Expanding the Prostate Cancer Cell Line Repertoire with ACRJ-PC28, an AR-negative Neuroendocrine Cell Line Derived From an African-Caribbean Patient

Prostate cancer cell lines from diverse backgrounds are important to addressing disparities in prostate cancer incidence and mortality rates among Black men. ACRJ-PC28 was developed from a transrectal needle biopsy and established via inactivation of the CDKN2A locus and simultaneous expression of human telomerase. Characterization assays included growth curve analysis, immunoblots, IHC, three-dimensional cultures, immunofluorescence imaging, confocal microscopy, flow cytometry, whole-genome sequencing (WGS), and RNA sequencing (RNA-seq). ACRJ-PC28 has been passaged more than 40 times in vitro over 10 months with a doubling time of 45 hours. Short tandem repeat profiling confirmed the novelty and human origin of the cell line. RNA-seq confirmed the expression of prostate specific genes alpha-methylacyl-CoA racemase (AMACR) and NKX3.1 and neuroendocrine specific markers synaptophysin and enolase 2 (ENO2) and IHC confirmed the presence of AMACR. Immunoblots indicated the cell line is of basal-luminal type; expresses p53 and pRB and is androgen receptor (AR) negative. WGS confirmed the absence of exonic mutations and the presence of intronic variants that appear to not affect function of AR, p53, and pRB. RNA-seq data revealed numerous TP53 and RB1 mRNA splice variants and the lack of AR mRNA expression. This is consistent with retention of p53 function in response to DNA damage and pRB function in response to contact inhibition. Soft agar anchorage-independent analysis indicated that the cells are transformed, confirmed by principal component analysis where ACRJ-PC28 cells cluster alongside other prostate cancer tumor tissues, yet was distinct. The novel methodology described should advance prostate cell line development, addressing the disparity in prostate cancer among Black men. Significance: Cell line development continues to attract less than 10% success rate. More than 98% of prostate cancer cell lines are from White men. This may contribute to the poorer response by Black men with prostate cancer to therapy compared with White men with prostate cancer, increasing overall survivorship among White men. The methodology described here to develop ACRJ-PC28, should advance the presence of Black prostate cancer cell lines thereby addressing prostate cancer disparity.


Introduction
The use of in vitro cancer and normal models continues to be a yardstick for understanding the biological drivers of cancers with the goal of developing diagnostic, prognostic, and theragnostic tools. While in vivo models still provide more accurate predictions of human response, they are often used with human-derived two-dimensional (2D) models in the form of cell line-derived xenografts. Moreover, many factors support the complete transition to the application of in vitro models for cancer research; the protection of animals, more expedited and accurate identification of drug leads warranting clinical trials, and more cost effective and less labor-intensive approaches to anticancer research. Still, in vitro models have a long way to go before they can fully recapitulate in vivo models, whereby systems such as the immune response (1) and extracellular matrix (ECM; ref. 2) are factored in when developing tools for prostate cancer diagnosis and treatment. Technologies like organ on a chip (3) are a step in this direction, notwithstanding, the use of simple 2D models, namely cell lines, remain crucial to these. While there is an understanding that three-dimensional (3D) in vitro models better recapitulate the phenotype of cells in their original microenvironment (4), the cost of maintaining 3D models for long-term interrogation remains a limitation of their utility (5). Humanderived 2D models still remain relevant in basic science and are still the most common research model (6).
Prostate cancer has one of the greatest disparities in incidence and mortality rates (7-9) and we previously described (8) the possible reasons for this. What remains clear is the lack of equal representation of Blacks in in vitro models (10) and in clinical trials (11) when ascertaining the drivers of the disease and during the process of developing treatments. The rate of prostate cancer mortality is on the decline in the United States (12), but this is more evident among White men (13) while Black men in regions such as the Caribbean and Africa continue to experience elevated and increasing incidence and death from prostate cancer (14)(15)(16).
We believe that the underrepresentation of Black men at all stages of research, preclinical and clinical, is a contributing factor to current chemotherapy drugs being more effective in treating prostate cancer among White men (17), https://corporate.dukehealth.org/news-listing/new-non-hormonaltarget-identified-advanced-prostate-cancer. Although the survival rate of prostate cancer among Black and White men is significantly improved when the disease is diagnosed early, Black men are known to still have poorer outcomes than their White counterparts (11). Designing therapies with more effective outcomes must consider a comprehensive assessment of patients, including their geographic locale and ethnic background. The use of cell lines in large numbers is believed to hold the key to predicting drug leads more accurately (18,19). The absence of cell lines before now from regions where the incidence and mortality rates of prostate cancer are among the highest is unfavorable to addressing the disparity of prostate cancer among Black men in general, especially those from the Caribbean. We report the first cell line from the Caribbean, a region presenting with the second highest incidence and mortality rates of prostate cancer among Black men.

