such as newborn bovine serum (frequently taminated with mycoplasma; see Section 2.1) ( Figure S1).
As already mentioned above, the PAV-1 cell line is a spontaneously immorta HSC line derived from cultured primary rat HSCs originating from male Wistar rats These cells resemble activated HSC and share many characteristics with pri HSC/myofibroblasts. In particular, they acquire a distinct fibroblast-like morpholo culture, supporting the notion that they originate from connective tissue (Figure low density, PAV-1 cells appear in fusiform, elongated spindle-shaped struc whereas at higher density, the cells form a uniform, dense cell layer. Electron microscopy revealed that PAV-1 cells possess ultrastructural features that are hallmarks of cells originating from HSC, including large, electron-dense nuclei, elongated mitochondria, distinct rough endoplasmic reticulum, pronounced lysosomes, and, most importantly, a large abundance of intracellular lipid droplets ( Figure 2).
Furthermore, in agreement with the proposed myofibroblastic phenotype, we could show that these cells form a robust network of cytoplasmic microfilaments, as assessed by Rhodamine-Phalloidin staining ( Figure 3). Electron microscopy revealed that PAV-1 cells possess ultrastructural features are hallmarks of cells originating from HSC, including large, electron-dense nu elongated mitochondria, distinct rough endoplasmic reticulum, pronounced lysoso and, most importantly, a large abundance of intracellular lipid droplets ( Figure 2). Furthermore, in agreement with the proposed myofibroblastic phenotype, we co show that these cells form a robust network of cytoplasmic microfilaments, as asse by Rhodamine-Phalloidin staining ( Figure 3).   . F-actin cytoskeleton staining in PAV-1. Cultured PAV-1 cells were stained with a Rhodamine-Phalloidin conjugate (red) and nuclei counterstained with DAPI (blue). Images were taken with a Nikon Eclipse E80i fluorescence microscope at 400× or 600× magnification. As previously shown, PAV-1 cells express large quantities of α-smooth muscle actin (Acta2) [23].
HSC typically contain large quantities of triglycerides, cholesterol ester, cholesterol, phospholipids, and free fatty acids that are stored in lipid droplets [38]. Importantly, the size and number of these lipid-storing vesicles is strongly dependent on dietary fatty acid intake, suggesting that these fat-storing cells have the capacity to uptake respective compounds [39]. The PAV-1 line was introduced as an activated rat HSC line with the ability to uptake, store, and metabolize fatty acids and vitamin A, thereby reversing the activated phenotype of PAV-1 into a quiescent phenotype [40]. In agreement with these findings, we demonstrated that the incubation of cultured PAV-1 with oleic acid (250 µM) resulted in the enlargement of lipid droplets, as demonstrated by BODIPY staining (Figure 4). intake, suggesting that these fat-storing cells have the capacity to uptake respective compounds [39]. The PAV-1 line was introduced as an activated rat HSC line with the ability to uptake, store, and metabolize fatty acids and vitamin A, thereby reversing the activated phenotype of PAV-1 into a quiescent phenotype [40]. In agreement with these findings, we demonstrated that the incubation of cultured PAV-1 with oleic acid (250 µM) resulted in the enlargement of lipid droplets, as demonstrated by BODIPY staining (Figure 4).

Western Blot Analysis and Reverse Transcriptase PCR
We subsequently performed Western blot analysis for typical HSC markers using the rat HSC line HSC-T6 as a control. This analysis showed that PAV-1 cells express collagen type I, desmin, caveolin-1, vimentin, fibronectin, collagen type IV, and GFAP (Figure 5).

