Derivation of xeno-free and GMP-grade human embryonic stem cells--platforms for future clinical applications

Abstract

Clinically compliant human embryonic stem cells (hESCs) should be developed in adherence to ethical standards, without risk of contamination by adventitious agents. Here we developed for the first time animal-component free and good manufacturing practice (GMP)-compliant hESCs. After vendor and raw material qualification, we derived xeno-free, GMP-grade feeders from umbilical cord tissue, and utilized them within a novel, xeno-free hESC culture system. We derived and characterized three hESC lines in adherence to regulations for embryo procurement, and good tissue, manufacturing and laboratory practices. To minimize freezing and thawing, we continuously expanded the lines from initial outgrowths and samples were cryopreserved as early stocks and banks. Batch release criteria included DNA-fingerprinting and HLA-typing for identity, characterization of pluripotency-associated marker expression, proliferation, karyotyping and differentiation in-vitro and in-vivo. These hESCs may be valuable for regenerative therapy. The ethical, scientific and regulatory methodology presented here may serve for development of additional clinical-grade hESCs.

Figure 1. Characterization of human feeders.

Human feeders derived from three types of human tissues showed similar properties. They were immunoreactive with anti-vimentin (A; immunostaining, nuclei counterstained with 4', 6-diamidino-2-phenylindole (DAPI)), CD44 and human Fibroblast antigens (B; FACS analysis). They had similar doubling times (C) and normal karyotypes (D). γ-irradiated cord feeders (WCB2) were mitotically inactive as indicated by lack of KI-67 expression and BrdU incorporation (E; immunostaining (green), nuclei counterstained with DAPI (blue)). Mitomycin-C treated and mitotically-active non-treated feeders served as negative and positive controls, respectively. (B & C present analysis of three lines from each feeder type).

Figure 2. The development of a clinical-grade hESCs culture system.

Feeders derived from the three types of human tissues were equally effective in maintaining undifferentiated pluripotent growth of hESCs (HES-1[2]) for a minimum of 5 passages within research-grade KO medium. (A) The hESCs cultured on the three types of feeders formed colonies with typical morphology (phase contrast images), expressed alkaline phosphatase (AP), and were immunoreactive with anti-Oct-4 (nuclei were counterstained with DAPI). They were pluripotent as demonstrated by their potential to differentiate in-vitro into progeny representing the three germ lineages. Immunofluorescence staining showed differentiated cells expressing beta-tubulin III (ectoderm), muscle actin (mesoderm) and alpha-Feto-Protein (AFP, endoderm). (B) FACS analysis showing that the percentage of cells expressing markers of pluripotent human stem cells, and the level of background differentiation (percentage of SSEA-1 expressing cells) was similar (n = 3). (C) The hESC doubling time did not differ significantly after culture on the three types of feeders (n = 3). Cord feeders were further used for the development of the clinical grade culture system, and the phenotype of hESCs was compared after culture within research-grade KO medium system and following replacement and supplementation with xeno-free and GMP-grade reagents. The following culture compositions were compared: hESCs cultured on porcine gelatin (porc-gel) or in recombinant gelatin (rec-gel) in KO medium, SCGM and SCGM supplemented with HSA (SCGM-HSA). (D) With all culture conditions, the hESCs colonies had similar and typical morphological characteristics (phase contrast images), and the stem cells expressed alkaline phosphatase activity (red) and Oct-4 (green; blue-DAPI nuclear counterstaining; immunofluorescence images). (E) The percentage of cells expressing Tra-1-60 and Tra-1-81 was analyzed by FACS and compared between the various culture conditions (n = 4). Scale bar in (D) represents 200 um.

Figure 3. Process flow diagram of the approach taken in deriving xeno-free clinical-grade fibroblast feeders and hESCs.

The scheme indicates the ethical, scientific, and regulatory steps taken in deriving our clinical-grade fibroblast feeders and hESCs.

Figure 4. Characterization of HAD-C 100 1° cell bank.

The hESCs colonies had the typical morphology of human pluripotent stem cell colonies with clear distinguishable borders from the cord feeder cells (A; phase-contrast image). The cells expressed alkaline phosphatase (B; AP, fluorescence image). Indirect immunofluorescence staining showed that the hESCs were immunoreactive with anti-Oct-4 (C; nuclei counterstained with DAPI, D), TRA-1-60 (E), TRA-1-81 (F), SSEA-3 (G), and SSEA-4 (H). FACS analysis showed that the majority of cells expressed markers of pluripotency TRA-1-60 (I), TRA-1-81 (J) and SSEA-3 (K) (Data from a representative experiment). The hESCs had normal karyotype (46, XY; L) and could differentiate in vitro and in vivo into cells representing the three embryonic germ layers (M-R). Immunofluorescence staining showing in vitro differentiated cells expressing beta-tubulin III (ectoderm, M), muscle actin (mesoderm, N) and sox-17 (endoderm, O). Hematoxylin-eosin stained histological sections of teratoma tumors showing neural rosettes (ectoderm, P), cartilage (mesoderm, Q) and villi structures with columnar glandular epithelium and goblet cells (endoderm, R). Scale bar represent 50 um for (A, B, Q and R), 100 um for (C, D, M, N, O and P) and 200 um for (E–H).

Figure 5. Gene expression analysis of undifferentiated and differentiated progeny of the three hESC lines.

Pairwise correlation of NANOG expression with that of all tested genes of the panel of HADC-100, HADC-102 and HADC-106 hESCs is presented in (A). Gene expression is presented by the DeltaCt value of each gene, normalized against the average Ct of three endogenous control genes (Table S9). The Pearson correlation coefficient was calculated from the combined data of DeltaCt values of undifferentiated and differentiated hESCs of the three lines. The results were plotted in descending order of the correlation coefficient values. Clustering analysis of genes that were differentially expressed in the three lines in the undifferentiated and differentiated state is presented in (B). An unpaired t-test for differential gene expression between undifferentiated and differentiated samples yielded 35 genes with a P- value <0.05. Two-way hierarchical clustering with Euclidean distance was performed on the 35 genes and six samples. Green colors in the heat map represent low DeltaCt or high expression. Red colors represent high DeltaCt or low expression.