Sample Collection
Once the project obtained ethical approval (ECP, 91, 13/14) from the local institution board, Mona Campus Research Ethics Committee patients provided their written informed consent in accordance with the Declaration of Helsinki.
We selected patients with PSA levels >4 ng/mL who were undergoing guided transrectal needle biopsy (TRNB) of the prostate. Also selected were patients with locally advanced or metastatic prostate cancer undergoing transurethral resection of the prostate (TURP) to relieve bladder outlet obstruction that corresponded to the inclusion criteria: males >18 years of age, Jamaican by birth or heritage of naturalization, patient diagnosed with prostate cancer, patients consenting to the study, and patients being followed at either The University Hospital of the West Indies or The Kingston Public Hospital. Patients who were diagnosed with a contagious illness, which might put the researcher at risk (e.g., HIV/AIDS), and those who were undergoing medical therapy known to affect serum PSA levels were excluded from the study. Samples were either collected from biopsy cores that were selected on the basis of abnormal PSA levels and/or palpable abnormality of the prostate suspicious for cancer on digital rectal examination or from TURP samples that were selected on the basis of a previously confirmed diagnosis of prostate cancer. All samples (TURP and TRNB) were collected and transported in RPMI (ATCC; #30-2001) supplemented with 5% FBS (ATCC; #30-2020), 1% penicillin/streptomycin (ATCC: #30-2300), and 0.1% amphotericin (ATCC; #30-2301) and processed within 3-4 hours of collection. Specifically, ACRJ-PC28 was developed from a TRNB sample. A total of 28 samples were collected over 5 years utilizing different culture conditions to realize ideal culture conditions for prostate cancer cells (Supplementary  Table S1).
After rinsing three times with PBS, DAPI was applied to samples for 15 minutes. Samples were protected from light until imaging with a Leica TCS SP8 confocal microscope with 63x lens. Androgen receptor (AR) immunofluorescence of cells growing on glass slides was performed as described previously (21). Briefly, the cells were seeded on coverslips and cultured with medium containing DHT for 24 hours before fixation with 4% paraformaldehyde, followed by permeabilization with 0.1% Triton X-100 in PBS and blocking with 3% BSA. The cells were incubated with AR primary antibody for 1 hour at room temperature and secondary antibody as well as phalloidin for 45 minutes. Coverslips were mounted with mounting buffer containing DAPI to the slides and sealed with nail polish before imaging.

