Stem cells are clonogenic cells with self-renewal and differentiation properties, which may represent a major target for genetic damage leading to prostate cancer and benign prostatic hyperplasia. Stem cells remain poorly characterised because of the absence of specific molecular markers that permit us to distinguish them from their progeny, the transit amplifying cells, which have a more restricted proliferative potential. Human CD133 antigen, also known as AC133, was recently identified as a haematopoietic stem cell marker. Here we show that a small population (approximately 1%) of human prostate basal cells express the cell surface marker CD133 and are restricted to the α2β1hi population, previously shown to be a marker of stem cells in prostate epithelia (Collins, A. T., Habib, F. K., Maitland, N. J. and Neal, D. E. (2001). J. Cell Sci. 114, 3865-3872). α2β1hi/CD133+ cells exhibit two important attributes of epithelial stem cells: they possess a high in vitro proliferative potential and can reconstitute prostatic-like acini in immunocompromised male nude mice.
Considerable research efforts have been directed towards the identification of markers associated with the initiation and progression of prostate cancer, yet there is little consensus about the target cell within prostate epithelium that is susceptible to malignant transformation. Stem cells may represent a major target for mutations leading to cancer as their longevity assures continued presence during the long latency between exposure and cancer development (Pierce and Wallace, 1971; Reya et al., 2001).
The existence of stem cells in the prostate is probably best illustrated by animal studies investigating the effect of androgen on the prostate. Castration leads to rapid involution of the gland, but once androgen levels are restored, the gland completely regenerates. As this cycle of involution-regeneration can be repeated many times, a population of long-lived, prostatic epithelial stem cells must exist (Isaacs et al., 1987). In human prostate, stem cells can be distinguished from transit amplifying cells (daughter cells that have a more limited proliferative capacity), by two- to threefold higher surface levels of integrin α2β1 (Collins et al., 2001). In addition, a third population of basal cells that are destined to differentiate after a few rounds of cell division can also be identified (Collins et al., 2001; Hudson et al., 2001). α2β1 integrin expression can be used to enrich for stem cells directly from prostate tissue using differential adhesion to type I collagen. Depending upon the methodology used, between 1-15% of the basal cells exhibit high levels of α2β1 integrin expression (Collins et al., 2001). Clearly, the availability of further cell surface markers for prostate basal cells would greatly facilitate the isolation and characterisation of human prostate epithelial stem cells (hPE-SC). In the haematopoietic system, multilineage stem and progenitor cells are isolated using monoclonal antibodies against cell surface markers for enriching rare subpopulations that are clonal self-renewing and multipotent (Morrison and Weissman, 1994; Spangrude et al., 1988). This strategy was used to identify and isolate a candidate hPE-SC. The monoclonal antibody CD133 recognises the CD133/1 antigen: a five-transmembrane glycoprotein, localised to membrane protrusions or microvilli. Antibodies to CD133/1 have been used to enrich for human haematopoietic stem cells (Yin et al., 1997), endothelial cells (Peichev et al., 2000), neurons and glial cells (Uchida et al., 2000). CD133 is also expressed by the intestine-derived epithelial cell line Caco-2 where it is down-regulated upon differentiation (Corbeil et al., 2000). CD133 is believed to be the human orthologue of mouse Prominin, a protein expressed on the apical surface of neuroepithelial cells as well as several other embryonic epithelia, and on brush border membranes of adult kidney proximal tubules (Weigmann et al., 1997).
Here we show that CD133+ selected cells significantly expand in culture and form acini with evidence of prostatic-specific differentiation when grafted subcutaneously into the flanks of male, athymic nude mice: properties consistent with a stem cell origin.
Materials and Methods
Tissue collection and isolation of hPE-SC
Human prostatic tissue was obtained, with patient consent, from 38 patients (age range 53 to 99) undergoing transurethral and retropubic prostatectomy for benign prostatic hyperplasia and cytoprostatectomy for bladder cancer. The condition of benign prostatic hyperplasia was confirmed by histological examination of representative fragments.
