Immortalization of human primary prostate epithelial cells via CRISPR inactivation of the CDKN2A locus and expression of telomerase

Immortalization of primary prostate epithelial cells (PrEC) with just hTERT expression is particularly inefficient in the absence of DNA tumor viral proteins or p16INK4A knockdown. Here, we describe the establishment of immortalized normal prostate epithelial cell line models using CRISPR technology to inactivate the CDKN2A locus concomitantly with ectopic expression of an hTERT transgene. Using this approach, we have obtained immortal cell clones that exhibit fundamental characteristics of normal cells, including diploid genomes, near normal karyotypes, normal p53 and pRB cell responses, the ability to form non-invasive spheroids, and a non-transformed phenotype. Based on marker expression, these clones are of basal cell origin. Use of this approach resulted in the immortalization of independent clones of PrEC that retained normal characteristics, were stable, and non-transformed. Thus, this approach could be used for the immortalization of normal primary prostate cells. This technique could also be useful for establishing cell lines from prostate tumor tissues of different tumor grades and/or from patients of diverse ethnicities to generate cell line models that facilitate the study of the molecular basis of disease disparity.


Introduction
Prostate cancer is the most common cancer diagnosed in men, accounting for an estimated 20% of new cancer cases and 10% of cancer deaths in US males in 2020 [1]. Prostate cancer also represents one of the largest cancer disparities of mortality rates between non-Hispanic blacks and whites [2]. However, the molecular basis of prostate cancer development and these disparities remains unclear, at least in part due to the limitations of available cellular models. Unlike other cancer types, it has been particularly challenging to develop prostate cell line models in culture [3]. The most widely used cell lines, PC-3, DU145, and LNCaP, were established spontaneously and all derived from metastases [4]. Moreover, they have already acquired numerous genomic alterations [5]. To help define the minimal relevant genetic alterations needed for prostate cell transformation and 3D organoid aberrant growth, it is necessary to generate a collection of immortal normal prostate cell lines and develop a methodology that can be easily used for the efficient establishment of primary prostate cancer cells that often fail to establish in culture.
Prostate epithelial cells (PrEC) can only go through a limited number of passages before they become senescent, which is defined as a process in which cells stop dividing and undergo distinctive phenotypic changes [6]. Immortalization allows cells to evade senescence and continue to divide. Primary PrEC have been previously immortalized using a human telomerase reverse transcriptase (hTERT) transgene with limited success [7][8][9]. Most of these cell lines are unavailable, as they have been lost or do not show their described original characteristics. One of these cell lines, EP156T, has apparent normal epithelial features [8], but exhibits karyotypic abnormalities of unknown effects [10], which is consistent with further alterations being needed for immortalization. Of note, in two studies that generated hTERT-immortalized PrEC with near diploid karyotypes, the resulting cells lost expression of CDKN2A/p16 INK4A 8,9 . In one case, this was correlated with increased methylation of the p16 INK4A promoter [9]. In accordance with a potential requirement for p16 INK4A loss, Hahn and collaborators have shown that PrEC can be immortalized with ectopic expression of hTERT with co-expression of large T antigen, which inactivates the tumor suppressors pRB and p53 [11]. pRB is downstream of p16 INK4A , and both pRB and p53 are needed to establish a permanent cell cycle arrest in response to senescence signals [12,13]. Others have shown that p16 INK4A expression increases with PrEC passage [14][15][16]. Consistent with these observations, shRNA mediated knockdown of p16 INK4A cooperates with hTERT expression in the immortalization on PrEC [16]. More recently, it has been shown that hPrEC at very early passage (<7) can be immortalized via expression of hTERT alone when p16 INK4A expression is relatively low [14], but this limits the window of opportunity for immortalization of primary cells. Moreover, inactivation of p53 function, presumably downstream of p14 ARF , with a dominant negative mutant generates immortal cells that grow more efficiently [16]. Therefore, elimination rather than attenuation of p16 INK4A /p14 ARF expression may increase the efficiency of immortalization and at the same time eliminate pressure for selection of genetic alterations that could attenuate senescence as well as passenger alterations that are not required for immortalization but contribute to the selection of karyotypically deficient cells.