IHC
To compare the original tissue sample to the cell line and to further confirm that the cells are prostate in origin, several antibodies [AR, chromogranin A, synaptophysin, ER, E-cadherin, keratin 8-18, NKX3.1 (NK3 homebox), PSA, p16, p27, p53, p63, Prostate Cocktail (AMACR/p63/HMWCK), CD44, c-MYC, Ki67, and MUC1] were evaluated via IHC on the original TRNB prostatic cores and the derived ACRJ-PC28 cells at passage 7. The original TRNB cores were treated by submerging in 10% neutral buffered formalin for preservation after which the cores were processed, embedded in wax, and serially sectioned to a thickness of 4 μm then stained with hemotoxylin and eosin stain. After prostate cancer diagnosis, screening for the antibodies was outsourced to Vitro Molecular Laboratories. Tissue pretreatment was performed in either citrate pH 6-buffer or online (BondIII), followed by incubation with antibodies, polymer-based peroxidase detection (Novocastra), and DAB in an automated platform (Autostainer, Thermo Fisher Scientific or Bond III, Leica). Histopathologic analysis of the antibodies was done using a Nikon Eclipse Ci research microscope (Nikon Instruments Inc., Americas).
Given the immunophenotypic expression of the original patient sample, the ACRJ-PC28 cells were subjected to an IHC panel, aimed at possible clarifying cellular lineage, and identifying cellular markers indicative of malignancy. The stains utilized were nuclear and cytoplasmic prostatic epithelial cell markers (AR, NKX3.1, and PSA), indicators for neuroendocrine (NE) cell differentiation [chromogranin A and synaptophysin (SYP)], a marker of cellular proliferation (ki67), and a marker for basal cell expression (p63). In addition, surrogate markers for neoplastic transformation, that is, cellular adhesion and intranuclear dysregulation (CD44, c-MYC, and BCL-2) were also utilized. Cells were cultured until 80%-90% confluence was achieved to receive at least 10 7 cells or more. Adherent cells were then collected and fixed in 10% formalin then resuspended in PBS at 4°C. Cells were immobilized in an agarose plug and transferred to a histology cassette for downstream processing and paraffin embedding, to facilitate sectioning of 3-5 μm FFPE tissue sections for histochemical (HC) and IHC testing using required HC stains, biochemical markers, or mAbs. The slides were sent to Vitro Molecular Laboratories for biomarker analysis mentioned prior in this section.

Soft Agar Assay
Soft agar assay was used to assess the tumorigenicity of ACRJ-PC28 (23). A stock of 1.8% noble agar (Millipore Sigma, #A5421) was prepared in Hank's Balanced Salt Solution and autoclaved for sterilization. Bottom agar (0.6% noble agar) was prepared by mixing 1.8% stock solution with medium and then aliquoted into a 24-well plate to set for 10 minutes. Top agar (0.3% noble agar) was prepared by combining 1.8% noble agar with medium and ACRJ-PC28 cells at passage 7 (250, 1,000, or 5,000 cells). Top agar was solidified for 30 minutes after which 100 μL of medium with 10% FBS was added to each well. Media were replenished weekly, and colonies grown for 3 weeks after which they were fixed in methanol and visualized using 0.15% crystal violet.

Contact Inhibition, Cell-cycle Analysis, and Response to Doxorubicin Treatment
On day 0, 150,000 cells were seeded in 6 cm plates and collected at the indicated times as cell density increased. The percentage of cells at different phases of the cell cycle was determined by measuring DNA content via propidium iodide (PI) staining followed by flow cytometric analysis (24 To conduct PCA, sequence reads of 50 The Cancer Genome Atlas Prostate Adenocarcinoma (TCGA PRAD) tumor and normal pair samples were aligned to human genome (hg38) using STAR (30). The number of raw counts in each known gene from the Ensembl database was enumerated using htseqcount from the HTSeq package (31). Differential expression between normal prostate and ACRJ-PC28 samples was assessed for statistical significance using the R/Bioconductor package DESeq2 (32). Genes with P value <0.01 and a fold change ≥1.5 were used for PCA including ACRJ-P28 RNA-seq data.

Cytotoxic Response to Endemic Jamaican Cannabis Sativa Tetrahydrocannabinol Extract
The cytotoxicity of three tetrahydrocannabinol (THC) extracts (THC_EX1, THC_EX2, and THC_EX3) from the Jamaican Cannabis sativa plant was assessed using the MTS assay as described in G Malich, B Markovic, and C Winder (33) against novel ACRJ-PC28 cells and established PC-3 cells. ACRJ-PC28 cells and PC-3 cells were grown to 90% confluency, after which they were harvested and later reseeded at 20,000 cells per well in a 96-well plate followed by the addition of the extracts at concentrations 3-80 μg/mL for 24-hours.
Absorbance readings were taken at 490 nm.

Data Availability
RNA-seq expression and WGS data were uploaded in the Gene Expression Omnibus (GEO; GEO:GSE213223) and SRA (SAMN30844989) databases, respectively. Until the cell line, ACRJ-PC28 is available on the ATCC and ECACC platforms, it can be accessed upon request from the corresponding author.