The tissue was prepared as described previously (Collins et al., 2001). Briefly, collagenase digestion released epithelial structures (organoids; glands and ducts) which where subsequently separated from the stromal fraction by repeated unit gravity centrifugation. Organoids were dis-aggregated into a single cell suspension by incubation with trypsin/EDTA (InVitrogen, Paisley, Scotland) for 30 minutes, at 37°C. Luminal cells were depleted from the sample by linking anti-CD57 antibody to magnetic beads (Dynal Biotech, UK).
Basal cells were further fractionated on the basis of adhesion to type I collagen, as described previously (Collins et al., 2001). Cells that had adhered after 5 minutes constituted the stem cell-enriched fraction (approximately 3% of the basal population). The non-adherent (transit amplifying and post-mitotic basal cells) were also recovered during washing (Collins et al., 2001). Anti-human epithelial antigen, linked to MACS microbeads (Miltenyi Biotec Ltd, Surrey, UK) was further used to purify the non-adherent population. CD133 cells were selected from the above fractions using MACS microbeads linked to anti-human CD133, according to the manufacturer's instruction. CD133-positive (CD133+) cells were not detected within the CD133-negative (CD133–) fraction and the purity of the MACS system was calculated as 98%.
Colony forming and long-term serial culture assays
Selected basal cells were counted and plated on collagen-coated (BD-Biocoat™) plates (BD Biosciences, Oxford, UK) for the determination of colony forming efficiency (CFE) and long-term proliferative potential. After plating, CFE was determined by marking single cells and examining at intervals up to 28 days when they were subsequently fixed and colonies scored under a phase contrast microscope. Colonies were scored if they contained >32 cells. As the number of cells selected was small, irradiated (60 Grays) STO (mouse embryonic fibroblasts) cells were added as feeders. For the long-term serial culture assays, used to determine the proliferative potential of various basal subpopulations, cells were passaged continuously until their growth capacity had been exhausted. At every passage, the number of cells generated by each fraction was determined by cell counts. The cumulative total cell output from the initial 5000 cells was calculated at the end of the experiment, assuming that all the cells from each passage had been replated. The results obtained were subjected to a paired t-test to determine statistical significance. The duration of each experiment was dependent on the individual culture, but was generally between 55-75 days.
Immunofluorescent staining of tissue sections
Prostate cells expressing CD133 antigen were identified by direct immunofluorescent staining using anti-CD133 mouse monoclonal antibody (clone 293C3, Miltenyi Biotec Ltd) directly conjugated with phycoerythrin (PE). A series of 1 μm optical sections through the entire thickness of the tissue was obtained using a 60× objective of the confocal microscope, and a Z series was constructed from these sections.
Confocal microscopy and FACS analysis of Isolated basal cells
Cells were processed for triple colour staining for confocal microscopy by fixing in ice-cold methanol (20 minutes) and permeabilising in 0.4% Triton X-100 (Sigma) containing 0.3% normal goat serum (NGS; Sigma) for 10 minutes. After blocking [1:5 dilution of NGS in Tris-buffered saline (TBS) for 10 minutes] primary antibody (diluted in blocking serum) or a non-specific isotype control was added for 1 hour. Cells were washed three times in TBS and incubated with goat anti-mouse Ig conjugated to tetramethylisothiocyanate (TRITC) for 45 minutes. After blocking free Ig-binding sites with normal mouse serum (1:5 dilution for 30 minutes), the second primary antibody, directly labelled with PE or fluorescein (FITC), was applied and the cells incubated for 1 hour. After washing, cells were mounted in the anti-photobleaching medium Vectashield containing diamino-2-phenylindole DAPI (Vector Laboratories, Peterborough, England).
The antibodies used for confocal analysis of cell fractions were 34βE12, which identifies keratins 1, 5, 10 and 14 of the basal cell compartment in prostate (Dako Ltd), RCK103 (BD PharMingen) and LL002 (Serotec Ltd), which react with keratins 5 and 14 respectively, CY-90 designated K18, which reacts with keratin 18 and identifies differentiating epithelium in prostate (Sigma) (Robinson et al., 1998) and RCK 108 (designated K19) which binds to basal and luminal epithelial cells and recognises keratin 19. In all cases a Leica Microscope, running spot advanced software, was used for image capture.