Thus, we devised a strategy to rapidly attenuate senescence signals simultaneously with the expression of hTERT to prevent spontaneous alterations that are dispensable for immortalization via CRISPR-mediated inactivation of the CDKN2A locus, which directs the expression of both p16 INK4A and p14 ARF [12,17]. This method facilitates direct immortalization of PrEC without depending on the spontaneous silencing of p16 INK4A expression or p16 INK4A /p14 ARF shRNAs, which may not be able to stably maintain low expression under all growth conditions. The immortal cell clones described here are of basal prostate cell origin and retain the characteristics of normal cells including normal p53 and pRB pathways, and near normal karyotypes.
For p53 pathway analysis, cells were treated with 150 nM flavopiridol, 60 µM etoposide, or 1 µg/mL doxorubicin and collected at the indicated time points. For contact inhibition assays, 150,000 cells were seeded in 6 cm plates, collected at indicated days and analyzed for western blotting and cell cycle DNA-content by PI/flow cytometry as described earlier [18].
For 3D spheroid culture, procedures were adapted from [19]. For hPrEC spheroids, Growth Factor Reduced Matrigel Matrix (Corning, REF 356231) was supplemented as described previously [20]. The bottom layer of Matrigel was 75% while the top was 50%. 2,000 cells of EP156T, hPrEC T-ΔN2A clone 1 or clone 2 were seeded in each well inbetween the Matrigel layers. Supplemented medium was added on top and replaced every other day for about two weeks.
Anchorage independent assays and clonogenic assays were performed as described earlier [21]. The bottom layer contained 0.6% agar. For clonogenic assays, 170 cells were seeded into a 6-well plate.
Telomere length was quantified by qPCR using the Absolute Human Telomere Length Quantification qPCR Assay Kit (ScienCell, Cat #8918) as per manufacturer instructions.
To verify CDKN2A targeting in immortalized clones, PCR was performed using primers targeting CDKN2A exon 2 (CDKN2A FWD: CTG TGC TGG AAA ATG AAT GC; CDKN2A REV: CTG GAA GCA AAT GTA GGG G) with annealing temperature at 55°C. Globin primers were used as control (Globin FWD: CAA CTT CAT CCA CGT TCA CC; Globin REV: GAA GAG CCA AGG ACA GGT AC).

Immunoblots, immunofluorescence imaging and confocal microscopy
Western blot analysis was performed as previously described [23], using antibodies indicated in Supplementary Tables 1 and 2. Cells for 3D spheroid culture were seeded in iBidi µ-slides (ibidi #81501). For immunofluorescence detection, spheroids were washed with PBS and fixed with 30 µl fixation/permeabilization solution consisting of 2% paraformaldehyde, 0.3% Triton X-100, 5 mM EGTA, and 1 mM MgCl 2 in PBS at room temperature for 20 min. Spheroids were washed with PBS three times, followed by blocking with 20% horse serum in PBS with 0.05% Tween 20 (PBS-T) at room temperature for 60 min. Samples were incubated with primary antibodies in PBS-T for 60 min at room temperature or at 4°C overnight, washed with PBS three times, followed by fluorescence-conjugated secondary antibodies mixed with phalloidin for 60 min at room temperature. After rinsing three times with PBS, DAPI was applied to samples for 15 min. Samples were protected from light until imaging with a Leica TCS SP8 confocal microscope with 63x lens.

