Many mouse cell lines have undergone spontaneous immortalization or transformation in tissue culture [35–39], but in general, cells from most of the higher species do not undergo spontaneous immortalization or transformation . The problem is further compounded in RPE research because the cells are post-mitotic in-vivo and therefore must first re-enter the cell cycle. Also, removal of primary RPE cells from Bruch's membrane causes a majority of cells to die in culture. Apart from the commonly utilized ARPE-19, there are few accounts of spontaneously immortalized RPE cell lines such as D407, RPE-J and BPEI-1 [41, 42]. Additionally, it is common practice for primary cells to be transformed by viruses, such as SV40 or HPV, or other genetic manipulation such as activating a telomerase gene [28–30, 43, 44]
While our cell line, tRPE, still shows many of the in-vivo characteristics, including expression of RPE-specific genes, further experiments would have to be performed to determine whether the spontaneous immortalization that occurred had changed the expression of other genes.
Primary RPE cell isolation can often become contaminated with choroidal melanocytes or fibroblasts. The melanocytes are derived from neural crest embryonic tissue whereas RPE are optic cup derived [1, 45]. These melanocytes can often be confused with the pigmented RPE, but melanocytes are difficult to culture in normal unsupplemented media. TPA, bFGF, phorbol esters, or concurrent growth with keratinocytes are required to grow the melanocytes [46, 47].
The tRPE cell line had similar morphological characteristics as the passaged primary cells. Although the tRPE cells were never observed to have the hexagonal morphology as seen in-vivo cells, they were polygonal. Growth of RPE cells on BCE-ECM and Matrigel has been established and used by many investigators [24, 48–51]. Other researchers have shown that primary RPE and other RPE cell lines grow better on BCE-ECM than Matrigel [49, 50]. Because ECM is integral in the differentiation of epithelial cells, both types of ECM were used to determine the optimal matrix material.
Morphologically, tRPE cells and primary cells grown on BCE-ECM were epithelioid. When Matrigel was used, the cells were elongated and disorganized, and did not seem to inhibit the proliferation of RPE cells (data not shown). Based upon morphological characteristics, BCE-ECM seemed to be a better substrate for tRPE cells.
It is important to have the cells melanized in order to regain an in-vivo like state. There are many publications that state melanin is a protective agent to light mediated damage and also acts as a free radical quencher . Remelanization has been attempted by feeding the cells α-MSH and thymidine dimmers, and irradiation with UV (data not shown). Boulton and Marshall , and Denton et al.  have shown that directly feeding purified melanosomes to the cells causes the cells to endocytose exogenous melanin. In our research, we have found that immortalized RPE cells will take up melanosomes granules when fed in culture, but a control cell line, HeLa, does not. The exogenously melanated tRPE cells were similar in morphology to primary cells as seen in Figure 1. Although speculative, the melanin granules may cause the tRPE cells to re-differentiate, but this needs further study.
Addition of growth factors was attempted as a mechanism to increase the differentiation of RPE cells. Although serum-free media has been established for RPE cultures, the cost and time for preparing it prohibited its use [51, 54]. Some of the ingredients were supplemented into the DMEM media as a cost-effective approach. The supplementation of the media along with growth on BCE-ECM was a requirement to fully differentiate the tRPE cells in order to determine if they produced RPE-specific proteins.
The tRPE cell line was characterized to establish that they are indeed RPE origin. Four genes were used as identifying markers performed by RT-PCR. RPE65 and CRALBP are RPE-specific genes [17, 55–61]. RPE65 expression is rapidly lost in cells that have been cultured without any external signals [17, 31, 62], and growth of RPE on BCE-ECM has been shown to produce an increase in CRALBP expression . We have seen that the tRPE cells will express both RPE65 and CRALBP when grown on BCE-ECM with supplemented media. This indicates that the tRPE cells may have the ability to re-differentiate when grown under these conditions. As expected, the RPE65 expression was lost in the tRPE cells cultured on tissue grade plastic, and CRALBP expression was diminished.
Vimentin functions as an intracellular scaffold, and it is expressed in cultured RPE cells and primary cells from ARMD patients [19, 63]. We show that both tRPE and primary RPE cells express the gene. Finally, β-actin is a housekeeping gene and was used as a measure to quantitate the relative levels of expression. We believe that the RT-PCR data support that the clones derived from the primary explant are of RPE origin.