Basal cells were processed for single (CD133 (CD133-1)- PE; Miltenyi Biotec Ltd) or double (CD133 (CD133-1)- allophycocyanin (APC; Miltenyi Biotec Ltd) and Ki-67 (MIB-1) – FITC; DakoCytomation) staining along with appropriate negative controls and single colour positive controls. Double staining was performed on cells fixed and permeablised in 1% saponin (Sigma). Cells were analysed using Beckton Dickinson FACScan software. At least 10,000 events where acquired for each sample.
Transplantation of CD133+ sorted cells into athymic nude mice
To determine the ability of selected CD133+ cells to induce prostatic morphogenesis, growth and functional cytodifferentiation in vivo, α2β1hi/CD133+ cells were combined with cultured stromal cells (passage 1-4) and injected subcutaneously into 6- to 8-week old male, nude mice (strain ICRF-nu) at a ratio of 1:10 epithelial (1×104 to 5×104) to stromal cells. α2β1hi/CD133– cells (with stroma) were used as a control. After 8 weeks, grafts were removed, fixed in phosphate-buffered formalin, embedded in paraffin, and cut into 4 μm thick sections. Serial sections 35 μm apart were stained with Mayer's Haematoxylin and counterstained with Eosin. The capacity of grafted cells to differentiate in vivo to the secretory phenotype was taken as evidence of an epithelial stem cell population.
Immunocytochemical analysis of transplanted CD133 cells
Paraffin embedded serial sections (4 μm) taken from xenografts were incubated with primary antibody. A biotinylated secondary antibody was applied to the specimens followed by incubation with avidin-biotin complex reagents (Dako Ltd., Bucks, UK). The staining was developed with diaminobenzidine tetrahydrochloride (DAB; Sigma). Primary antibodies used were 34βE12, K18, anti-PAP (clone PASE/4LJ: a marker of secretory luminal cells; Dako Ltd) and anti-androgen receptor (clone AR27) antibody (Novocastra Laboratories Ltd, Newcastle upon Tyne, UK).
CD133+ cells are located within the basal layer of prostate epithelium
We initially determined the location of CD133 cell within prostatic epithelium using anti-CD133 directly conjugated to PE (Fig. 1). Two hundred cross sections of acini, from five patients samples, were examined for CD133 expression. Only five percent (10/200) of sections examined for CD133 expression were found to contain positive cells; which were always located in the basal layer. Between 0 and 5 CD133+ cells were present in each cross section of an acinus and were either found alone, or when clustered together, were often situated at the base of a budding region or branching point (Fig. 1). There was no significant difference in the number of CD133 cells and the size of acini (P<0.05), i.e. CD133 cells were not confined to tips of acini, for example, but were randomly located throughout each acinus.
Candidate hPE-SC-expressing CD133 are restricted to the α2β1hi population and represent a quiescent subpopulation of the basal layer
Basal cells were selected at 5 minutes (α2β1hi) and 20+ minutes (cells that had not adhered within 20 minutes; α2β1low) on type I collagen-coated plates and were subsequently incubated with anti-CD133 and analysed by confocal microscopy. Cells were also stained for expression of 34βE12 to confirm that we had isolated basal epithelium. All cells, from both populations, stained positively for 34βE12, whilst CD133 expression was restricted to the α2β1hi population (Fig. 2A). Expression of CD133 was punctate, similar to that described by Corbeil and co-workers (Corbeil et al., 2000) in developing epithelia. Expression of CD133, in the absence of 34βE12 expression was not observed. Both populations were also analysed by flow cytometry (Fig. 2B,C). Similarly, CD133 expression was only observed within the α2β1hi population (25% of cells within the α2β1hi population expressed CD133; 0.75% of the total basal population).