Expression of Hepatic Stellate Cell Markers in PAV-1 3.2.1. Western Blot Analysis and Reverse Transcriptase PCR
We subsequently performed Western blot analysis for typical HSC markers using the rat HSC line HSC-T6 as a control. This analysis showed that PAV-1 cells express collagen type I, desmin, caveolin-1, vimentin, fibronectin, collagen type IV, and GFAP ( Figure 5). When compared to HSC-T6, which represents another immortalized HSC line, the expression of all these markers at protein level was significantly higher in PAV-1 cells. Only the expression of collagen type IV was slightly reduced in PAV-1 cells. As expected, the expression of the Simian virus large T-antigen (SV40T) was absent in PAV-1 cells, while the expression of SV40T was high in HSC-T6 that were immortalized with this Cell protein extracts were prepared from AML12, PAV-1, and HSC-T6 cells and analyzed by Western blot analysis (40 µg protein/lane) for expression of collagen type I, hepatocyte nuclear factor 1α (HNF-4α), desmin, caveolin-1, α-smooth muscle actin (α-SMA), vimentin, fibronectin, collagen type IV, glial fibrillary acidic protein (GFAP), and Simian virus (SV40) large T antigen (SV40T). Ponceau S staining and probing with antibodies specific for β-actin and glyceroaldehyde 3-phosphate dehydrogenase (GAPDH) served as controls to document equal protein loading. In this analysis, the cell line AML12 was used as a positive control for HNF-4α expression. * As a further control for GFAP expression, extracts prepared from rat brain (10 µg protein/lane) served as a positive control.
When compared to HSC-T6, which represents another immortalized HSC line, the expression of all these markers at protein level was significantly higher in PAV-1 cells. Only the expression of collagen type IV was slightly reduced in PAV-1 cells. As expected, the expression of the Simian virus large T-antigen (SV40T) was absent in PAV-1 cells, while the expression of SV40T was high in HSC-T6 that were immortalized with this dominantacting oncogene. In addition, PAV-1 cells were negative for the nuclear receptor hepatocyte nuclear factor 1α (HNF-4α), representing a transcription factor that regulates the expression of several hepatocyte-specific genes during hepatic differentiation and proliferation [41]. Interestingly, AML12 cells that were used as a control for HNF-4α expression showed a faint expression of GFAP-an astrocyte marker that, within the liver, is typically only expressed in HSC and generally suggested as an early marker of stellate cell activation [42,43]. The respective band was found with two different antibodies. To document the identity and rule out cross-contamination with other cells in our AML12 culture that was used as a control, we performed STR profiling in AML12 using the CellCheck TM Mouse system, including 19 species-specific STR markers, from which 18 are also implemented in the Cellosaurus cell line knowledge resource [31,32]. This analysis revealed that the AML12 cells used in our study had a unique STR profile and were not contaminated by other cells ( Figure S2, Table S1). The highest similarities regarding the 18 Cellosaurus markers were assigned to CVCL_0120 (3T-Swiss albino, 46. The expression of collagen type I (Col1α) and fibronectin (Fn1) was also confirmed by conventional reverse transcription polymerase chain reaction (RT-PCR) ( Figure 6A). However, in both PAV-1 and HSC-T6, we found only low mRNA quantities of Gfap that were about 64-fold lower in PAV-1 than in HSC-T6, as indicated by the different C t values in RT-qPCR that differed in their cycle threshold values by a factor of 2 6 ( Figure 6B). This finding contrasted the GFAP protein expression levels that were much higher in PAV-1 cells. were about 64-fold lower in PAV-1 than in HSC-T6, as indicated by the different Ct values in RT-qPCR that differed in their cycle threshold values by a factor of 2 6 ( Figure 6B). This finding contrasted the GFAP protein expression levels that were much higher in PAV-1 cells.

Transcriptomic Analysis of PAV-1 Cells
To measure the general expression of individual genes in PAV-1 cells, we performed bulk RNA-sequencing (mRNA-seq). In the respective analyses, we showed that PAV-1 cells express a total of 23765 different mRNA species (Table S2). The five highest expressions were found for ribonuclease pancreatic beta-type (LOC103690354, 17612.5 TPM), ferritin heavy chain 1 (Fth1, 17487.4 TPM), eukaryotic translation elongation factor 1-α1 (Eef1a1, 8875.58 TPM), ferritin light chain 1-like (LOC100360087, 7138.67 TPM), and vimentin (Vim, 7038.71), respectively. Importantly, PAV-1 cells expressed a large number of HSC markers that were also found in the rat HSC lines HSC-T6 and CFSC-2G (Table 2). Interestingly, the bulk mRNA sequencing confirmed our RT-PCR results, showing that Gfap mRNA is expressed in HSC-T6 at a rate that is about 60-fold higher than in PAV-1 (0.60103 TPM vs. 0.0115619).