Cytogenetic analysis and chromosome microarray analysis (CMA)
To determine the karyotype of the immortalized clones, exponentially growing cells were treated with 0.01 µg/ml colcemid as previously described [24]. Chromosome spreads were prepared and G-banded according to standard procedures [25]. For each clone, at least 10 metaphases were examined. CMA was performed using Thermofisher microarray CytoScan™ HD array (https://www.thermofisher.com/ us/en/home/life-science/microarray-analysis/cytogenetics-ana lysis-microarrays.html), which contains more than 2.6 million copy number markers of which 750,000 are "genotype-able" SNPs and 1.9 million are nonpolymorphic probes with 250 ng of total genomic DNA from each test sample as described earlier [26].

Generating immortalized human prostate epithelial cell line
Our attempts to establish immortalized primary normal human PrEC via ectopic expression of hTERT alone failed to generate any immortal clones following selection. This indicated that the introduction of a single hTERT transgene is very inefficient as a single hit for immortalization of PrEC, supporting why long-term successful stable immortalization of PrEC has been rarely accomplished [8]. Thus, we rationalized that ablation of the CDKN2A locus (Fig. 1a), which encode p16 INK4A and p14 ARF , two gene products known to be activated in response to senescence signals that drive pRB and p53 activation, would prevent senescence and facilitate rapid immortalization precluding spontaneous alterations. We also wanted to promote immortalization without using viral oncogenes that may have additional targets [27]. Moreover, we also rationalized that by not directly altering the pRB and p53 tumor suppressors, other properties of normal cells including cell cycle arrest in response to growth to high density, DNA damage, and other cellular stresses would be preserved.
To ablate CDKN2A in hPrEC, we designed a guide RNA (sgCDKN2A) for Cas9 to target exon 2 of CDKN2A, which contributes sequences for both p16 INK4A and p14 ARF (Fig. 1a). hPrEC were transduced with lentiviruses expressing hTERT and/or sgCDKN2A and selected with hygromycin and/or puromycin. One CDKN2A knockout clone (ΔN2A) and two clones of the combination of ectopic hTERT expression and CDKN2A knockout (T-ΔN2A) were successfully selected. No clones survived selection after transduction with hTERT lentiviruses alone upon multiple attempts. Non-transduced primary hPrEC grew sparsely with cells varying in sizes at passage 3 (Fig. 1b). Starting from passage 7, the majority of hPrEC showed senescence characteristics with enlarged cell nuclei and flat morphology. At passage 8, few cells remained attached. The ΔN2A clone started to show signs of senescence at passage 8 and became fully senescent at passage 9, leaving an insufficient number of cells for further characterization experiments. In contrast, two T-ΔN2A clones proliferated relatively fast, showing compact and smaller cells (Fig. 1b). The derived cell lines have been continuously grown and are currently at passage 36 without showing any signs of senescence.

Characterization of the immortalized cell lines
We next verified that the immortalized cell clones had the intended CDKN2A exon 2 deletion and expressed the hTERT transgene. qPCR performed with primers targeting the telomere and a single copy reference using genomic DNA showed that the relative telomere length in the immortalized clones ranged from 1.29-to 1.99-fold greater than the reference human genomic DNA, while the corresponding hPrEC ratio was only 0.52-fold (Fig. 1c). This strongly suggests that ectopic hTERT expression elongates/ maintains telomeres in immortalized hPrEC T-ΔN2A clones. Next, we used PCR to amplify the region of CDKN2A exon 2 flanking the guide RNA targeting site cut by Cas9. Comparing with the band amplified from hPrEC genomic DNA, clone 1 produced a CDKN2A PCR product migrating slightly slower than the control PCR product, indicating the presence of a small insertion into the Cas9 lesion during DNA repair (Fig. 1d). There were no amplified PCR products from clone 2 genomic DNA, suggesting that the genomic region deleted is larger than the region selected for amplification (Fig. 1d). As expected, primers for the human globin gene amplified a band of the expected size for all samples. Therefore, the CDKN2A gene was efficiently targeted by Cas9 and edited in both clones.
