Producing Human Amniotic Epithelial Cells-only Membrane for Transplantation

Human amniotic epithelial cells (hAECs), as pluripotent stem cells, have characteristics of immune privilege and great clinical potential. Here, we produced hAECs membrane consisting of single-layer hAECs and basal membrane (BM) of human amniotic membrane (hAM). In conventional methods, hAECs were isolated from hAM by repeated trypsin digestion. In this study, collagenase I and cell scraper were used to remove human amniotic mesenchymal stem cells (hAMSCs) from hAM and hAECs-only membranes were produced. These hAECs on the membranes were evaluated by surface biomarkers including epithelial cell adhesion molecule (EpCAM), stage-specific embryonic antigen 4 (SSEA4) and endoglin (CD105), transcriptional level of biomarkers including POU class 5 homeobox 1 (OCT4), sex determining region Y-box 2 (SOX2), fibroblast growth factor 4 (FGF4), immunofluorescence of cytokeratin-8 (CK-8), alpha smooth muscle actin (α-SMA) and collagen type I alpha 1 chain (ColA1). Finally, the hAECs membrane were transplanted on skin-removed mice to evaluate its effect on wound healing. In comparison to the hAECs isolated by the conventional methods, the cells isolated by this proposed method had higher purity of hAECs, expressed higher in pluripotency related genes, and maintained an epithelium construction in a long-term culture. In mice application, the hAECs membrane effectively improved the skin wound healing. An efficient method was successfully established to produce hAECs membrane in this work which not only held promise to obtain hAECs in higher purity and quality, but also showed practical clinical potential.

have showed that hAECs have therapeutic potential on multiple sclerosis, immunomodulation, 59 inflammation suppression, angiogenesis promotion, oxidative stress inhibition, neurogenesis 60 induction, matrix metalloproteinases regulation, and remyelination stimulation [12][13][14]. Thus, 61 hAECs can be applied to cellular transplantation for clinical use. However, there are still some 62 obstacles in clinical application of hAECs. 63 One of the obstacles is that the purity of isolated hAECs still need improvement. In 64 conventional methods, hAECs were isolated from hAM through multiple treatment of trypsin 65 [15][16][17][18][19][20][21]. hAM are consist of three layers: single-layer of hAECs, a thick layer of BM which 66 mainly made up of collagen and hyaluronic acid, and a thick layer of hAMSCs. Trypsin 67 simultaneously affects the hAECs and hAMSCs. In most methods, the cumulative period of 68 trypsin digestion was over 1 hour. Thus, many hAMSCs could be isolated together with hAECs 69 which decreased the purity of hAECs. To address the issue, some researchers used hAECs and 70 hAMSCs together in therapy [12,22]. However, hAECs and hAMSCs had different transplantation. Although many researches had proved that hAMSCs could be capable of 76 restricting the organ fibrosis and inhibiting the epithelial-to-mesenchymal transition, hAMSCs, 77 as mesenchymal cells, could cause organ fibrosis by themselves in a long term [26][27][28]. Third,78 there are a lot of uncertain factors of the safety and reliability of hAMSCs in clinical application.

79
For instance, in clinical trials, hAMSCs have risk of tumorigenicity [29,30], causing adverse 80 responses like pain, swelling or heat [31][32][33], and even pneumonia-related death [34]. hAECs 81 could also have some potential safety risks. Probably, it will increase much difficulty of safety 82 risk evaluation in clinical trials when the test cells were mixed with hAECs and hAMSCs. Thus, 83 it is significant to establish a novel approach to provide us with highly pure hAECs.

84
Another obstacle is that the hAECs should be made into epithelial sheet in epithelium 85 transplantation therapy. To address the problem, some studies used several scaffolds for 86 producing hAECs sheet including electrospun poly (lactide-co-glycolide), poly (ε-87 caprolactone), poly (lactic acid) scaffolds [35], and silk fibroin scaffold [36]. These techniques 88 still faced some barriers to produce hAECs sheets of epithelium construction for reliable clinical 89 application, including the cost of production, the immunogenicity of scaffold material, the cell 90 adhesion capability of transplantation, the reduction degree of epithelium construction and etc.

