Postnatal periodontal ligament as a novel adult stem cell source for regenerative corneal cell therapy

Abstract Corneal opacities are a leading cause of global blindness. They are conventionally treated by the transplantation of donor corneal tissue, which is, restricted by a worldwide donor material shortage and allograft rejection. Autologous adult stem cells with a potential to differentiate into corneal stromal keratocytes (CSKs) could offer a suitable choice of cells for regenerative cell therapy. Postnatal periodontal ligament (PDL) contains a population of adult stem cells, which has a similar embryological origin as CSK, that is cranial neural crest. We harvested PDL cells from young adult teeth extracted because of non‐functional or orthodontic reason and differentiated them towards CSK phenotype using a two‐step protocol with spheroid formation followed by growth factor and cytokine induction in a stromal environment (human amnion stroma and porcine corneal stroma). Our results showed that the PDL‐differentiated CSK‐like cells expressed CSK markers (CD34, ALDH3A1, keratocan, lumican, CHST6, B3GNT7 and Col8A2) and had minimal expression of genes related to fibrosis and other lineages (vasculogenesis, adipogenesis, myogenesis, epitheliogenesis, neurogenesis and hematogenesis). Introduction of PDL spheroids into the stroma of porcine corneas resulted in extensive migration of cells inside the host stroma after 14‐day organ culture. Their quiescent nature and uniform cell distribution resembled to that of mature CSKs inside the native stroma. Our results demonstrated the potential translation of PDL cells for regenerative corneal cell therapy for corneal opacities.

mimecan) that regulate collagen fibril alignment and interfibrillar spacing, which are crucial for stromal matrix organization, corneal mechanics and transparency. 2 Developmentally, CSKs are derived from the ocular mesenchyme of cranial neural crest (NC) origin. 3,4 In adult corneas, they remain quiescent and extend dendritic cell processes to neighbouring keratocytes forming a highly organized syncytium. Besides collagens and KSPGs, they also express stromal crystallins, such as aldehyde dehydrogenases (ALDH, type 1A1 and 3A1), a-enolase, lactic dehydrogenase and transketolase, which match the refractive indices between inter-and extra-cellular regions, contributing to transparency. 5 Trauma to the corneal stroma or infectious keratitis will activate CSKs to transform into stromal fibroblasts (SFs) with a loss of keratocyte features. SFs are proliferative and acquire a different set of tissue healing-related genes compare to CSK's, for example fibronectin, a5-integrin and/or a-smooth muscle actin (aSMA) (when SFs transit to myofibroblasts during scar formation). 6 SFs mediate extracellular matrix (ECM) contraction and disrupt the organized lamellar alignment, causing diminished transparency.
"Corneal opacities" are a leading cause of worldwide blindness and are estimated to affect over 10 millions of people (information obtained from WHO global blindness data and WHO 2002 subregional causes). 7,8 Conventionally, advanced stages of opacities are treated by keratoplasty using donor corneas. 9 However, the use of cadaveric donor tissue is limited with a worldwide shortage, and there is a risk of allograft rejection in the longer term (up to 38% 10-year graft failure rate). 10,11 Hence, new strategies should be developed to replace the defective CSKs and to eliminate or prevent stromal scarring. 12 Autologous cell therapy involving an adult stem cell source is an attractive option.
Many adult stem cell types are known to transdifferentiate into cell types other than the tissue they reside. Unlike induced pluripotent stem cells, the conversion of adult stem cells can be direct, efficient and bypass the pluripotent cell state, which often elicits safety concerns and risk of tumorigenesis. 13 Many studies have reported transdifferentiation via the forced expression of transcription factors or the provision of an appropriate niche and trophic factors. [14][15][16][17] Although the exact mechanism is undetermined, the presence of suitable environment and enhancement of functional activity of specific niche (with signal mediators) may suffice to drive the differentiation of adult stem cells towards the desired phenotypes.
In recent years, dental stem cells have received much attention in cell regeneration research because of its easy accessibility, plasticity and applicability in regenerative medicine. 18,19 Among the known dental stem cells, 20,21 periodontal ligament-derived stem cells (PDLSCs) and dental pulp stem cells (DPSCs) originate from the cranial NC and share similar developmental pathways as CSKs. 22,23 Isolation of PDL tissue by scraping from tooth root surface is relatively easier than that for dental pulp cells, which requires cracking the tooth and this could induce primary cell death. Besides DPSCs, the loose dental pulp tissue contains nerve and blood vessels. Tooth extraction induces lesions to the pulp nerve and endings, which cause the release of inflammatory cytokines, such as IL-1 and TNFa, and this would impact on DPSCs' physiological status, such as survival and multipotency. 24,25 A recent study has illustrated that DPSCs could differentiate into CSK-like cells expressing keratocan (KERA), a unique marker for CSKs, and KSPGs. 26  on ice to the culture facility within 6 hours after extraction. PDL cell isolation protocol was followed as previously reported. 35 Briefly, after extensive PBS washes to remove blood traces, PDL tissue was scraped mechanically along the middle one-third of root surface, finely cut and digested with 100 lg/mL collagenase I <8 hours at 37°C. Primary culture of activated keratocytes was established using low serum ERI protocol as reported earlier. 37 At P4, activated keratocytes were kept in serum-free ERI culture to obtain CSKs and in 5% serum culture to obtain SFs, respectively, for 7 days. After characterization by specific marker expression, these cultures were used as control cell types. 38 Similarly, primary CSK cultures from adult monkey (n = 4) and rabbit (n = 5) were established using the same ERI protocol and P4 cells were used in the experiments.