Transcriptomic Analysis of PAV-1 Cells
To measure the general expression of individual genes in PAV-1 cells, we performed bulk RNA-sequencing (mRNA-seq). In the respective analyses, we showed that PAV-1 cells express a total of 23765 different mRNA species (Table S2). The five highest expressions were found for ribonuclease pancreatic beta-type (LOC103690354, 17612.5 TPM), ferritin heavy chain 1 (Fth1, 17487.4 TPM), eukaryotic translation elongation factor 1-α1 (Eef1a1, 8875.58 TPM), ferritin light chain 1-like (LOC100360087, 7138.67 TPM), and vimentin (Vim, 7038.71), respectively. Importantly, PAV-1 cells expressed a large number of HSC markers that were also found in the rat HSC lines HSC-T6 and CFSC-2G (Table 2). Interestingly, the bulk mRNA sequencing confirmed our RT-PCR results, showing that Gfap mRNA is expressed in HSC-T6 at a rate that is about 60-fold higher than in PAV-1 (0.60103 TPM vs. 0.0115619). Moreover, in line with previous reports and other rat HSC lines, we found that PAV-1 cell express many metabolic and nuclear retinoic acid receptors, including retinoic acid receptor (RAR)α, RARβ, RARγ, retinoid X receptor (RXR)α, RXRβ, retinol binding proteins, retinal dehydrogenases, and aldehyde dehydrogenases ( Table 3, Table S2).

Karyotype Analysis and Spectral Karyotyping of PAV-1 Cells
The counting of numerous Giemsa-stained metaphase spreads from different passages showed that the overall diploid chromosome numbers vary between 38 and 45, with most metaphases (>70%) displaying 43 chromosomes. However, in each culture passage, some cells were endowed with >80 chromosomes (nearly tetraploid, Figure S3). Interestingly, the number of tetraploid cells tends to outnumber the diploid cells in higher culture passages (Table 4). This phenomenon might be analogous to the increasing rate of genomic rearrangements in liver with age [44].
In each cell, a large metacentric and two small marker chromosomes were noted, which appear to be a typical feature of PAV-1 cells. The GTG banding pattern allowed us to classify the individual chromosomes and revealed several unique features of the PAV-1 karyotype. In GTG-banded karyotypes, the large metacentric chromosome was identified as an iso-chromosome of RNO5 ( Figure 7A). The other notable feature of the karyotype is the presence of three copies of RNO7 (trisomy) and one copy of RNO12 (monosomy). Additionally, in one of the homologs of RNO4, an extra chromosomal segment at the terminal region of the long arm can be noted. In each cell, a large metacentric and two small marker chromosomes were noted, which appear to be a typical feature of PAV-1 cells. The GTG banding pattern allowed us to classify the individual chromosomes and revealed several unique features of the PAV-1 karyotype. In GTG-banded karyotypes, the large metacentric chromosome was identified as an iso-chromosome of RNO5 ( Figure 7A). The other notable feature of the karyotype is the presence of three copies of RNO7 (trisomy) and one copy of RNO12 (monosomy). Additionally, in one of the homologs of RNO4, an extra chromosomal segment at the terminal region of the long arm can be noted.
The C-banded metaphases unequivocally identified the heterochromatic Y-chromosome ( Figure 7B), confirming that the investigated PAV-1 cell line was originally derived from a male rat. The remaining constitutively heterochromatic regions (C-bands) were confined to the centromeric regions of most chromosomes.  Figure 8 shows the silver staining of metaphase chromosomes, demonstrating six active nucleolar organizer sites (NORs). As expected, these NORs were located near the centromeres of RNO3, RNO11, and RNO12. Interestingly, the larger marker chromosomes had active NORs in each metaphase analyzed, suggesting that either an NOR was transferred to the marker chromosome, or alternatively, that the marker chromosome was derived from one of the three NOR-carrying rat chromosomes, RNO3, RNO11, or RNO12. The C-banded metaphases unequivocally identified the heterochromatic Y-chromosome ( Figure 7B), confirming that the investigated PAV-1 cell line was originally derived from a male rat. The remaining constitutively heterochromatic regions (C-bands) were confined to the centromeric regions of most chromosomes. Figure 8 shows the silver staining of metaphase chromosomes, demonstrating six active nucleolar organizer sites (NORs). As expected, these NORs were located near the centromeres of RNO3, RNO11, and RNO12. Interestingly, the larger marker chromosomes had active NORs in each metaphase analyzed, suggesting that either an NOR was transferred to the marker chromosome, or alternatively, that the marker chromosome was derived from one of the three NOR-carrying rat chromosomes, RNO3, RNO11, or RNO12.

Spectral Karyotype Analysis
SKY analysis is based on specific color codes assigned to each chromosome and allows for the accurate identification of both numerical and structural rearrangements. Figure 9 displays a PAV-1 metaphase spread and representative SKY karyotype with chromosome-specific paintings.