91
Thus, hAECs still need some other proper scaffolds for further therapeutic application.

92
To address the problems mentioned above, we produced hAECs membrane of epithelium 6 / 30 93 construction by collagenase and cell scraper instead of trypsin. In previous studies, collagenase 94 had been repeatedly tested, compared with trypsin, and turned out to be inefficient in isolation 95 of hAECs [18][19][20][21]. One of the main reasons could be that collagenase affected both the BM and 96 hAMSCs. Thus, a long term digestion by collagenase broked the BM into patches and released 97 much collagen lysate which made it hard to separate hAECs and hAMSCs. In order to 98 efficiently produce a sheet of highly pure hAECs membrane in this work, we first digested the 99 hAM less than 30 min, kept the integrity of BM, and then scraped off the loose hAMSCs 100 affected by collagenase I. Thus, a sheet of BM covered by hAECs (hAECs membrane) was 101 produced. The hAECs membrane generated by this developed novel approach was proven to 102 have a high purity and pluripotency of hAECs with an epithelium construction.

105
Producing hAECs membrane 106 The Human placenta samples were aseptically processed in biosafety cabinet of class 10,000 107 clean room area. The hAM samples were isolated and cut into pieces of 5 cm, and weighed, 108 performed as the established procedures [20]. The isolation methods of hAECs were tested 109 including using tubes or plates, trypsin or collagenase I, different digestion period, and 110 application of cell scrapers (Supplementary Table S1, S2, S3).

111
To produce hAECs membrane, some hAM samples were spread out in 9 cm plates with 112 their hAECs side facing the bottom. The hAM samples were covered by 0.2% collagenase I 113 reagent (Sigma, USA), followed by placing plates in rocking device (30-60 RPM) at 37℃ for 7 / 30 114 5 min as pre-digestion. Then the samples were transferred into a new plate with fresh 115 collagenase I reagent and incubated on rocking device (30-60 RPM) for less than 30 min (The 116 exact processing period varied according to the actual digestion condition and digestion was 117 terminated when hAMSCs began to peel off and the BM was undamaged). Collagenase I 118 reagent was discarded and samples were washed with HBSS (Gibco, USA) for 2 times. Then 119 cell scrapers were used on the hAMSCs side of the samples to remove the residual hAMSCs.    Total RNA from test groups was isolated using Invitrogen TM TRIzol TM (Thermo, USA), and  Table S6) 138 Immunofluorescence (IF) and Immunocytochemistry (ICC) 139 The cell samples being fixed with 4.0% paraformaldehyde solution (10-30min) were perforated 140 on membrane by Triton X100 (0.1%, for less than 10min), and washed with phosphate buffer 141 saline (PBS) for three times (10 min per wash). Later they were blocked with 5% bovine serum

187
The isolated cells by conventional methods could contain more hAMSCs and have lower purity 188 of hAECs. Results of hematoxylin-eosin (HE) staining showed both of the hAECs side and 189 hAMSCs side had obvious tissue defect (Fig. 1F). It proved that in general methods, the isolated 190 cells were consist of both hAMSCs and hAECs. Here, the hAM were digested by collagenase 191 I and scraped on the hAMSCs side. Results showed that most of the hAMSCs were removed 192 from the hAM and the hAECs were unharmed (Fig. 1F). This tissue consisted of BM and single-193 layer of hAECs was identified as hAECs membrane (Table 1). 194 To demonstrate that the hAECs membrane comprised pure hAECs, the isolated cells from that EpCAM can be used as the main specific biomarker of hAECs. According to the EpCAM 208 biomarker, the purity of hAECs on hAECs membrane (92.2%) was proven to be higher than 209 those isolated by conventional methods (86.1%). Furthermore, hAMSCs had relatively low 210 immune-positive result of SSEA4 (56.9%). The cells from hAECs membrane (91.2%) showed 211 higher SSEA4 + than those isolated by conventional methods (78.5%). Thus, it also implied that 212 the cells on hAECs membrane had higher purity of hAECs.

217
Therefore, it showed that the cells on hAECs membrane generated here were very likely to be 218 highly pure hAECs.

219
Regarding the viability of the cells isolated from hAECs membrane, the viability of the 220 isolated cells by general methods was 71.4% (recovered from first digestion) and 61.8% 221 (recovered from second digestion) ( Table 1). In comparison, the viability of the cells isolated 222 from hAECs membrane produced here was 79.1%. Thus, the cells on hAECs membrane had 223 higher viability.

