In this manuscript we describe a process that is readily scalable for production of functional dopaminergic neurons from hESCs for potential therapeutic applications. This process can be transferred to a GMP facility that could generate a qualified product for clinical use. The process can be broken into four steps: 1) derivation and propagation of ESC, 2) propagation and storage of NSC, 3) induction of a midbrain dopaminergic precursor population, and 4) maturation of the precursor to dopaminergic neurons. Our data suggest that each step could utilize xeno-free defined media that are suitable for scalable GMP manufacture of cells for clinical use. Neurons generated by this process appear to be authentic A9 dopaminergic neurons as assessed by in vitro (expression of midbrain dopaminergic markers) and in vivo (ability to survive after transplantation) assays.
Although it is not necessary to exclude xeno components in the culture and differentiation of hESCs for clinical applications, provided appropriate validation tests are performed, we believe that xeno components should be avoided as much as possible, because exposure to animal products raises unique concerns for medical use 
. Analyzing the components in our defined medium preparations, we identified three major sources of xeno material: BSA as a component of various media formulations, geltrex (which was used as a substrate in several stages of the process) and PA6-CM. In addition, there was the potential of exposure to xeno components as part of the hESC derivation process. We showed that hKSR (in completely humanized base medium with growth factors) could replace BSA and serum containing medium, and that a defined substrate - fibronectin (CellStart) - could substitute for geltrex in the culture of hESCs and hESC-derived NSC, and the process of inducing NSCs. Likewise, Shh and FGF8 could substitute for PA6-CM in the induction process and BDNF and GDNF along with TGFβ could substitute for the maintenance and survival effect of PA6-CM.
One advantage of our process of generating dopaminergic neurons from hESCs is the potential for scalability. As shown in the results, NSCs derived from defined cultured hESCs could be frozen and thawed, and NSCs cultured for prolonged periods retain the ability to differentiate into authentic A9 dopaminergic neurons both in vitro and in vivo. The efficiency of dopaminergic differentiation is sufficient for scalable production, as a high percentage of dopaminergic neurons can be obtained without enrichment.
A process amendable for GMP manufacture does not simply require the development of components, but also replicable tests to establish the quality of the end product. Toward this end, we have employed a large-scale microarray analysis and identified a panel of stage-specific markers that could be used to rapidly assess the quality of the end product. These include markers of NSCs, dopaminergic precursors and mature dopaminergic neurons, expression of which could be reliably assessed by real time PCR and/or immunocytochemistry.
Although the protocol we described works with multiple hESC lines, and we and others have shown that multiple lines can be adapted to serum free conditions, we acknowledge that we have not yet successfully derived a line in completely xeno-free defined conditions and shown that we can successfully generate dopaminergic neurons using this process. As an interim measure we have adapted a hESC line derived under nearly xeno-free condition 
to defined medium conditions using CellStart as a substrate at an early passage culture. We show that such an adapted line can be readily differentiated into dopaminergic neurons using our four-step protocol and the defined media and growth factors described. We believe that these results suggest that it will be possible to derive cell lines in hKSR on a defined substrate and that such lines will be similar to other hESC lines and will differentiate into dopaminergic neurons following this protocol.
We note as well that we have used growth factors and components that were themselves not manufactured under GMP protocols as would be required for a true GMP process, nor was this process tested in a GMP facility. In addition, we have not demonstrated a GMP-applicable process for purifying dopaminergic neuron cultures. There were several reasons why we did not undertake these efforts even though we believe that these are easily solvable problems. One important reason was cost, as developing GMP grade material such as antibodies and growth factors would be required prior to finalizing a protocol. The second was the lack of access to a GMP facility that had the requisite expertise in adherent cell culture and dopaminergic neuron differentiation, and the absence of consensus on what degree of purification (if any) is required for a final product. We have however initiated preliminary discussions with commercial providers who have assured us that such products can be developed when they are required.
In summary, we have shown that hESCs and NSCs can be maintained in xeno-free defined media for a prolonged period of time while retaining their ability to differentiate into authentic dopaminergic neurons. Our defined medium system provides a path to a scalable GMP-applicable process of generation of dopaminergic neurons from hESCs for therapeutic applications, and a ready source of large numbers of neurons for potential screening applications.