Pilot Study on Cytothesis for rabbit ocular surface defects using peripheral blood autologous mononuclear cells

At present, there is no effective treatment for severe ocular surface injury due to limbal stem cell deficiency. Stem cell transplantation offers hope for disease treatment, and the peripheral blood mononuclear cell fraction is a rich source of stem cells. In this study, rabbit autologous peripheral blood mononuclear cells were collected and combined with human amniotic membrane for repair of experimental ocular surface deficits. We observed good reparative effects, suggesting that autologous peripheral blood mononuclear cells are a promising stem cell source for the treatment of severe ocular surface injury.


Introduction
Many such studies have revealed that human (h)PBMCs are rich in pluripotent progenitor cells that can be differentiated into almost any triploblastic cell [1]. Compared to other adult pluripotent stem cells (e.g., from bone marrow and other tissues), PBSCs are easily obtained and easily separated [2]. Moreover, unlike embryonic stem cell (ESC)based therapy, autologous transplantation of PBMCs does not require long-term use of immunosuppressive agents and involves no ethical limitations [3,4], while unlike induced pluripotent stem cells (iPSs), PBMCs carry little tumorigenic risk.
PBMCs have many possible applications for the treatment of eye diseases. For instance, Liu, et al. [5], Zhang, et al. [6], Peng, et al. [7], and Xian, et al. [8] transplanted PBMCs into the eyes of retinitis pigmentosa (RP) mice and observed migration of transplanted cells to all layers of the retina, indicating possible use for restoring retinal function. Limbal stem cells (LSCs) propagate and differentiate into corneal epithelial cells, allowing continuous renewal and dynamic stability of the corneal epithelium [9][10][11][12]. Further, the limbus isolates the cornea from the conjunctiva to prevent intrusion of conjunctival cells and neovascularization towards the cornea, thereby maintaining corneal transparency [13][14][15][16][17]. Limbal stem cell deficiency impairs this barrier function, resulting in conjunctival cell intrusion, neovascularization, and visual impairment [18].
Adult stem cells isolated from PBMCs [1,6] are a promising alternative warranting further study. It is intended to provide a new ideology and approach for the treatment of severe ocular surface disorders caused by human limbal stem cell deficiency.
TUNEL staining (apoptosis): Some frozen sections (5-μm thick) were fixed in 4% paraformaldehyde and the cornea ruptured using 0.03% Triton-X for TUNEL staining with a FITC-conjugated TUNEL antibody kit (Beyotime, Jiangsu, China) according to the manufacturer's instructions. The nucleus counterstained with 0.5 g/mL Hoechst33342 (Sigma). Observation, photography, and image analysis were conducted using a laser scanning confocal microscope (Carl Zeiss Meditec, Germany).

Statistical
analysis of corneal transparency and neovascularization: Corneal transparency and neovascularization were evaluated by Score Criteria for Ocular Surface Evaluation (Supplementary Table 1) [17,31]. Score results were compared among groups by ANOVA using SPSS Statistics Software (IBM). Quantit/ative results are expressed as mean ± standard deviation (Supplementary  Tables 3 and 4) and change in mean over time ( Figures 5 and 6).

Morphology of rabbit PBMCs co-cultured with HAMs
HAMs were scraped of the epithelial layer and spread onto 24well plates to form a basement membrane, which was clearly observed under light microscopy ( Figure 1A). Rabbit PBMCs obtained from each animal of the experiment group (rPBMC Group, average (± SD) of 5.328×10 7 ± 1.071×10 7 cells; Supplementary Table 4) were small, round, and phase-bright in dispersed suspended culture ( Fig. 1-B). Cell numbers increased substantially through overnight co-culture with HAMs ( Figure 1C).
Sourcing, processing, and preservation of human amniotic membranes: All HAMs were sourced from the discarded placentas of healthy women (with no HIV or CMV). Epithelial cells were scraped off and the remnants washed with sterile saline (or PBS), suspended in DMEM-F12, and maintained in a 37 °C.
Separation and tags of rabbit PBMCs: Autologous venous blood samples (20 mL) were extracted from the ear margins of experimental Group (rPBMC) rabbits. Mononuclear cells in blood were separated with rabbit PBMC separation medium (Hanyang Biology, Tianjin, China). CM-DiI-labeled and suspended in 1 ml of stem cell culture medium.
Graft Preparation: HAMs were cut into the grafts of about 1.5 × 1.5 cm and laid with the scraped epithelium face-up in 24-well plates. Autologous CM-DiI-labeled rPBMC suspensions were seeded on these HAM grafts and the cells separated from 1 rabbit were transplanted onto a HAM graft and cultured overnight at 37 °C in an incubator with 5% CO 2 and 100% humidity. The excess culture medium was removed and rinsed three times with PBS to remove non-adherent cells before transplantation the next day.