224
The hAECs membrane was produced through removing the hAMSCs by combination of 225 collagenase I and cell scraper. Results indicate that the cells from hAECs membrane show 226 higher purity of hAECs, better pluripotency and higher viability compare to those isolated by 227 methods previously described. Conclusively, it was an alternative efficient approach of 228 isolating highly pure hAECs.

229
The hAECs membrane exhibited epithelial construction and performed better in 230 characteristic maintenance 231 In general methods, hAECs were isolated into single cells and cultured. In this study, the hAECs 232 were cultured as one-layer on BM which had potential to mimic an in vivo micro-environment 233 for these hAECs.

234
To identify the difference between the hAECs cultured as single cells and those on hAECs 235 membrane, two test groups were compared in specific biomarker, pluripotency related 13 / 30 236 biomarkers and genes. In group (hAECs membrane), a hAECs membrane was spread out on a 237 9 cm plate with BM side facing the bottom and cultured in SCM. In group (hAECs separated), 238 the single cells isolated from hAECs membranes by trypsin were cultured in SCM (2 × 10 4 239 cells/cm 2 ). After 15 d, the cells of cultured hAECs membrane orderly arranged with similar 240 cellular size and clear edge (Fig. 2B). On the contrary, the separated hAECs grew into patches 241 in different cellular size, irregular edge shape, and excessively attached to each other ( Fig. 2A). 242 Thus, the results indicated that the hAECs membrane had a more similar construction of 243 epithelium than those hAECs cultured as single cells.  (Fig. 2C, 2D). These results indicated that the cells on hAECs 249 membrane had better maintenance of their own main characteristic biomarkers. Human immune 250 system mainly recognizes the allogeneic cells through MHC II surface antigen. The immune-251 positive result of HLA-DR, as a MHC II antigen, was quite low in hAECs membrane cultured 252 for 15 d (Fig. 2C). It indicated that the cells on hAECs membrane had low immunogenicity and 253 maintained to be applicable for tissue transplantation after a long-term culture (Fig. 2C). The  (Fig. 2E, 2F). These results showed that the hAECs membrane had prolong 258 the maintenance of the pluripotency of the hAECs.  Thus, the hAECs membrane could be potential medical treatment for skin and endothelium 269 wound, psoriasis, lupus erythematosus, diabetes-caused ulceration, etc. In this study, the hAECs 270 membrane was used to heal mice skin wound.

271
Nine black mice were surgically excised of skin surface in thickness of 2 mm, and six of 272 them were transplanted of hAECs membrane. After 5 d, the hAECs membrane closely attached 273 on the wound area and prevented the dermis from exposure to the air (Fig. 3B). At 10 d, the 274 skin wound of control group was heavily scabby (Fig. 3D). In transplanted mice, no sign of 275 scab, scar formation, or exposed dermis were observed (Fig. 3C). In the tissue slider of mice 276 skin wound area, the transplanted hAECs membrane had conjugated to the dermis at 5 d ( Fig.   15 / 30 277 3F). At 10 d, there were corneum-like construction formed around hAECs (Fig. 3G). The BM 278 was degraded and fell off. In comparison, the surface of skin wound of mice in control group 279 were covered by thick scab, and there was no apparent regeneration of dermis, epithelium and 280 corneum (Fig. 3H). Thus, results showed that the skin wound performed better regeneration by 281 the transplantation of hAECs membrane.

285
However, the characteristic of these hAECs from different methods were quite different. One 286 of the main reason was these methods could have caused uneven efficiency and purity of hAECs.

287
Thus, we proposed a novel alternative approach to isolate highly pure hAECs by removing 288 hAMSCs from hAM with collagenase I and cell scraper. There were several key points 289 influencing the production of hAECs membrane.

290
First, based on the tested methods, it was important to use plates instead of tubes to remove 291 the hAMSCs. The cells isolated by collagenase I were more in number and lower in EpCAM + 292 result on plates (Method F, Step 1, 2.32% of EpCAM + in 3.21±0.09×10 6 cells/g hAM) than 293 those in tubes (Method B, Step 1, 4.52% of EpCAM + in 4.67± 0.53× 10 5 cells/g hAM) 294 (Supplementary Table S1, S3). Thus, the hAMSCs were more efficiently and selectively 295 removed when the digestion was performed on the plates. One of the main reasons was that the 296 hAM samples had an uneven digestion in tubes because of discrepant exposure of collagenase.  Table S3). These results indicated the hAMSCs were supposed to face up to 304 be efficiently removed. Probably, the hAMSCs side faced up, were more likely to be exposed 305 to fresh collagenase I and avoided the collagen lysate derived from BM and hAMSCs.