| Quantitative PCR
Samples in RLT buffer (Qiagen, Valencia, CA, USA) or Trizol (Sigma) were processed for total RNA extraction using RNeasy kit (Qiagen) and on-column RNase-free DNase kit (Qiagen) according to manufacturer's protocol. Total RNA (1 lg) was reverse transcribed using Superscript III RT-PCR kit (Invitrogen) with random hexanucleotide primer (10 ng/mL, Invitrogen). Gene expression was assayed with specific primer pairs (Table S3)

| Statistical analysis
Mann-Whitney U-test was used to compare the gene expression levels between treatment and control groups. Results were described as mean and SD. All statistical calculation was performed using SPSS 20.0 (SPSS, Chicago, IL, USA). P < .05 was considered statistically significant.

| RESULTS
We collected teeth from 41 Chinese individuals and successfully established 21 primary PDL cultures (success rate: 51.2%). The failed cultures were because of contamination issues or the absence of viable cells. Among those established cultures, we randomly selected five primary cultures for the CSK induction study and all of these donors were non-smokers and had good to excellent oral hygiene status reported by dental surgeon (GBT) ( Figure 1A). Each primary PDL culture was derived from PDL tissues of a single donor and we did not pool cells from different individuals. This would demonstrate the unique responsiveness of PDL cells towards CSK differentiation irrespective of individual variability.