Spectral Karyotype Analysis
SKY analysis is based on specific color codes assigned to each chromosome and allows for the accurate identification of both numerical and structural rearrangements. Figure 9 displays a PAV-1 metaphase spread and representative SKY karyotype with chromosomespecific paintings.

Spectral Karyotype Analysis
SKY analysis is based on specific color codes assigned to each chromosome and allows for the accurate identification of both numerical and structural rearrangements. Figure 9 displays a PAV-1 metaphase spread and representative SKY karyotype with chromosome-specific paintings. with the SKY probe cocktail that contains specific colored probes for each chromosome (B) Classified pseudo-colored image of the metaphase spread after hybridization with SKY paints. (C) Inverted DAPI-stained image. (D) Karyotype of the metaphase showing spectrally classified, pseudo-colored chromosomes (right) compared with its inverted DAPI-stained chromosomes (left) and corresponding RGB image (middle). Specific chromosomes involved in the rearranged chromosomes are indicated. Based on the SKY analysis, the PAV-1 karyotype can be designated as: 43, XY, der(4)t(4;11), i(5)(q10), +7, +der(11)t(11;20)+11(p), t (12;19), +mar.
In addition to the iso-chromosome of RNO5 and trisomy 7, there were three different chromosomes (RNO4, RNO11 and RNO12) containing segments of other chromosomes. For example, the spectral patterns of RNO19 and RNO20 were present on one homolog of RNO11 and RNO12-consistent with translocations between RNO11 and RNO12 and between RNO12 and RNO19, respectively. The second homologs of RNO19 and RNO20 maintained their morphology as individual chromosome units in the karyotype. Additionally, one of the homologs of RNO4 has a tiny segment from RNO11 at the terminal end of its long arm, implying that RNO11 material was translocated to RNO4. Based on its spectral pattern, the larger marker chromosome carrying a NOR represents duplicated material from RNO11. The spectral pattern of the smaller marker chromosome could not be assigned to a specific chromosome. The copy number changes of RNO5 and RNO7, as well as the above-mentioned structural changes in the diploid karyotype, were observed to occur twice in the tetraploid karyotype ( Figure S4). In a few tetraploid metaphases, additional sporadic translocations were also noted. Both common and sporadic structural changes in diploid and tetraploid metaphases are depicted in Table 5. Sporadic structural changes were detected in less than 50% of cells. Table 5. Common and sporadic structural chromosomal rearrangements in PAV-1 cells identified through spectral karyotyping (n = 20).

Short Tandem Repeat Profiling
To establish a characteristic STR profile of PAV-1 cells, we genotyped 31 established polymorphic markers that we previously used to profile two other rat HSC lines, namely HSC-T6 [19] and CFSC-2G [20]. The sum of generated STR profiles for PAV-1 cells revealed a unique pattern of allele sizes for the 31 markers that was different from those of CFSC-2G and HSC-T6 that we have recently established (Table 6, Figure S5). This analysis confirmed that the PAV-1 cell line is rat-derived and is free from mammalian interspecies contamination. Table 6. Comparison of STR profiles from PAV-1, CFSC-2G, and HSC-T6 using the 31 species-specific STR markers 1 .