In vivo cell kinetic studies have established that prostate stem cells in the mouse are largely quiescent and do not proliferate at high rates, whereas transit amplifying cells are actively cycling (Tsujimura et al., 2002). To determine whether this is true for human prostate, basal cells were labelled with Ki-67 and CD133 and analysed by flow cytometry (Fig. 2D). The results obtained from three separate experiments shown in Fig. 2D demonstrate that the majority of actively cycling basal cells (i.e. cells expressing Ki-67) reside in the CD133– fraction, whereas the CD133+ population contain significantly more quiescent cells.
Clonogenic cells can be enriched by CD133 selection from the α2β1hi population
To determine the proliferative potential of putative hPE-SC, α2β1hi basal cells were separated into two populations (CD133+ and CD133–) using MACS immunomagnetic beads, and plated to determine CFE (Fig. 3A). CD133+ cells displayed a 10.6±0.45-fold greater CFE than the non-selected basal population and CD133– cells had a 4.5±1.2-fold greater CFE than non-selected basal cells. The other important difference between the CD133+ and CD133– founded colonies was the time taken by the cells founded by the CD133+ cells to begin a phase growth in vitro, perhaps reflecting the slow cell cycle time of stem cells in vivo. For example, colonies founded by CD133+ selected cells first appeared 5±1.5 days after colonies founded by the CD133– population. 87.5% of CD133+ founded colonies contained >32 cells, whereas only 36% of the colonies founded by CD133–-selected cells were >32 cells in size.
We compared the long-term proliferative capacity of the CD133+ and CD133– populations by assaying total cell output following serial passage until all growth potential was exhausted (typically 55-75 days). The total cell output from selected cells demonstrated clearly that basal cells with the greatest, long-term proliferative capacity reside within the CD133+ population: the CD133+ population generated more than twofold more cells than the CD133– population (Fig. 3B; P<0.05).
The phenotype of α2β1hi/CD133+ cells was examined and the results are shown in Fig. 4. All CD133+ cells examined expressed the basal cell specific marker, 34βE12 (Fig. 4A). Within this population, 75±8.2% of cells expressed K14 with K5 (Fig. 4B). A minority of cells (11±2.6%) expressed K18 or K19 (Fig. 4C), but never with K14 (Fig. 4D,E). Markers of prostate-specific differentiation, such as prostatic acid phosphatase (PAP), prostate-specific antigen (PSA) and the androgen receptor were not expressed within the CD133+ population (results not shown).
CD133+ selected cells form fully differentiated acini in immunocompromised mice
One of the properties attributed to stem cells is the ability to regenerate the different cell types that constitute the tissue in which they exist (Morrison et al., 1997). We therefore determined whether the candidate hPE-SC, isolated directly from the basal layer, had the potential to self-renew and form fully differentiated glands when grafted into athymic male mice. Basal cells were selected for α2β1hi/CD133 expression and were immediately grafted, together with human stromal cells, into the flanks of mice. After 8 weeks, a fully formed epithelium was retrieved from 2/10 (20%) mice grafted with CD133 selected cells. From 4/10 (40%) mice, epithelial nests were recovered. In the remaining mice epithelium was not found, but a dense inflammatory infiltrate was observed. In mice grafted with CD133– cells, epithelium was not observed (0/10), but a fibromuscular stromal matrix was apparent in 7/10 of these grafts. In the four grafts containing epithelium, variable gland formation was observed, with morphologic and immunohistochemical evidence of both secretory and squamous cell differentiation (Fig. 5). The xenografts often consisted of multiple acini within connective tissue. Lumina were either well defined (Fig. 5A) or when less apparent, containing large, squamous-like cells with secreted material (Fig. 5D). In the well-defined acini, flattened or cuboidal cells lined the periphery and stained intensely with the basal cell-specific marker 34βE12 (Fig. 5B) and lacked expression of the differentiation marker K18 (Fig. 5C). The cells facing the lumen or adjacent to the basal layer were either morphologically squamous or columnar and stained with antibodies against K18 (Fig. 5C), PAP (Fig. 5D), a marker of functional cytodifferentiation, while lacking expression of the basal cell marker 34βE12 (Fig. 5B). As the secretory luminal cells are dependent upon androgen for survival, we expected that androgen receptor expression would also be present in these grafts; indeed this was the case (Fig. 5E). Expression was confined to the nucleus of a proportion of basal cells as well as the columnar or squamous-like epithelium. It is also worth noting that AR was also focally expressed within the nucleus of stromal cells surrounding the epithelium (Fig. 5E).