| PDL cell characterization
The primary cells were adherent and displayed clonal growth forming loose colonies with migratory cells at the periphery, whereas cells at central region displayed more intercellular connections via dendritic cell processes ( Figure 1B) Figure 1C). Cells at P3 were characterized for specific gene expression representing NC-derived cells using qPCR. Figure 1D  Comparing between supplements, culture added with 5% CEE generated more intact spheroids than that supplemented with 20 ng/mL EGF, a protocol reported to expand palatal NCSCs. 39 The centrally located cells of CEE-generated PDL spheroids expressed nestin, Sox2 and Sox10, suggesting the NC identity ( Figure 2). However, Sox2 and Sox10 were negligibly expressed in EGF-generated spheroids as well as in single PDL cells.
When the spheroids were cultured under CSK induction with bFGF (20 ng/mL), TGFb3 (1 ng/mL) and L A 2 P (1 mmol/L) on low attachment surface for 5-7 days, they expressed CD34, LUM, ALDH3A1, Col8A2 and low levels of KERA and its synthesizing enzymes (B3GNT7, CHST6) as shown by immunostaining and qPCR analyses ( Figure 3A,B). LUM up-regulation was not clear under qPCR, and this could be because of the high basal expression levels. Without induction, these CSK-associated genes were barely expressed ( Figure 3A top panel). In particular, CD34 immunoreactivity was detected in part of the treated spheroids, similar to that observed in primary human CSK ( Figure 3A bottom panel). We found almost half of human primary CSK population expressed CD34. Flow cytometry also revealed that 42.5% CSKs were CD34 + ( Figure 3C). Concomitantly, all human CSKs were positive to LUM, KERA and ALDH3A1 ( Figure 3A). However, human SFs did not express any CSK genes and only 0.5% cells were CD34 positive as revealed by flow cytometry ( Figure 3A,C).
When spheroids were dissociated into single cells using collagenase, followed by induced differentiation to CSK lineage on collagen I-coated surface, the cells attached and displayed dentritic morphology with thin processes extending and interacting with neighbouring cells ( Figure 4A). They had mildly up-regulated expression of ALDH3A1 and KERA ( Figure 4B). When cells were differentiated on AM stromal side, stronger ALDH3A1 and KERA expression was noted while the cells maintained highly dendritic morphology, which was not observed for PDL cells differentiated on AM stroma without spheroid formation ( Figure 4C). The percentage of KERA-expressing cells in five random regions was gradually increased, reaching 42.4 AE 6.15% for spheroid-dissociated cells on AM stroma, compared to 9.2 AE 0.9% for cells on collagen 1-coated surface, at day 15 ( Figure 4E). CSK induction was significantly improved when intact spheroids were directly seeded on AM stroma ( Figure 4D can-rich ECM via integrin-mediated attachment. 44 We studied two NCSC enrichment protocols for PDL cells and found that the supplementation with CEE resulted in spheroids expressing nestin, Sox2 and Sox10 while the protocol added with 20 ng/mL EGF derived spheroids expressing nestin only. The latter protocol was reported to expand palatal NCSCs 39 ; however, it seems not applicable for PDL cells, even though both are originated from cranial NC developmentally. 45 Hence, further optimization will achieve better protocols to obtain a refined population of NC-derived stem cells from oral and craniofacial regions. Having enriched NCSCs, these free-floating spheroids attained CSK phenotype after further treatment with bFGF, TGFb3 and L A 2 P, a reported CSK differentiation protocol. 46 The spheroid cells positively expressed CD34, ALDH3A1 and Col8A2, markers for CSKs, 47,48 and they also had low expression F I G U R E 3 Characterization of periodontal ligament (PDL) spheroids under corneal stromal keratocyte (CSK) differentiation. Spheroids were generated by protocol having 5% chick embryo extract and treated with bFGF, TGFb3 and L A 2 P on low attachment surface for 7 d. A, By confocal microscopy, the expression of CD34, Lum, KERA and ALDH3A1 was visualized in treated spheroids, and this was similarly observed in primary human CSKs but not in stromal fibroblasts (SFs). B, qPCR analysis showing the up-regulated fold changes of CSK genes (LUM, ALDH3A1, Col8A2, B3GNT7, CHST6) in treated spheroids compared to control PDL cells. C, Flow cytometric dot plots and histograms showing the event profiling of CD34 expression in human primary CSK (42.5% CD34 + cells) and SF (0.5% CD34 + cells) levels of KERA and its synthesizing enzymes (B3GNT7, CHST6). In particular, CD34 immunoreactivity was detected in almost half of cell population inside the treated spheroids. This staining pattern was similar to that of primary human CSK culture and flow cytometry analysis further revealed CD34 expression in 42.5% CSKs (Figure 3).
CD34 is a transmembrane phosphoglycoprotein with L-selectin (CD62L) as its most commonly described ligand. 49  Several cell types have been reported for potential use in regenerative corneal therapy. 59 The recent success in propagating primary human activated keratocytes and the reversion to quiescenct CSKs has "in principal" resolved the issue of CSK shortage which has been a challenging topic in corneal cell therapy. 37  Future identification of unique markers will help in a more efficient isolation of pure CSSCs and a better control of differentiation. 65,66 The report of in vitro differentiating human ESC into CSK via NC progenitors has supplied valuable information for protocol design 46 ; however, a number of limitations, such as induction efficiency, control of differentiated cell purity and the risk of tumorigenesis, remain to be solved before its translational application. Adult bone marrowand adipose-derived MSCs has been shown to differentiate towards CSK phenotype when intrastromally injected into mouse corneas, [67][68][69] but it is difficult to control the final cell fate. Dental stem cells have been an attractive option in cell regeneration research. The harvest of PDL cells and DPSCs from extracted teeth is a significant advantage when compared to bone marrow, umbilical cord and adiposederived MSCs, which require more invasive procedures. Dental SCs are also present in all teeth, of which the third molar is a common source, because of its large tooth volume and surface area. It is the last tooth to develop and is often impacted and remains buried in an "unused," but healthy state. Third molar extraction is one of the most commonly performed procedures in oral surgery. Impacted third molars are removed to prevent risks of caries, periodontal disease, pericoronitis, odontogenic cysts and dental crowding. Extracted teeth are merely disposed as medical waste. 70 It is thus under minimal ethical concerns to utilize these cells and future translational application with donor matching or autogenic usage can be supported through "dental cell banking" concept. 19 Our study highlighted the plasticity and differential capacity of PDL cells to generate CSKs. These cells are readily propagated ex vivo, hence offering an ample amount of viable cells for CSK differentiation for regenerative corneal cell therapy and stromal tissue engineering applications in the management of corneal opacities and corneal biomechanical pathologies.