Study approval, design, and subjects

The study encompassed a prospective interventional randomized design, forming a consecutive series of non-masked cases. This collaborative effort involved the Research, Development, and Innovation Department of Vissum Instituto Oftalmologico de Alicante (Grupo Miranza), Miguel Hernandez University (Spain), Optica General (Lebanon), Laser Vision Center (Lebanon), and REVIVA Research and Application Center at Middle East Hospital (Lebanon). The Institutional Review Board (IRB) Ethical Committee of REVIVA granted prospective approval for the study, which adhered to the principles of the Declaration of Helsinki. Informed written consent was obtained from all participating patients for the procedures outlined. The study was officially registered with ClinicalTrials.gov (Code: NCT02932852).

A total of fourteen patients were enrolled and subsequently allocated randomly to three groups: Group 1 (G-1) consisted of patients who received implantation of ADASCs (n = 5); Group 2 (G-2) received human dCL, each 120 µm thick (n = 5); Group 3 (G-3) received human ADASCs-rCL (n = 4). The distribution of laminas in G-2 and G-3 occurred randomly after the decellularization procedure.

Follow-up appointments occurred first at 1 week post-operation, and then at 1, 3, 6, and 12 months post-operation. Thirteen patients successfully completed the full one-year clinical follow-up. Only one patient from G-1 discontinued participation after the first month post-operation, citing personal reasons unrelated to the study. The results of interim analyses conducted at 6 and 12 months have already been published [7,8,9,10,11].

Patient selection criteria

The study’s inclusion criteria encompassed patients with advanced keratoconus, defined as stage ≥ IV based on the RETICS keratoconus classification [24], who were already designated as candidates for corneal transplantation due to disease severity and associated comorbidities. Participants needed to be at least 18 years of age and exhibit negative serology for human immunodeficiency virus (HIV), hepatitis B (HBV), and hepatitis C (HCV). Conversely, patients with corrected distance visual acuity (CDVA) < 0.1 in the contralateral eye, a history of previous corneal hydrops or central corneal scars, active concomitant inflammatory eye disease, other sight-threatening ocular comorbidities, prior corneal surgical interventions, including collagen cross-linking, pregnancy or breastfeeding status, or a history of systemic malignancy, were excluded from the study.

Isolation, characterization, and culture of autologous ADASCs

Standard liposuction was conducted on each patient, yielding approximately 250 ml of fat mixed with local anesthesia. The adipose tissue was processed in accordance with previously published protocols [2, 4, 25,26,27], where autologous ADASCs were isolated, cultured, and characterized by flow cytometry analysis as per the recommendations of the International Federation of Adipose Therapeutics (IFATS). Cell quiescence was induced by reducing the serum concentration to 0.5% for 60 to 80 h prior to transplant, while the absence of apoptosis and aneuploidy was ensured through propidium iodide labeling (Invitrogen, USA) and cell cycle flow cytometry analysis. In G-1, a total of 3 × 106 cells were suspended in 1 ml of phosphate-buffered saline (PBS) and subsequently transplanted intrastromal. In G-3, ADASCs were cultured on decellularized corneal lamina (dCL) of the stroma for 24 h [7, 9] with a seeding density of 0.5 × 106 cells per 1 ml of PBS (12 h on each surface of the lamina).

Laminas

Human donor corneal stroma characterized by negative viral serology and suitability for human complete corneal transplant were provided by the “Banco de Ojos para el tratamiento de la Ceguera, Centro de Oftalmología Barraquer” in Barcelona (Spain), according to the regulatory directives 2004/23/EC and 206/17/EC established in Spain.

The donor corneas were cut in laminas using a femtosecond laser with a diameter of 9.0 mm and a thickness of 120 µm. The corneal laminas (CL) were subsequently decellularized as before [4]. The efficacy of the decellularization procedure was verified using three distinct approaches: Biochemical digestion in proteinase K, DNA extraction and quantification carried out using a Picogreen Assay kit, and histological nuclear DAPI staining and fluorescence microscopy, alongside hematoxylin and eosin staining [4, 28].

Surgical procedureAutologous ADASCs implantation

The technique for implanting mesenchymal stem cells has been outlined in preceding publications [7]. Briefly, a 60-kHz IntraLase iFS femtosecond laser (AMO Inc, Irvine, CA) was employed to create a recipient corneal lamellar dissection using a single-pass approach. This generated an intrastromal laminar pocket with a diameter of 9.5 mm, situated at a medium depth corresponding to the thinnest preoperative pachymetry point as measured by anterior segment optical coherence tomography (AS-OCT) (Carl Zeiss, Germany). The anterior side-cut incision was executed at 30°, spanning a 3 mm arc length incision. Parameters mirroring those used in LASIK surgery were employed in the femtosecond laser process. Following this, the intrastromal pocket was opened via blunt dissection, employing a Morlet lamellar dissector (Duckworth & Kent, England). Subsequently, 3 million ADASCs suspended in 1ml PBS were injected into the pocket using a 25-G cannula. Before the cell injection, a 1 mm corneal paracentesis was applied to reduce intraocular pressure and allow a greater volume to be injected into the stromal pocket, while no more than 10% to 30% of the injected volume was estimated to remain within the corneal stroma. Notably, no patients underwent corneal suturing as part of this procedure [7, 9]. A topical antibiotic and steroids (Tobradex; Alcon) were applied at the end of the surgery and were applied every 6 h for 1 week, followed by a descending dose of topical dexamethasone 0.1% (Maxidex, Alcon) for 3 more weeks [7].