306
Third, the repeated treatment with combination of collagenase digestion and cell scraping 307 improved the efficiency and success rate in producing hAECs membrane. In digestion, collagen 308 lysate released, filled the gap between cells, and constructed the digestion process. Thus, cell 309 scraper was necessary to remove the loosened hAMSCs and collagen lysate from the surface. Finally, the residue of red blood cells should be washed off as much as possible. In culture, 320 it was observed that the red blood cells obstructed the attachment and migration of the hAECs 321 or hAMSCs. It was a critical point to clean off the blood at the time of sample collection.

322
Characterization of the hAECs membrane 323 In this study, a hAECs membrane was produced by collagenase I and cell scraper. Based on the 324 results, the cells from the hAECs membrane showed higher viability, purity, and better 325 pluripotency than the hAECs isolated by the general methods.

326
In conventional methods, the hAM samples were repeatedly treated by trypsin and the 327 whole digestion period reached 75 min. Based on the results, the viability of the isolated cells 328 treated by trypsin was 71.4% for 45 min, was for 61.8% for 75 min ( Table 1). In comparison, 329 the hAECs isolated by collagenase I in 35 min and trypsin for 15 min was 79.1%. Thus, the 330 isolated hAECs had much higher viability by the methods established in this study. There were 331 two main reasons: 1. Collagenase I had little influence on the hAECs, but worked through 332 digesting the collagen on the surface structure of hAMSCs [38,46,47]. Thus, the application of 333 Collagenase I did not cause much damage to the hAECs; 2. The digestion period in producing 334 hAECs membrane was much shorter than the conventional methods. In conventional methods, 335 the hAM was treated by trypsin for 15 min, 30 min, and then another 30 min. The hAECs were 336 considered to suffer a certain damage on their cell membrane or even disrupted. In the approach 337 to produce hAECs membrane, the samples merely suffered from 0.2% collagenase I for 30 min.
18 / 30 338 Yet, these hAECs were still unharmed. The viability of the isolated cells did not just indicate 339 the living cells` portion, but also used to identify the global state of these cells. In the isolated 340 cells with low viability, these living cells could have suffered different levels of damage which 341 caused negative results in proliferation, survival, pluripotency and differentiation capability. To 342 accord with this view, results showed that the hAECs from hAECs membrane had higher 343 pluripotency marker SSEA4 + result (91.2%) than those by general methods (78.5%).

365
In this study, it was found that hAECs can maintain in a good condition when cultured as 366 hAECs membrane. The immune-positive results of some specific markers EpCAM, CD29, 367 CD44, and CD56, pluripotency markers SSEA3, SSEA4, Tra1-60, and Tra1-80 had maintained 368 in a high level (Fig. 2C, 2D, 2E). In addition, the transcription level of pluripotency related 369 genes OCT4, SOX2, FGF4, REX1, CFC1, NANOG, DPPA3, PROM1, and PAX6 was much 370 higher in the cells of hAECs membrane (Fig. 2F). Accordingly, the hAECs membrane was 371 speculated to provide a better micro-environment for the maintainance of hAECs. There could 372 be three main reasons. First, the hAECs membrane immobilized the hAECs and kept them in 373 an epithelium construction, exactly as how they grow in vivo. In general methods, hAECs were 374 cultured as single cells. Based on the results, the hAECs separated as single cells disorderly 375 grew into single-layer ( Fig. 2A). In the meanwhile, the cells on hAECs membrane still had an 376 epithelium construction (Fig. 2B). Thus, these hAECs could have a proper surface contact on 377 the hAECs membrane.  Thus, the hAECs membrane was speculated to provide good micro-environment for   Table 1 The Methods of producing hAECs membrane and isolating hAECs 2.32% Step 2 The samples were scraped on the hAMSCs side. Washed twice with PBS and retrieve all the solution for cell collection.
85.6% 1.97±0.39 ×10 6 6.54% Step 3 The tissue left was hAECs membrane. hAECs membrane was digested by trypsin for 15 min and the cells were collected for characterization. 86.1% Step 2 The samples were digested in 0.2% Trypsin/EDTA in tubes for 30 min.