Next, we wanted to confirm that no products were expressed from the targeted CDKN2A locus. HEK293T cells were used as a positive control for p16 INK4A and p14 ARF expression, and a panel of PCa cell lines and EP156T, the only available hTERT-immortalized prostate epithelial cell line [8], as controls for other cell markers. We detected p16 INK4A expression in hPrEC at passage 4, but no expression of p16 INK4A was detected in either immortalized clone (Fig. 1e), suggesting that the CRISPR knockout efficiently prevented p16 INK4A expression. p14 ARF and p16 INK4A expression were absent in the two ΔN2A clones. p14 ARF expression was also not detected in hPrEC at passage 4, which may not be activated at early cell passages (note that these cells show p53 activation likely via a p14 ARF -independent mechanism). However, p14 ARF was readily detectable in PC3 and DU145 PCa cell lysates. These results confirm that CRISPR deletions targeting exon 2 efficiently eliminated expression of the two CDKN2A protein products.
hPrEC and its derived T-ΔN2A clones expressed high levels of basal cell markers cytokeratin 5 and p63 The sgRNA targets exon 2, which encodes a segment of both proteins using alternative frames. b Morphology of primary hPrEC in culture and clones obtained upon selection with antibiotics. Passage numbers are indicated. c Telomere length qPCR with relative and absolute length of telomeres in primary hPrEC and immortalized clones. d PCR to amplify CDKN2A exon 2, which was targeted by CRISPR sgRNA. e Western blot of p16 INK4A , p14 ARF and additional cell markers including those for basal (CK5) and luminal (CK18) tissue origin.
( Fig. 1e and Supplementary Fig. 1A) and undetectable levels of luminal cell marker cytokeratin 18 and AR ( Fig. 1e and Supplementary Fig. 1B). AR was not detected even after growth in the presence of 2 nM DHT and upon 24 h following EGF withdrawal [20]. This indicates that hPrEC-T-ΔN2A clones, like EP156T cells, are derived from prostate basal epithelial cells.
To determine if the immortalized T-ΔN2A clones maintained the features of normal cells, we examined their cell cycle parameters and karyotype. Cells from exponentially growing T-ΔN2A clones were stained with propidium iodide and analyzed by FACS. Similarly to immortalized EP156T cells, both T-ΔN2A clones lacked significant polyploidy, suggesting that the cells were diploid (Fig. 2a). Cytogenetic analysis showed that the karyotype of Clone 1 is mosaic: 46,XY,der (19) (19), that involves an unbalanced translocation with the distal end of the long arm of chromosome 5, 5q23.2→qter. A representative karyotype of the cells in Clone 1 is shown in Fig. 3b, which also includes an inset showing the der(19) from another metaphase with more elongated chromosomes. Chromosome microarray analysis (CMA) confirmed that the der (19) was an unbalanced rearrangement with a gain of 5q23.2→qter, as there were three copies of nucleotides chr5:121,598,528-180,719,789 and loss of one copy of a small segment in chromosome band 19p13.3, which includes nucleotides chr19:260,911-683,931 ( Fig. 2b and Supplementary Fig. 2). Notably, the CMA profile also revealed a focal deletion in 9p21.3, (9p21.3)x1, which appeared to result in loss of one copy of the CDKN2A locus in about 50% of the cells (Fig. 2b and Supplementary Fig. 2). Cytogenetic analysis of Clone 2 was limited by the low number of mitotic cells but revealed a mosaic karyotype in two separate metaphase harvests: 47,XY, +der(2;20)(p10; q10), +0-1mar [9]/46,XY [1]. This is consistent to what was observed by CMA analysis, which showed gains of 2p and 20q, i.e., (2p)x3 and (20q)x3 (Fig. 2c). The CMA analysis also revealed a gain of 9q31.1q34.3, which might be present in an unidentifiable marker chromosome noted in some metaphases on both occasions. Another abnormality seen in the CMA analysis, (4q34.3q35.2)x1, could not be identified with certainty in the available metaphases. Interestingly, the CMA profile also revealed a focal deletion in 9p, (9p21.3)x0, which resulted in homozygous loss of the CDKN2A locus. The abnormalities of 4q and 9p are shown in Fig. 2c and Supplementary Fig. 3.