In the present study we have shown that a subpopulation of α2β1hi basal cells express the CD133 antigen and that this expression correlates with a high proliferative potential and ability to regenerate a fully differentiated prostatic epithelium with expression of prostatic secretory products in vivo.
In the past decade, haematopoietic stem and progenitor cells have been identified using monoclonal antibodies against cell surface markers to enrich for rare subpopulations that are clonally self-renewing and multipotent (Spangrude et al., 1988; Yin et al., 1997). One such antigen, CD133, has not only been used to isolate haematopoietic stem cells (Yin et al., 1997), but it also has an expression restricted to the stem cell population within the central nervous system (Uchida et al., 2000). CD133, although it has no known function, is expressed by developing epithelial cells, and is rapidly down regulated upon differentiation (Corbeil et al., 2000). In this study, we examined the position and number of CD133-expressing cells within prostatic acini. CD133 expression was restricted to a small population of cells within the basal layer, but was randomly located throughout acini. High surface expression of integrin α2β1 has been shown to be associated with stem cells in the prostate (Collins et al., 2001), as well as other tissues systems such as the epidermis (Jones et al., 1995). We have previously shown that α2β1hi cells (which make up 3% of the total basal population) are also randomly located throughout acini (Collins et al., 2001). In contrast, Tsujimura and co-workers (Tsujimura et al., 2002) found that stem cells (label-retaining cells) in the rodent prostate are concentrated in the proximal region of the prostatic ducts. However, they also detected label-retaining cells in more distal locations, albeit in lower numbers. They proposed a model whereby proximally located stem cells give rise to a population of transit-amplifying cells that migrate distally. However, this model does not take into account the nature of the cells in the intermediate regions and the label-retaining cells that are also present in the distal regions of the duct.
Unlike α2β1hi cells, CD133+ cells were often found in clusters, most frequently at the base of a budding region or branching point. This discrepancy may reflect the strict criteria that were used to identify α2β1hi cells (those cells with a fluorescence intensity across cell borders which was at least twice the average of the other cells in the transect). Using these criteria, only 1% of basal cells were regarded as α2β1hi (Collins et al., 2001). CD133+ cells were restricted to the α2β1hi population and had a greater colony-forming ability and proliferative potential in vitro than CD133– cells within the α2β1hi population. The enrichment of colony forming cells seen with CD133 was similar to that observed by Yin et al. (Yin et al., 1997) for haematopoietic stem cells. Although the total cell output was relatively low compared to rapidly renewing tissues, such as the epidermis and bone marrow, this may reflect the age of the patients in this study (average age 71 years). It is well documented that the proliferative capacity of stem cells decreases with age (Marley et al., 1999). Moreover, compared with tissues that undergo rapid cell turnover, the adult prostate is slow growing [epithelial cells within the adult gland have a doubling time of approximately 200 days (Isaacs and Coffey, 1989)], thus the proliferative capacity would be expected to be lower. CD133+ cells also founded colonies at a slower rate; possibly reflecting the slow cycling times of stem cells in vivo; resembling the type I colonies previously described by our group (Collins et al., 2001). Indeed, the majority of CD133+ basal cells did not express the proliferation-associated marker, Ki-67.
In vivo, epithelial cells exhibit ordered pairs of keratins. Thus, K5 and K14 are expressed by basal cells, whereas K8 and K18 are predominantly expressed by the luminal cells in the prostate (Sherwood et al., 1990; Sherwood et al., 1991). An intermediate population has also been identified within the basal layer that expresses basal and luminal keratins (Hudson et al., 2001; van Leenders et al., 2000) suggesting the presence of committed cells within the basal layer. All CD133+ cells directly isolated from prostate tissue reacted with antibody 34βE12, which identifies keratins 1, 5, 10 and 14. Whilst the majority of cells expressed K5 and K14, a minority also expressed K18 or K19. It is possible that K18 and K19 are not as robust indicators of differentiation as previously believed. For example K19 is a marker of skin stem cells (Michel et al., 1996). Several investigators have hypothesised that K5/14 stem cells give rise to a transit amplifying population that express K5 with K18 and K19 (van Leenders et al., 2000). As the K5/K14-expressing cells made up the largest proportion within the CD133 population our results suggest that stem cells express this phenotype. CD133, although enriching for the stem cell population may also detect an early transit amplifying population.