Results
We processed 28 prostate cancer samples over a 5-year period before successfully developing ACRJ-PC28 that was obtained from a 69-year-old Jamaican male with a PSA of 441.2 ng/mL, diagnosed with prostate cancer adenocarcinoma (Gleason score; 4+4) with perineural involvement. To confirm the ancestry of the cell line, admixture analysis using YRI and CEU from 1000 genome project was used as a reference, which indicated 97% African ancestry and male by birth. TRNB cores obtained from the patient's prostate were successfully grown to generate a primary prostate cancer explant which was then transduced with pLV-hTERT-IRES-hygro and lentiCRISPRv2-sgCDKNA lentiviruses to establish an immortal ACRJ-PC28 cell line. We adapted numerous protocols (34)(35)(36)(37)(38)(39)(40)(41)

Generation of Primary Explants
Where mentioned in the literature, we recognized that the primary means of dissociating prostate cancer biospecimens applied in protocols (10) that describe the 32 currently established prostate cancer cell lines is that of enzymatic digestion followed by xenograft models. Attempting enzymatic dissociation with collagenase and hyaluronidase cocktails at various concentrations as detailed in Niranjan (34) and other protocols (35-41), gave us 0% success to

FIGURE 1
Steps utilized to generate ACRJ-PC28. The development of ACRJ-PC28 was primarily achieved by overcoming three challenges, (i) the identification of the appropriate media combination and ECM, (ii) overcoming the presence of TAF, and (iii) inducing immortalization using CDKN2A knockdown and hTERT upregulation.
develop our primary explants on the first 13 prostate cancer TURP samples.
This led us to omit initial digestion steps from our protocols similar to Latimer (42) to keep prostate cancer fragments in an organoid-like form with the intention of mimicking in situ tumor microenvironment. Combining this step with the use of fibroblast feeder cells resulted in an almost 100% success rate in generating prostate primary explants. Interestingly, when we attempted to combine digestion steps with the use of the feeder cells, we observed no prostate primary explants. Perhaps, the use of digestion to generate primary explants for cell lines Bob, CA-HPV-10, MDA PCa 2b, RWPE-1, PZ-HPV-7, and HPE-15 was successful given the more resilient nature of the tumors they arose from.  Fig. 2. Figure 2 also includes some of our successfully generated explants, ACRJ-PC13 and ACRJ-PC17, which did not transition into cell lines. Collagen type 1 was also an effective inducer of tissue fragment attachment, generating microexplants when used to coat culture vessels at 0.01%-0.02%, albeit less efficiently than Swiss albino 3T3 irradiated feeder cells with a 20%-30% attachment rate.

Managing Tumor-associated Fibroblast Growth
The second hurdle we encountered was the presence of tumor-associated fibroblasts (TAF) that we observed at approximately 10 days or more among our primary explants. These TAF restricted the expansion of our explants through space ( Fig. 2A) and resources. We observed that higher serum concentrations favored the viability of both our prostate primary explants and TAF. Decreasing the FBS concentration to 10% or below did reduce the presence of TAF but had negative effects on the morphology of our prostate primary explants. We observed a distorted morphology (Fig. 2B), and a 2-fold increase in cytoplasmic mass combined with earlier senescence (passage 3-4) compared with samples that were maintained in 15%-20% FBS that lasted up to the sixth passage. Next, we attempted mechanical removal of TAF, but this was ineffective as residual cells repopulated the culture. Differential trypsinization was also ineffective as   which was typically after our prostate primary explants (∼2-3 days). We then mechanically removed tissue chips at around day 5.

Immortalizing Prostate-derived Explants-establishing ACRJ-PC28
The final hurdle was overcoming the eventual senescence of our cells. Of all the 28 samples we processed, none were spontaneously immortal, which is expected as spontaneous immortalization is a 1 in 10 7 occurrence (43) (20), that is believed to provide a less transformative approach to immortalization compared with HPV-16/18 or SV-40 large-T/Small-t transduction methods given the normal features of the cell cycle are maintained when applying this method of immortalization to normal prostate epithelial cells.
What we routinely observed was that normally, our prostate extended explants senesced between passages 5-7. Described in Zhao and colleagues (20), was the principle that expression of p16 INK and p14 ARF from the CDKNA locus and the lack of hTERT expression induce culture stress and replicative senescence, respectively. Thus, targeting both aids cells to bypass senescence. This proved successful in generating ACRJ-PC28 shown in Fig. 2C. CDKNA ablation or hTERT upregulation individually proved insufficient to immortalize our prostate-derived explants. We also attempted the transduction of dominant-negative TP coupled with hTERT expression, but this was also ineffective.