Modeling of rabbit ocular surface deficiency and Transplantation:
Rabbit's corneal epithelium was scraped off using microsurgery forceps. Then, the tissues inside and outside of the corneal limbus with the width of 1−2 mm and the thickness of 100-150 μm were cut to build LSCD model. The prepared graft was placed over the corneal surface deficiency with the scraped epithelium/seeded side facing the lesion. In the negative control group (LSCD Group), the lesion was left untreated, while the AM group received unseeded HAM transplantation.
Observation of ocular surface signs: After transplantation, the rabbit operative eye was observed under a slit lamp once a week. Conjunctival congestion, secretion, corneal transparency, and neovascularization were observed and evaluated, the integrity of the corneal epithelium assessed using sodium fluorescein, and the eye photographed.
Confocal laser microscopy of living corneas: Changes in corneal epithelium were assessed under anesthesia in living cornea by confocal laser microscope (Heidelberg HRT3-RCM, Germany) on postoperative Weeks 1, 2, 3, and 4.
Histology: After completion of in situ examinations on Week 4 (Sections 2.7−2.8), rabbits were sacrificed and corneas isolated for hematoxylin and eosin (HE) staining, then observed and photographed under an optical microscope (Axio Imager A1 and AxioCAMMRc5; Carl Zeiss, Jena, Germany) for image processing and analysis (AxioVision; Carl Zeiss).

Immunocytochemistry and immunohistochemistry:
Fresh rabbit PBMCs and some frozen corneal sections were stained for K3/K12, β-1 integrin, and CD45 protein to monitor changes in cell phenotype. The concentration of separated rPBMCs was diluted to 2−3 × 10 6 /mL and cell smears prepared from 100-μL samples at 800 rpm with moderate acceleration for 2 min using a centrifugal smear machine (Thermo Shandon CYTOSPIN USA). Frozen corneal sections were fixed with 4% paraformaldehyde for 10 min and stained with mouse anti-rabbit K3/K12 (ab68260 abcam, UK), mouse anti-rabbit β-1 integrin (ab78502 abcam, UK), and (or) mouse anti-rabbit CD45 (AM02304PU-S, Acris Antibodies Inc USA). Immunolabeling was visualized with FITC-conjugated goat anti-mouse IgG (ab6785 abcam USA) and the nucleus was counterstained with 0.5 g/mL Hoechst33342 (Sigma, St. Louis, MI, USA). Observation, photography, and image analysis were conducted using a laser scanning confocal microscope (Carl Zeiss Meditec, Germany).

Changes of ocular signs
The ocular surface of untreated rabbits showed no conjunctival congestion or secretion, high corneal transparency, and no neovascularization or fluorescein staining (indicating intact corneal epithelial layers) (Figure 2A and 2C). On the day of surgery, a few rabbit subjected to corneal surface lesion modeling exhibited conjunctival congestion but without secretions. Compared to normal eyes, lesioned eyes were significantly edematous, with reduced corneal transparency and extensive fluorescein staining indicative of a breached epithelium, but without neovascularization ( Figure 2B and 2D Table 3), corneal transparency scores were obtained for each rabbit (Supplementary Table 5) and compared among groups by ANOVA ( Figure 5A and 5C, Supplementary Table  6, Supplementary Figure 1). At all four postoperative observation times (1W, 2W, 3W, and 4W), corneal transparency differed significantly among the rPBMC, AM, LSCD Groups (P<0.05). The rPBMC Group had a significantly lower evaluation score at all four observation times than either the AM or LSCD Group, as well as the highest corneal transparency, while the AM Group had higher evaluation score and corneal transparency at all 4 observation times than the LSCD Group. Moreover, transparency scores (higher numbers indicating greater    Table 3), corneal neovascularization score for each rabbit (Supplementary Table 7) was obtained and compared among groups by ANOVA ( Figure 5B and 5D, Supplementary Table 8, Supplementary Figure 2). At all four postoperative observation times (1W, 2W, 3W, and 4W), corneal neovascularization differed significantly among groups (P<0.05), with significantly lower levels in the rPBMC Group than AM and LCD Groups (P<0.05). In turn, AM Group scores were lower than LSCD Group scores. Further, like transparency, neovascularization continued to increase with time in the LSCD group.

Corneal changes under confocal microscopy
Under confocal microscopy, the cell bodies of normal corneal epithelium were highly reflective, with centrally located and less reflective nuclei ( Figure 6M-6P). In the rPBMC Group at postoperative 1W, corneal epithelial cells were clearly reflective but smaller than normal corneal epithelial cells ( Figure 6I). At postoperative 2W, there was no increase in cell number, and cells exhibited clearly reflective nuclei and unclearly reflective somata ( Figure 6J). At postoperative 3W and 4W, the number of cells was significantly increased and more densely arranged, with smaller somata and clearly reflective nuclei ( Figure 6K and 6L). In the AM Group at postoperative 1W, no cells with high brightness and clear boundaries were observed in the field of view, but a few dark dot-like substances were observed ( Figure 6E). At postoperative 2W, small cells with unclear nuclei were observed ( Figure  6F). At postoperative 3W, the number of cells was not significantly increased and cells were still of smaller size with unclear nuclei ( Figure  6G). Cell number still had not dramatically changed by postoperative 4W (Figure 7H). In the LSCD Group at postoperative 1W, numerous irregular bright reflective puncta without obvious cell morphology were observed under the microscope ( Figure 6A). At postoperative 2W, the number of these reflections was reduced and no round cells could be observed under the microscope ( Figure 6B). At postoperative 3W and 4W, a large amount of disorderly arranged fibrous tissues which were observed, with occasionally bright reflections. Again, no morphologically clear cells could be observed under the microscope ( Figure 6C and 6D).