One important property attributed to stem cells is their ability to regenerate all of the cell types specific to the tissue from which they are derived. Therefore, transplanted stem cells should have the potential to self renew and to produce progeny that differentiate into a fully functional epithelium. By definition, only stem cells have this capacity. α2β1hi/CD133+ where able to fulfil these criteria, as they were able to produce a fully differentiated epithelium, expressing markers associated with prostate-specific differentiation when grafted into nude mice. In contrast, α2β1hi/CD133– were incapable of forming prostate epithelium in vivo.
Prostate development and maintenance of function is dependent upon androgen via interaction with the stroma (Cunha et al., 1983). Therefore, in our experiments epithelial cells were grafted together with stroma, into male mice, to induce epithelial morphogenesis and cytodifferentiation. Often within the same graft, varying degrees of organisation were observed. Some acini had a fully formed lumen and a clear distinction between the basal and lumen compartment was observed, yet in adjacent acini the distinction between these compartments was not as clear and lumen were not present. It is possible that prostatic regeneration may have been more effective in the presence of a powerful inducer of prostate development; urogenital sinus mesenchyme (UGM) (Cunha et al., 1987). Tissue recombination and grafting experiments have demonstrated that epithelial budding, branching morphogenesis, and the functional differentiation of luminal secretory cells are dependent on mesenchymal androgen receptor (Donjacour and Cunha, 1993; Cunha and Lung, 1978; Lang et al., 2001). In all cases, in our study, stroma surrounding acini expressed the androgen receptor. However the signals that are involved in lumen formation are less well understood. Epithelial cells have default apoptotic machinery and require survival signals to avoid engaging in a suicide program (Raff et al., 1993). There is now good evidence that survival signals are provided by soluble factors, including hormones and cytokines, and more importantly through cell contacts with extracellular matrix (ECM) proteins and neighbouring cells (Frisch and Francis, 1994; Pullan et al., 1996) In skin, growth and differentiation of the overlying epidermis is dependent upon synthesis of an ECM by the supporting fibroblasts (Demarchez et al., 1992). In the prostate, the differences observed in acini formation may represent the time taken for the prostatic fibroblasts to form a functional basement membrane.
It has been shown in Chinese hamster ovary three-dimensional cultures, apoptosis is in part due to the loss of integrin signalling (Zhang et al., 1995). Integrin-regulated apoptosis has also been suggested as a possible mechanism of cavitation. This is accomplished by apoptosis, within the region where a space forms, caused by a lack of integrin signalling. This has been shown to occur in embryonal carcinoma cell lines that form hollow embryoid bodies (Coucouvanis and Martin, 1995). What is evident however, is that even within acini with no clear lumen or cellular compartmentalisation, basal and luminal cells could still be detected immunohistochemically. Similarly, we have recently reported successful tissue reconstructions from prostate cell lines that display a degree of polarisation and differentiation (Lang et al., 2001).
In conclusion, we have shown that cells expressing the α2β1hi/CD133+ phenotype have properties that indicate that they are equivalent to the stem cells of the prostate, as they have the potential to establish and maintain a prostate epithelium similar to that found in vivo, with associated secretory activity. The degree of enrichment of stem cells attainable by the use of CD133+ will allow further analysis of the stem cell population, such as the pathways governing proliferation and differentiation. Most importantly, stem cell research may provide novel therapies against prostate tumours whose long-term cure rate remains stubbornly low.
This work was supported by the Harker Foundation, University of Newcastle, and Yorkshire Cancer Research.