Antibiotics Use is Associated with Earlier Onset of Senescence
We also sought to minimize the use of antibiotics in culture as numerous studies indicate that beta-lactam antibiotics, which include penicillin and aminoglycosides such as streptomycin, may affect gene expression and regulation, inducing antiproliferative effects in eukaryotic cells (44,45). We observed that culturing prostate explants in media supplemented with penicillin and streptomycin experienced earlier onset of senescence, as early as passage 3 compared with samples grown without penicillin/streptomycin that senesced at around passages 5-7. As such, we adopted a model that limited the use of penicillin/streptomycin and amphotericin to only at the initial processing of tissue, after which all cells were grown in antibiotic-and antimycotic-free media.

Authentication Profile of ACRJ-PC28
STR profiling (Supplementary Fig. S1) characterized ACRJ-PC28 as a novel human cell line by comparison with the DSMZ Human and Animal Cell Lines Database. Mycoplasma testing also proved that ACRJ-PC28 was negative for Mycoplasma contamination. We were unable to compare the original patient sample with the cell line as there was insufficient and fragmented DNA yield from the original FFPE block, which resulted in a lack of amplification of the human markers.

Immortalization Effects
Further characterization assays confirmed the absence of p16 and p14, downstream products of the CDKNA gene that was knocked down (46) during the immortalization process. At the protein level, complete inactivation of p16 and p14 was achieved as ACRJ-PC28 (Fig. 3) had undetectable levels compared with positive controls DU-145, LNCaP, PC-3, and HEK293T (Fig. 3A). CNVnator indicated a possible heterozygous deletion at region chr9:21970801-22006400, which corresponds to the genomic region of exon 1 and exon 2 of CDKNA. Interestingly, at the protein level, using IHC on the original patient sample, p16 expression and nuclear staining were limited to <1% of the cancer cells, with focal expression identified in some benign epithelial cells as well as stromal cells ( Supplementary Fig. S2).
Because pRB protein is readily expressed in ACRJ-PC28 cells, which was also confirmed using RNA-seq data, we assessed whether high cell density caused G 1 -G 0 cell-cycle arrest (Fig. 4A), which is known to prompt activation of pRB and upregulation of p27 (47,48). ACRJ-PC28 cells demonstrated sensitivity to increased cell-cell density comparable with immortalized normal hPRECs (EP156T and hPrEC clone 1), as evidenced by accumulations of cells with G 0 -G 1 DNA content and a corresponding depletion of cells in S-G 2 -M at day 4 and most prominent at day 7 when cells reached full confluency (Fig. 4A). Consistent with these results, the levels of total pRB and mitotic cyclin B1 expression were depleted at day 4 and undetectable at day 7, while p27 expression peaked at day 7 (Fig. 4B). We also observed that when cells remained confluent for an extended period, cell death is detected (e.g.: hPrEC clone 1 in this experiment).
Altogether, our results show that despite their cancerous origin, ACRJ-PC28 cells can be arrested by cell contacts when grown to confluency, which is consistent with their epithelial origin. This observation has been made for other cancer cell lines including T98G glioblastoma cells, which are transformed but demonstrate a normal pRB checkpoint (49,50). Becaise ACRJ-PC28 cells express p53, we determined whether these cells exhibit a normal p53 response to genotoxic stress. As shown in Fig. 4C, treatment with the cytotoxic agent doxorubicin led to clear upregulation of p53 expression, although it was comparatively lower to the effect seen in normal immortal PrEC and LNCaP cells. The migration of the pRB band increased after 24 hours, which is indicative of dephosphorylation and associated with its activation downstream of p53.