Lenticule implantation

The same 60-kHz IntraLase iFS femtosecond laser in single-pass mode was utilized in this aspect. A 50° anterior cut assisted with corneal dissection, and the arc length incision measured 4 mm. The intrastromal pocket was opened using blunt dissection by a Morlet lamellar dissector (Duckworth & Kent, England). The lamina was inserted, centered, and gently unfolded through careful tapping and massaging from the epithelial surface of the host cornea. In instances where ADASCs-rCL was utilized (G-3), the pocket was irrigated with a solution containing an additional 1 million autologous ADASCs suspended in 1 ml of PBS, administered via a 25-G cannula. Similar to G-1, a temporal limbal paracentesis was performed just before implantation to reduce the intraocular pressure. Post-insertion, the incision was closed using an interrupted 10/0 nylon suture, which was subsequently removed one week after the surgery. Patients were treated during surgery and post-operatively like ADASCs implantation patients alone [8, 9, 11].

In vivo confocal microscopy (IVCM) protocol for corneal assessment

IVCM was conducted using a HRT3 RCM (Heidelberg) with a Rostock Cornea Module. This confocal microscope employed a coherent diode laser as its light source, characterized by a wavelength of 670 nm and a minimum resolution of 1024 × 768 at 16-bit [29]. The utilization of coherent light facilitated enhanced image contrast and quality, in particular for visualizing the corneal stroma and its cellular components [30].

For the confocal microscopy assessment, a drop of topical anesthetic (Oxybuprocaine 0.4%) was administered to the eye under evaluation where the confocal microscope was adjusted to + 12 D, and a high-viscosity gel (2.0 mg/g of carbomer) was applied to the front surface of the microscope lens of the Rostock Cornea Module (RCM). A Tomocap was then positioned atop the RCM objective. Patients were guided to maintain a stable position, aligning their gaze with the confocal microscope’s lens and focusing on the light within. The initial focal position was reset to “0” at the superficial epithelial cells of the examined eye. Subsequently, the RCM was rotated either clockwise or anticlockwise within a range of 0 ± 50 μm, and the focal plane was meticulously adjusted to the desired cell layer. To ensure comprehensive visualization, a minimum of four images were captured at intervals of 50 μm in depth, as previously published [29]. These images, integral to the analysis in this study, were obtained at varying stromal depths. However, it’s worth noting that all images were derived from the central diameter region, with a limit of ≤ 9 mm, and situated below the measurement of the intrastromal diameter pocket [7,8,9,10, 31].

ImageJ-based cell characterization and classification

The ImageJ analysis program, developed by the National Institutes of Health, was used. Distinctive morphological features guided the selection of infiltrated cells, setting them apart from host keratocytes characterized by their fusiform shape and specific nuclei traits within different stromal regions [31, 32] Furthermore, infiltrated cells were differentiated from ADASCs that, within the initial 6 months post-surgery, exhibited a unique round shape and distinctive luminosity and refringence compared to regular keratocytes [10, 31].

The meticulous task of cell identification was executed by a panel of four authors (MEZ, MPDM, NM, J.L.A), the inter-concurrence for the first expert was 97%, for the second was 99%, for the third was 92%, and for the fourth was 98%. The intra- concurrence between the four experts was 90%. When all experts agreed on the reproducibility of the ICs identification, each one of the identified ICs underwent independent area measurements by two experts (MEZ and KAJ), being the inter-concurrence 98%, and 96.5% respectively. The intra-concurrence was 92% between the two experts.

Within the ImageJ system, the following sequence was employed: Analysis < Tools < Roi Manager (Fig. 1). When calculated area differences exceeded 25% between the two experts, the analysis was reiterated. In categorizing infiltrated cells based on area, the study classified cells as lymphocytes if their area ranged from 50 to 80 µm2, granulocytes if within 110 to 220 µm2, and monocytes/macrophages for areas exceeding 300 µm2. Cells with structures smaller than 50 µm2 were designated as cell remnants. In cases falling beyond these size brackets, cell classification was determined by structural attributes and luminance, nuclei, and morphological aspects.

Fig. 1

Measurement of the area of the ICs using the Image J software. Open image pressing the icon File (blue arrow), then in the tool bar choose the irregular shape (green arrow), then manually delineate the nuclear periphery of the cell with some squares (red arrow). From the Tool bar, choose Analysis < Tools < Roi Manager and Press Add < Measure, the area measurement is obtained in µm2 (yellow arrow)

Statistical analysis and significance evaluation

The statistical evaluations were conducted using Excel 2019, SPSS, and GraphPad Prism 9. Results were presented as mean ± SEM or percentage and assessed for statistical significance using Student’s T-test and F-test, where the latter gauged the variance difference between the two samples. Significance was attributed when both t-test and F-test P-values were < 0.05 within the same comparison. Differences among multiple groups were ascertained through a one-way analysis of variance (ANOVA). In all statistical tests, α and P values were two-tailed, with the level of significance set at 0.05.