Since clone 1 had a significant proportion of karyotypically normal cells, a chromosome 5q32 PDGFRB break apart FISH probe (MetaSystems, Newton, MA) was then used to screen subclones of Clone 1 that were generated by single cell dilution. However, all subclones exhibited 3 fused orange-green signals, indicating three copies of the 5q32 region ( Supplementary Fig. 4), with no subclones being present without this alteration. Therefore, it seems that the cells with the der (19) had outgrown the normal diploid cells by the time the cells were diluted to generate single cell clones. Note that single cell cloning started at passage 34 and involved >20 additional population doublings with no signs of senescence observed.
To determine if the immortalized hPrEC clones retain the properties of normal cells, we determined their response to signals that activate the p53 and/or the pRB pathway. First, we treated the cells with flavopiridol to induce p53 expression [28]. The expression of p53 dramatically increased 2 h after flavopiridol treatment, returning to near basal levels by 24 h (Supplementary Fig. 5), suggesting that p14 ARF -independent p53 signaling is not affected. We also assessed if the p53 pathway was intact in T-ΔN2A clones upon treatment with DNA-damaging agents. As shown in Fig. 3a, treatment of T-ΔN2A clones 1 and 2 with etoposide resulted in p53 upregulation by 2 h and subsequent potent upregulation of p21 by 6 h. Doxorubicin treatment resulted in potent upregulation of p53 expression that peaked at 6 h, followed by p21 upregulation at 24 h (Fig. 3b). This demonstrates that T-ΔN2A clones retain normal p53/p21 signaling in response to DNA damage.
Next, we determined whether growth to high cell density resulted in G1/G0 cell cycle arrest, which is known to trigger activation of pRB and related proteins [29,30]. Accumulation of cells with a G0/G1 DNA content was already observed with increased cell density at day 4 prior to reaching cell confluence, and prominent G0/G1 arrest was observed by day 7 when cells were fully confluent (Fig. 3c). Consistent with these results, the levels of mitotic cyclin B1 were undetectable and total pRB expression was downregulated while p27 expression was upregulated at day 7 compared to day 4 in both T-ΔN2A clones (Fig. 3d). We have consistently observed that these cells are highly sensitive to cell-to-cell contact inhibition of proliferation and when seeded at low concentration, they form colonies that become quiescent even if these colonies do not fully cover the plate surface (data not shown). Thus, our results indicate that the cells are strongly arrested by cell contacts, which is consistent with their epithelial origin.
To explore the differentiation potential of these immortalized cells and their ability to form normally shaped spheroids, we seeded them in Matrigel. T-ΔN2A clones and EP156T cells were cultured according to methodology based on previously described work [19,20]. EP156T and hPrEC-T-ΔN2A clones were able to establish round and similarly sized spheroids in Matrigel (Fig. 4a). After 13 days in culture, spheroids were fixed and immunofluorescence stained for prostate epithelial cell markers. The spheroids consisted of multiple cells, and individual cells were enclosed with F-actin. The boundary layer of the spheroids was high in CK5 and low in CK18 (Fig. 4b), indicating that both hPrEC-T-ΔN2A clones were polarized and of basal cell origin (consistent with the results in Fig. 1e).
We also stained the spheroids for Epithelial-Mesenchymal Transition (EMT) markers. It was reported that EP156T exhibited high E-cadherin and low vimentin expression in 2D culture [10]. Under our spheroid culture conditions, we observed positive E-cadherin staining in EP156T and both T-ΔN2A clones. Of note, vimentin staining was positive in a subset of cells facing the ECM, exhibiting clear polarity in EP156T and both immortalized T-ΔN2A clones (Fig. 4c). We did not detect expression of AR, even after EGF withdrawal [20].