ACRJ-PC28 Cells are Transformed and Form Organoids
ACRJ-PC28 demonstrated successful anchorage-independent growth on soft agar, generating colonies at seeding densities of 5,000 cells per well in a 24well plate as seen in Fig. 5. Lower densities of 250 and 1,000 cells per well failed to produce any surviving cell aggregates over the 3-week growth period. Plating efficiency was 23% with colonies growing to maximal size of 40 μm.
Furthermore, ACRJ-PC28 organoids (Fig. 6) expressed protein expression of ecadherin, integrin β1, and laminin β1 that were encased in the organoids seen in Fig. 6D. This was consistent with the original patient sample that also expressed E-cadherin (Table 1; Supplementary Fig. S2).
Doubling time assessments demonstrated that ACRJ-PC28 had a doubling time of 45 hours as seen in Fig. 3C, which was slower than popular prostate cancer cell lines, PC-3 (∼29 hours) and DU-145 (∼34 hours), but as fast as LNCaP (∼28-60 hours), while being almost twice as fast as that of popular African American cell line MDA PCa 2b (∼80 hours).

As Prostate Cancer
To confirm that ACRJ-PC28 is a prostate cancer cell line, we first examined at the protein level, the presence of prostate specific biomarkers, AR, AMACR, NKX3.1, and PSA (Table 1). AR levels were undetectable using western blot ( Fig. 2A) and immunofluorescence (IF) staining (Fig. 6B). The formation of ACRJ-PC28 organoids in Matrigel and expression markers as described previously was also used to determine the presence of AR. ACRJ-PC28 organoids were not invasive as compared with PC-3 organoids that started to develop protrusions by days 6-7 (Fig. 6A). In agreement with ACRJ-PC28 cells grown in 2D culture, AR expression was not detected (Fig. 6B). The remaining biomarkers (AMACR, NKX3.1, and PSA) are routinely used in IHC analysis and so were evaluated in the original patient sample and ACRJ-PC28 (Table 1) (52)] were more highly expressed in ACRJ-PC28 (Fig. 7). Despite seven of nine expressed NE genes, the absence of the benchmark NE markers such as chromogranin and SYP begs for further interrogation into the possible NE phenotype of ACRJ-PC28 and validates the need for diverse prostate cancer cell lines. IHC confirms that the cell line demonstrated strong diffuse nuclear expression for CD44 and c-MYC in 100% of the cells and approximately 50% of the cells expressed MUC1. The Ki67 indicated proliferative activity in roughly 20% of the ACRJ-PC28.
Next, we conducted a PCA using the RNA-seq data of 50 TCGA prostate normal and tumor paired samples and ACRJ-PC28. The analysis was adjusted by race and Gleason score. We used the top 169 genes that varied the most between normal and tumor samples. Results as seen in Fig. 8 showed that ACRJ-PC28 closely clustered with tumor samples from TCGA PRAD rather than normal samples. Interestingly, even though ACRJ-PC28 was among the tumor samples, it stood alone indicating the uniqueness in the cell line's genetic profile.

Discussion
We believe the methodology described here should increase the success rate of prostate cell line development from biospecimens obtained from either

FIGURE 7
Seven of nine neuroendocrine markers are more highly expressed in ACRJ-PC28 compared with PC3.
TURP, TRNB, or radical prostatectomy surgeries. We report >90% success rate in generating primary explants that survived up to passage 7, and the immortalization protocol described in Zhao and colleagues (20) is believed to provide a reduced transformative approach to prostate cell line development. This is especially important considering the prior absence of cell lines from the Caribbean, furthermore, the only 2 men of African ancestry (MAA) prostate cancer cell lines commercially available for research are from the United States.