SN
Marker Name

Discussion
The authentication of a novel immortalized cell line for research and clinical use is mandatory when using a cell line in biomedical research [45]. Cell misidentification and cross-contamination are fatal issues that lead to error-filled publications, false data, irreproducible results, and a substantial waste of money [12,46]. By using complementary search strategies, Horbach and Halffman identified 32,755 articles in 2017 that reported research results with misidentified cells that, in turn, were cited by approximately half a million other papers [12]. Moreover, there are estimates that 16.1% of all published papers use problematic cell lines that are either contaminated or misidentified [47]. Other estimates even suggest that misidentified cell lines in biomedical research can be close to 50% in some areas of the world [48,49]. In particular, experts in this field estimate that one-third of all human cell lines are thought to be misidentified [46].
In liver research, there are many examples of misidentified cell lines, including Chang liver cells, GREF-X cells, LO2, WRL 68, and many others [10]. Nevertheless, continuously growing cell lines are still useful model systems for modern medical research because they provide an indefinite source of biological material for experiments [50]. In addition, well-characterized cell lines have the potential to replace some animal experiments, thereby fostering the ethical 3R (Replacement, Reduction, Refinement) framework proposed by William M. S. Russell and Rex L. Burch in 1959 [9]. Nowadays, many journals and funding agencies, including the National Institutes of Health (NIH), request that the cells used in experiments be subjected to authentication testing prior to publication or before providing funding or grants.
In the past, we determined genetic details of different continuous HSC lines, including human LX-2 [18], mouse cell lines GRX [21] and Col-GFP HSC [22], and rat lines HSC-T6 [19] and CFSC-2G [20]. Herein, we extended these studies and characterized the rat HSC line PAV-1 in regard to its genetic characteristics. This cell line was originally described as a convenient model to study aspects of vitamin A metabolism in HSC that can produce functional retinoids from retinol [23,[51][52][53]. Moreover, a subsequent study found that treatment with palmitic acid alone or in combination with retinol significantly decreased cell proliferation and α-SMA expression, suggesting that PAV-1 cells might be suitable to study processes of HSC deactivation [40].
PAV-1 cells were originally introduced as an immortalized HSC line model with the capacity to convert retinol into retinoid acid [23]. The bulk mRNA-sequencing performed in our study confirmed that PAV-1 cells express important genes implicated in the building, storage, and hydrolysis of retinyl esters. Most important is the expression of retinoid acid receptors and retinol-binding proteins. Furthermore, important key genes that mediate the breakdown of β-carotene metabolism to vitamin A and genes involved in the bidirectional transport of vitamin A between extra-and intracellular RBPs are expressed in PAV-1. For example, Bco1 is expressed in PAV-1 and is the only vitamin A-producing enzyme mediating the central cleavage across the C15,C15 double bond adjacent to a canonical β-ionone ring site of carotenoids and β-apocarotenoids [54]. The expression of all these genes strongly confirms the notion that PAV-1 cells are an ideal experimental tool that can be used to study retinol metabolism.
In our study, we found that PAV-1 cells express GFAP both at mRNA and protein level. Although the expression of Gfap at mRNA level was extremely low, as assessed by the high C t value in RT-qPCR, Western blot analysis showed that the cells expressed GFAP protein quantities that were comparable to that found in HSC-T6. In contrast, in the initial study describing the establishment of PAV-1, the cells were found to be negative for GFAP in immunohistochemistry [23]. In the previous study, the authors used a rabbit polyclonal GFAP antibody (#G9269, Sigma-Adrich, St._Quentin Fallavier, France) directed against rat and human GFAP, while, in the present study, we used another antibody (#ab7779, Abcam, Berlin, Germany) that detected a protein that was~55 kDa in size and further confirmed with another antibody (#sc-33673, Santa Cruz Biotech., Santa Cruz, CA, USA). The overall reactivity of the antibody that was used in the previous study was significantly lower, preventing the identification of GFAP in immunocytochemistry. In this context it is worth noting that the rat gene Gfap produces several GFAP splice variants (GFAPα, GFAPβ, GFAPδ, GFAPε, GFAPκ). The difference between the splice variants lies either in the 5 -untranslated region in the case of GFAPα, GFAPβ, and GFAPγ or the 3 -untranslated region for GFAPδ, GFAPε, and GFAPκ [55,56]. Importantly, quiescent and culture-activated primary rat HSC GFAPα is the predominant form, while GFAPβ predominates in the SV40-immortalized cell line HSC-T6 [56]. All splice variants have a different half-life and are differentially regulated on a transcriptional level [56]. In our expression analysis, we used primers that bound to regions located in exons 7 and 8. With these primers, it is possible to amplify the transcripts GFAPα and GFAPβ, while the three other Gfap mRNAs that lack exons VIII are not amplified. Potentially, the inconsistency is based on the fact that, in PAV-1 cells, one of the other (GFAPδ, GFAPε, GFAPκ) mRNAs that are not amplified by our PCR strategy is expressed at higher levels than in HSC-T6.
Unlike the two previously published karyotypes of HSC lines, HSC-T6 and CFSC-2G, PAV-1 is an authentic spontaneously transformed hepatic cell line. In the current study, the application of conventional banding methods and SKY analysis was able to reveal several striking chromosome alterations in the PAV-1 cell line that include isochromosomes of RNO5, trisomy RNO7, monosomy RNO12, partial trisomy of RNO11, two marker chromosomes (one derived from RNO11), and cryptic translocations to RNO4, RNO11, and RNO12. At least four different chromosome breakages and reunions must have occurred during the formation of iso-chromosomes of RNO5 and the translocations involving RNO4, RNO11, and RNO12. However, due to technical limitations, small deletions, duplications, and inversions cannot be excluded by SKY analysis. The structural rearrangements identified in this study on RNO4, RNO11, and RNO12 appear to be specific to PAV-1 and are not observed in the karyotype of two HSC lines previously analyzed [19,20]. Interestingly, the gain of one copy of RNO7 and the loss of one copy of RNO12 are also noted in the previously analyzed HSC-T6 but not in CFSC-2G. Therefore, the analysis of different HSC lines can provide information on the recurrence of RNO7 trisomy and monosomy 12 during transformation. We have not addressed the impact of these chromosomal alterations on cellular features. However, it should be noted that the cell lines PAV-1 and HSC-T6 cells grow rather fast, while CFSC-2G cells grow significantly slower and require non-essential amino acids for optimal growth. Whether there is a link between the gain of one copy of RNO7 and the loss of one copy of RNO12 needs to be investigated in future studies. In particular, it will be essential to identify genes affected by these alterations and to clarify if this is only a coincidence in PAV-1 and HSC-T6 or if these changes are a prerequisite for the immortalized phenotype of these two cell lines.
In our study, we further analyzed the expression of typical HSC markers, showing that the PAV-1 cell line is capable of expressing collagen type I, desmin, caveolin-1, α-smooth muscle actin (α-SMA), vimentin, fibronectin, collagen type IV, and GFAP, while lacking the expression of the typical hepatocyte marker HNF-4α. Therefore, PAV-1 cells are suitable as an alternative HSC in vitro model for various scientific issues in contrast to isolated primary HSCs, which are associated with higher overall costs and always require the use of animals [1,9]. In addition, we showed that PAV-1 lacks the expression of SV40T, which is used to immortalize many other different HSC lines, including LX-1, SV68c-IS, A640-IS PAV-1 cells were initially described as a spontaneous immortalized cell line. However, it might be possible that PAV-1 cells were immortalized by the activity of mycoplasma infection. It is well known that the persistent infection of cells with mycoplasma can lead to malignant transformations in cell culture and provoke prominent chromosomal changes [64,65]. In our study, we performed mycoplasma clearance for two weeks and showed that PAV-1 can be effectively propagated without the presence of mycoplasma, suggesting that this bacterial contaminant is not required to maintain the immortalized phenotype.
Mycoplasma contamination can exhibit negative effects on cell morphology, proliferation, gene expression, and responsiveness to stimuli [66]. We cannot rule out that the clearance of mycoplasma has affected the properties of the original PAV-1 cell. However, many other studies have unequivocally shown that the cytotoxic effects of antibiotics against Mycoplasma species are reversible. For example, Plasmocin treatment for five passages has not produced any obvious alterations in human embryonic stem cells [67].
The elimination of mycoplasma in PAV-1 culture enables the safe application of this cell line in many laboratories and deposition in suitable cell line banks. Nevertheless, as recommended for all other established cell lines, regular testing for inter-and intra-specific cross-contamination and bacterial impurities is still highly recommended to prevent falsified research results, misleading publications, and waste of research money [68].