To further expand the evidence that the cells are of RPE origin and retain the in-vivo characteristics, immunocytochemistry was performed. Reverse transcriptase PCR was already performed, but it does not show that the cell will produce protein. There are post-transcriptional regulations that can inhibit the translation of the protein. Multiple groups have confirmed that cultured RPE cells will express ZO-1, RPE65, and CRALBP under specific conditions [8, 24, 64, 65]. Expression was determined by immunocytochemistry to show that the cells do produce proteins specific to RPE cells and some that are characteristic of monolayer epithelial cells.
Because Nicoletti et al. have shown that the RPE65 mRNA has AU rich elements on the 3' UTR that will allow down-regulation of the RPE65 protein, , the presence of the protein needed to be established. Only after the cells have redifferentiated when grown on an extracellular matrix, is the expression seen again [17, 31]. Ma et al. have shown that a Sf9 cell line transfected with a plasmid-containing recombinant human RPE65 will have a distribution throughout the cell when detected with an anti-RPE65 antibody . Figure 3 shows that the tRPE cells produce RPE65. The green fluorescence is seen throughout the cells, but has a mottled appearance and higher intensity toward the edges of the cells, which may be indicative of protein on the membrane and in the cytosol.
Vimentin is usually used as a marker for the determination of cell types of mesenchymal origin [66–68], while cytokeratins are found in epithelial tissue almost exclusively . Both of these proteins are frequently co-expressed with in RPE cultures. Rapidly dividing cells show cytokeratin diffusely, while quiescent cells will show a strongly staining perinuclear cage-like structure surrounded by profuse finger-like projections extending out from the nucleus . Immunocytochemistry showed that most of the cells were positive for vimentin and cytokeratin, but not all. The tRPE cells were reactive against antibodies to both proteins in a typical pattern of the cytoskelatal network found in RPE cultured cells, which further indicates that the cells are of RPE origin.
Long culture conditions are needed to form adherins junctions in RPE cells [64, 65], which forms an adhesion belt around each cell . Zonula Occludens-1 (ZO-1), also known as tight junction protein (TJP1), is utilized in adherins junctions for cell polarization, transport and cell-to-cell signaling [71–73]. ZO-1 immunofluorescence shows that the tRPE cells form a nearly hexagonal structure on the periphery of the cells, as previously seen in Figure 1a. This indicates that the tRPE cells form adherins junctions and are a polarized epithelial monolayer, as expected of differentiated RPE.
Bestrophin is heavily detected in cells of epithelial and endothelial origin because of the requirement of these cells in transport of nutrients through the cell [74, 75]. Marmorstein et al. have shown bestrophin to localize in the basolateral side of RPE cells to control the transport of chloride and water between the RPE and the choriocapillaris . The observed expression of this protein indicates the tRPE cells have a polarized morphology and are a monolayer of epithelial cells.
Initially, the tRPE cells were thought to be spontaneously immortalized because they had no foci formation after the cells became confluent. Generally, as cells increase in passage number, they have the ability to become transformed due to mutations accumulating in their DNA. Specific tests were done to establish that the tRPE cells were not transformed even at population doubling 102. Additionally, the experiments also show that the cells still retained specific characteristics of a differentiated state.
The tRPE cells are spontaneously immortalized after being passaged up to 150 population doublings. Growth beyond 100 population doublings is considered immortal . The cells did not have the ability to grow in the lack of serum or under low serum conditions and did not form foci (data not shown), which are two characteristics of untransformed cells.
Bovine cells have a normal diploid chromosome number equal to 60 . All of the autosomes are telocentric except the X and Y chromosomes, which are metacentric . Transformed cells have been shown to be aneuploid or heteroploid . Most of the chromosomes that were counted were within the 55–60 range, with many at 60 chromosomes.
Growth studies of cells in suspension or soft agar were performed to determine whether the cells were anchorage-independent and exhibited the loss of contact inhibition . The tRPE cell numbers dramatically decreased in suspension at the beginning of the experiment, indicating they require a substrate to attach. The tRPE cells were unable to grow in soft agar, while the control CHO AA8 cells were able (data not shown).
Pulse labeling was performed to confirm that the cultures stopped or slowed down proliferation after confluence, a process called density-dependent inhibition of cell growth . The number of dividing cells was never greater than 1.5%, with most times being less than 1%. Comparatively, actively dividing tRPE cells were expected to have a rate of labeling around 25%, and transformed cells could show up to 50% of cells labeled with 3H-thymidine. The tRPE cells did not exhibit significant cell division at confluency. Also, the tRPE cells remained as a confluent culture for one year and still were viable, whereas primary cells are difficult to keep alive for any extended period of time.