AACRJournals.org
Cancer Res Commun; 2(11) November 2022 Considering this, our novel method could be further tested with other tissue types, increasing the complement of tools for generation of primary cell lines from multiple tissue types. A major contributor to the negative growth effects observed with penicillin/streptomycin use could be related to the connectivity between penicillin/streptomycin and activating transcription factor  (ATF).
AFT functions as a stress response mediator and is involved in multiple oncogenic pathways including AR signaling that controls prostate cell proliferation (55). Studies have reported an upregulation in the differentially expressed gene ATF in penicillin/streptomycin conditions (56). Considering our observations and evidence in literature, this further emphasizes the importance of omitting penicillin/streptomycin from primary cell culture protocols while optimizing culture conditions by using aseptic techniques in a sterile work area along with sterile reagents and media (57).
Ideally, the patient sample used to develop a cell line should also be used for STR profiling to confirm the source of the cell line in addition to being used for pathology reports and IHC staining. This model, however, seems more feasible when collecting tumor specimen from radical prostatectomy surgeries which could provide >5 g of tissue but would require a pathologist on-site to confirm the sample used for developing the cell line is the same for STR profile and IHC stains. This set-up continues to be a challenge for us at The UWI, Mona and for the duration of this project, we did not access samples from radical prostatectomies. This left us with TURPs and biopsies and for both, the lack of sufficient tissue sample persists because cell line protocols vary among patient samples requiring careful planning to ensure that the appropriate ECM/s, media, growth factors, and incubation times are employed. These can easily exhaust tissue samples which could otherwise be used for authentication and characterization purposes. Moreover, biopsied samples, the type used to generate ACRJ-PC28 yield relatively less tissue (200 mg vs. 1-2 g from TURPs) and because 12 cores are usually taken, it is likely that the sample used for IHC staining on the original patient sample is different from the core used to develop the cell line. It is important to address these so that a wider cross-section of prostate cancer cell lines can be generated, not only increasing the representation among races but prostate cancer grades as well. Perhaps, optimizing protocols for DNA extraction from FFPE is one route to addressing these. Despite our attempts, DNA content was either insufficient or highly fragmented resulting in our inability to confirm the source of ACRJ-PC28.
Not having confirmation that ACRJ-PC28 was from the original patient sample that definitively displayed tumor throughout, was the first challenge to appropriately demarcate the cell line as prostate cancer. The second was the lack of expression of AR at both the protein and RNA-seq levels. This led us to explore other prostate-specific biomarkers AMACR, NKX3.1 and PSA, and AMACR was expressed at the protein level while AMACR and NKX. were expressed at the RNA-seq level. Interestingly, both AMACR and NKX3.1 were expressed in the original tissue sample, and it is possible that the biopsied core used to generate the cell line is different from the one used for IHC analysis. However, if the cell line was developed from the same core analyzed, the lack of expression of NKX3.1 at the protein level could be due to protein degradation. In any event, the factors surrounding the relationship between mRNA and protein expression remain to be fully elucidated (58). The further RNA-seq expression of prostate related genes, PSCA and SLCA, are sufficient to deem ACRJ-PC28 a prostate-derived cell line, especially because NKX3.1 was the only prostate-specific biomarker detected in the Bob prostate cancer cell line (35).
The lack of expression of the prostate-specific genes (AR, NKX., and PSA) at the protein level prompted the investigation of a possible NE prostate can-cer phenotype, which was confirmed at the RNA-seq level when 2 NE specific marker (SYP and ENO2) were expressed at RNA-seq. Specific analysis of mRNA expression in normal prostate tissue revealed expression of both ENO2 and SYP albeit at low levels; however, no expression was observed for both following IHC staining (59) similar to our observation for ACRJ-PC28. The mechanism behind this incongruity remains to be elucidated. Furthermore, the overexpression of seven of nine NE-associated markers seen in ACRJ-PC28 compared with the PC-3 cell line further supports an NE phenotype of ACRJ-PC28 (52). It is common for a NE prostate cancer phenotype to not express AR and PSA and common for amplification of the genes aurora kinase A (AURKA) and N-myc (MYCN (51), both of which were overexpressed in ACRJ-PC28 at RNA-seq level and MYC was detected at the protein level. It is also possible that ACRJ-PC28 represents a prostate cancer NE phenotype as opposed to a normal NE phenotype as described by Yan and colleagues (60). This was addressed in four ways. First, a normal NE phenotype usually stains positive for basal markers like CK5 (61), while a prostate cancer NE usually stains for luminal markers like CK18 (62). Interestingly, ACRJ-PC28 expressed basal (p63 and CK5) and