Conclusions
In sum, the biochemical characteristics and the genetic features described in this study are suitable means for the identification of PAV-1 cells. Together with our recent publications in which our team reported unique elements for the rat HSC lines HSC-T6 and CFSC-2G, there are now three continuously growing HSC lines available with established genetic hallmarks allowing discrimination from each other. As such, the scientific value of results established with these cell lines will increase the biochemical research results and help to fulfill the demands of scientific journals and funding agencies when working with continuously growing cell lines. Nevertheless, one should always be aware that these continuous cell lines represent an immortalized model, and these may show different results compared to primary cells. The additional use of isolated primary cells should be critically evaluated depending on the scientific question.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/cells12121603/s1, Table S1: Short tandem repeat (STR) marker analysis for AML12 cells; Table S2: Gene expression in PAV-1 cells under basal culture conditions as determined by next generation sequencing; Figure S1: Testing for Mycoplasma spp. infection in PAV-1 cells; Figure S2: Short tandem repeat (STR) profiling of AML12 cells; Figure S3: GTG-banded karyotype of a tetraploid PAV-1 cell; Figure S4: Spectral karyotyping of a tetraploid PAV-1 cell line metaphase; Figure S5 Funding: The laboratory of R.W. is supported by grants from the German Research Foundation (projects WE2554/13-1, W2554/15-1, and WE 2554/17-1). The funder had no role in the design of this article or in the decision to publish it.

Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.
Data Availability Statement: This manuscript contains most of the original data generated during our study. However, additional datasets (e.g., fastq data files of NGS analysis, additional illustrations of SKY painting) and results of repetitions of individual experiments are available upon reasonable request from the corresponding authors.