Histological changes
In normal cornea, 4−6 layers of epithelial cells were densely arrayed [ Figure 7A (a)]. Similarly, limbal cells were also densely arrayed in multiple layers [ Figure 7A (b)]. Immediately after surgery, the corneal epithelium and limbus were largely devoid of cells [ Figure 7A (c) and 7A (d)].
At postoperative 4W, the rPBMC Group exhibited a growing corneal epithelium with 3−4 layers of epithelial cells in a dense array [ Figure 7B (a)]. In contrast, the corneal epithelium of the AM Group grew poorly and remained disordered at W4. In addition, a few vacuolar cells could be observed and numerous inflammatory cells had infiltrated the basement layer of AM Group corneas [ Figure 7B (b)]. In the LSCD Group, the epithelial layer also grew poorly. Epithelial cell layers were disordered, numerous vacuolar cells were observed, and a large number of inflammatory cells had infiltrated the basement layer [ Figure 7B

Dynamic changes of CM-DiI labeled rabbit PBMCs after transplantation
In addition to the changes in corneal epithelial cell marker expression and apoptosis, changes in rabbit autologous PBMC numbers were assess by CM-Dil staining. At postoperative 1W, numerous CM-DiI-labeled cells were observed in the rPBMC group. These numbers decreased with time post-surgery, and CM-DiI-labeled cells were no longer detectable at postoperative 4W ( Supplementary Figures 4-6).

Discussion
Recent studies affirm that ESCs, iPSCs, and a variety of adult stem cells can be transdifferentiated into various triploblastic cells [32][33][34][35][36][37][38][39][40][41]. The application of stem cells for the prevention and treatment of human diseases has been also studied extensively, including for myocardial infarction, pancreatic disease, and liver disease [42][43][44][45]. Compared to other stem cells, PBMCs have several advantages, most importantly the potential for abundant and non-invasive acquisition from the The model group receiving rabbit PBMC plus HAM transplantation exhibited significantly higher corneal transparency, lower neovascularization, more complete recovery of the corneal epithelium, and lower apoptotic rate than groups receiving no transplantation or HAMs alone. These results strongly suggest that autologous PBMCs can facilitate the recovery of severe ocular surface disorders caused by limbal stem cell deficiency. One possible mechanism is inhibition of inflammation, which could reduce progressive ocular damage and inhibit neovascularization by secreted angiogenic factors. In addition, stem/progenitor cells in PBMCs may have transdifferentiated into corneal epithelial cells within this favorable corneal microenvironment, thereby promoting recovery [26][27][28][29]30].
Confocal laser microscopy can specifically detect laser light reflection from the focal plane of living animal cornea, while reflected or scattered light outside the focal plane is blocked or filtered, yielding high-definition corneal images. Such dynamic in situ observation revealed that epithelial cells appeared earlier and expanded faster on corneal surfaces treated with rPBMCs.
Rabbit PBMCs do not express K3/K12 or β-1 integrin (supplementary Figure 3). Rather, K3/K12 is a specific marker for corneal epithelial cells, while β-1 integrin is expressed by epithelial stem/progenitor cells [38]. At 1W after cell transplantation, very few of K3/K12﹢ and β-1 integrin﹢ fluorescent cells were observed on the corneal surface of operative eyes in the rPBMC Group, but fluorescent immunopositive cell numbers increased with time such that the corneal surface was extensively covered by K3/K12﹢ andβ-1 integrin﹢ cells by W4. In contrast, few such cells were detected in AM and LSCD Groups, suggests more rapid and complete recovery of corneal epithelium in the rPBMC Group.
In contrast to this increase in K3/K12﹢ and β-1 integrin﹢ cells, the numbers of CM-DiI-labeled autologous PBMCs decreased with time, and no CM-DiI﹢ cells were detected by W4. One possible explanation for this disappearance is dilution of dye due to proliferation of the stem/progenitor cell fraction coupled with elimination of labeled terminal cells such as lymphocytes through apoptosis and removal by ocular tears. Due to this reduction in fluorescence, CM-DiI may not permit long-term tracing of PBMCs [46][47], so the use of alternative markers is required to assess PBMC fate in this model.

Ethics approval and consent to participate
All animal experiments were approved by the Laboratory Animal Ethics Committee of Zhongshan Ophthalmic Center, Sun Yat-Sen University (Approval Number: 2013-060).

Consent for publication
All authors have approved the submission for publication.

Availability of data and materials
All data generated or analysed for this study are included in this published article.

Funding
This study was funded by the Science and Technology Projects of Guangdong Province, China (2017A030303013 and 2013B020400003), the Provincial.