Finally, clonogenic assays showed efficient colony formation by both T-ΔN2A clones, generating the expected number of colonies based on cell seeding. In contrast, hPrEC seeded at passage 7 failed to form visible colonies, consistent with these cells undergoing replicative senescence (Fig. 5a, b). Soft agar anchorage independence assays demonstrated that the cells are not transformed. Specifically, PC3 prostate cancer control cells formed numerous colonies as expected [21], while the immortalized clones remained mostly as single cells and were unable to form colonies (Fig. 5c, d).

Discussion
This report describes a method for the efficient immortalization of normal prostate epithelial cells that could be used to build a collection of immortal normal PrEC from men with diverse genetic ancestry. We have used the same methodology with a biopsy of primary cells from a prostate tumor and successfully established a cell line (Badal and collaborators, in preparation). Therefore, this methodology could also be used to immortalize patient-derived prostate cancer cells from patients with diverse genetic ancestry, which is a major limitation of the existing PCa cell line collection. Such cell lines could potentially serve as cellular models to help address PCa disparities [31,32].
Our preliminary data indicated that introduction of a single hTERT transgene is extremely inefficient as a single hit for immortalization of primary prostate epithelial cells from normal or cancer tissue. Of the three hTERT immortalized primary "normal" cell lines initially reported by others [7][8][9], one has a very abnormal karyotype [7], one is unavailable [9], and the one that is readily available shows the properties of normal cells [8] but exhibits some cytogenetic abnormalities and heterogenicity [10]. Two of these cell lines have lost expression of the tumor suppressor p16 INK4A during immortalization. Previous work has shown that senescent primary human fibroblasts and epithelial cells do not reenter the cell cycle upon hTERT expression. In contrast, senescent cells with low, but not high, expression of p16 INK4A reenter the cell cycle following inactivation of p53 [33]. Since primary PrEC transduced to express hTERT become senescent and fail to efficiently immortalize, and this can be bypassed by expression of SV40 large T antigen, which inactivates p53 and pRB [11], it appears that either hTERT cannot re-stabilize telomeres quickly enough to prevent senescence or that senescence is triggered by other environmental cues that activate p14 ARF , the upstream activator of p53 [34]. Consistent with these results, it has been shown that shRNA-mediated knockdown of p16 INK4A allows hTERT immortalization of PrEC, and in this context, elimination of p53 function with a dominant-negative p53 mutant generates immortal PrEC that proliferate faster [16], suggesting that telomere independent senescent signaling through p53, possibly mediated by p14 ARF upregulation, can still attenuate the proliferation of these cells. Based on these reports and our preliminary work, we hypothesized that expression of hTERT simultaneously with the co-inactivation of the CDKN2A locus would bypass the senescence signals that preclude immortalization in normal prostate epithelial cells. Our results support this hypothesis and provide a greatly improved methodology for immortalization that could also be applied to other cell types that prove difficult to immortalize. It remains to be determined if simultaneous inactivation of both p16 INK4A and p14 ARF is absolutely required, but based on previous observations this appears likely [11,16,33,34]. Moreover, alterations of the CDK2N2A locus are detected at about 2% of PCa [35][36][37][38], suggesting that these alterations are also selected in prostate tumors in vivo. It is tempting to speculate that primary cells from tumors with alterations in the CDKN2A locus that already express hTERT may spontaneously establish in culture.
The cell lines described here provide a starting model for stepwise transformation assays aimed at determining prostate-relevant oncogene/tumor suppressor gene cooperativity as well as ECM invasion in 3D organoids. Moreover, the immortalization methodology could be applied to the rapid immortalization of cells in tumors and adjacent tumor tissue for studies of organoid formation and drug response. Finally, immortalization results in stable cell lines that retain the characteristics of the cells of origin, but are much easier to grow and cost effective for long term passage and further manipulation. This should facilitate establishment of new cell lines even in laboratories with limited resources around the globe. Nakajima for primary PrEC expansion and the FCCC Cell Culture and Genomics Facilities.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.