luminal (CK18) markers in 2D culture, confirmed by both immunoblots and IHC stain. The lack of expression of the CK18 marker when assessed in 3D models is not surprising as the differential expression of proteins among 2D and 3D models is established (63). Furthermore, 3D models are believed to recapitulate more in vivo conditions. This possibly means that ACRJ-PC28 represents a transitional prostate cancer cell line. Second, the aggressive proliferative nature of the prostate cells (64) (65). Furthermore, the tumorigenicity of ACRJ-PC28 was determined by an anchorage-independent growth assay, which confirmed that the cells are transformed. Indeed, our ongoing characterization of the cell line will provide further insight, especially through in vivo studies. Third, a normal NE prostate phenotype typically does not express BCL-2 (66) and ACRJ-PC28 demonstrated overexpression of BCL-2. Furthermore, inhibitors of BCL-2 were more effective in treating prostate cancer NE than AR-positive prostate cancer (67). Also identified is MYCL as a novel driver of prostate cancer NE phenotype (67) which too was expressed in ACRJ-PC28 at RNA and protein levels. Fourth, AMACR is only expressed in prostate cancer NE cells, and this gene was expressed in ACRJ-PC28 at RNA and protein levels. Altogether, ACRJ-PC28 is more aligned with a prostate cancer NE phenotype and using PCA, the clustering of ACRJ-PC28 cells to prostate cancer versus normal prostate supports this claim. In addition, five prostate cancer-associated genes (EGF, PIKR, ATM, BRCA, and BRCA) with apparent damaging mutations, one of which is a rare mutation in ATM, a possible therapeutic target, further support this claim. Also of note is that numerous variants were detected in ACRJ-PC28 that are predicted to be deleterious.
Still, the cell line is atypical in expressing both basal (CK5 and p63) and luminal (CK18) markers and lacking the expression of chromogranin A, a typical NE biomarker, but expressing SYP and ENO2. The presence of ACRJ-PC28 further highlights the value of this cell line as currently, there is need for more research demystifying the transitional processes in the development of prostate cancer. Also, given that MAA are more likely to develop prostate cancer and die from the disease, having a transitional prostate cell line presents a valuable tool in understanding the dynamics of prostate cancer development and underlying factors that increases such risks.
Given that ACRJ-PC28 was from the TRNB of a patient diagnosed with late-stage prostate cancer adenocarcinoma, the initial hypothesis was that ACRJ-PC28 would be naturally immortal. However, prior to immortalization, these cells senesced at passage 6. Furthermore, none of the earlier 27 prostate cancer explants survived beyond passage 7. This indicated that these cells were susceptible to culture stress and/or replicative senescence under in vitro culture conditions. One possibility could be that our culture conditions induced cellular stress that predictably led to increased expression of p16/p14ARF. Alternatively, or additionally, our in vitro environment did not support expression of telomerase. To minimize alterations that could select if we only eliminated senescence pathways or ectopically expressed hTERT, we hypothesized that a two-hit immortalization through hTERT upregulation and CDKNA ablation as described in Zhao and colleagues (20) would be most effective for rapid immortalization of cells that otherwise senesce. In this regard, our prior attempts at immortalization through p53 (dominant negative R175H mutant)-pcw107-V5 and hTERT upregulation, which does not target p16, proved ineffective and as such were not attempted on ACRJ-PC28.
Immunoblotting results expectedly indicated that ACRJ-PC28 lacked expression from the CDKNA locus, as cells were negative for both p16 and p14 ARF , which was consistent with the WGS results that indicated the possible heterozygous deletion at region chr9:21970801-22006400, when using CNVnator. The protein expression of both p53 and pRB tumor suppressors appeared normal, which also seems to correlate with WGS data that identified no exonic mutations, but intronic variants with no association to protein function. Recent research (68) highlighted the challenges in treating AR active adenocarcinomas that were associated with a combined loss of p53 and pRB. The tumors were found to be nonresponsive to AR antagonists. On the contrary, ACRJ-PC28 presents the reverse profile, the absence of AR and the presence of p53 and pRB.
The Cannabis sativa extract THC_EX3 that was three times more potent in reducing the viability of ACRJ-PC28 compared with PC-3 not only confirms the need for diverse cell lines but also can be a novel lead to be further explored against Erβ. Although preliminary, this further reiterates the suggestion that MAA can potentially derive greater benefits from treatment options for prostate cancer if more research accommodates the use of cell lines derived from MAA or through greater recruitment of MAA for clinical trials (69). Also highlighted was the IC 50 value (12 μg/mL) of THC_EX3, which is significantly lower than widely used prostate cancer drug docetaxel (IC 50 of 483 μg/mL) in PC-3 cells (70). Further purification of THC_EX3 could potentially provide a new targeted compound for treating prostate cancer with possibility for greater treatment derived benefits for MAA with the disease.