Erika Lapinskas, Product Manager, Laboratory Technologies, Stem Cells & Tissue Engineering at Sartorius Stedim Biotech

Erika Lapinskas

What are the key challenges that face cell-based bioprocess manufacturers today? 

Several issues present challenges for manufacturers today, including purification and stabilisation, delivery, yields, contamination, process development and scaleup.  

Harvest of product cells from adherent cell bioreactors is one of the critical bottlenecks in the cell therapy manufacturing process.

Many processes utilise trypsin or other enzymatic digest methods to remove cells from cultivation surfaces, but the cells must be neutralised, washed, analysed for purity and formulated within a narrow window of time — as short a timeframe as possible and typically in less than 1 hour — to produce a reliable product. Many cell production processes are patched together from singleuse bioprocessing containers designed for supernatant rather than cell harvest. Factors such as particle loads from singleuse bags, smaller process volumes and lack of terminal sterilisation of the product all impact on process design, as these factors are not a concern for supernatant processing. There is limited equipment available that has been purpose-designed for sterile, scalable cell harvest, washing and concentration, although this situation is changing.  

Products can currently be formulated and filled into storage or delivery bags or vials for fresh or frozen inventory. Particularly if products are shipped fresh, the product should be quantified for stability, viability and preferably dose-response.  

The factors mentioned above all contribute to delivery concerns with both fresh and frozen product. In addition, the manufacturing model should be considered — whether the product is allogeneic or autologous, and manufactured in small batches close to the site of delivery or at scale at a central manufacturing location. Although fixed costs may be lower, operators may be better trained and have greater process experience, and the product quality and consistency may be easier to control at centralised sites, none of this matters if the product degrades unacceptably during transport. Each product needs to be assessed separately due to varying inherent cell lineage characteristics to determine whether fresh or frozen is acceptable and to determine parameters, that can even possibly be monitored throughout transport, which indicate the cell population viability. 

One of the more troubling cell therapy process development issues is trying to develop a consistent, safe and efficacious autologous stem cell therapy when inherent genetic, and perhaps epigenetic factors, cause cells from different donor/patients to grow and differentiate at different rates. This means that multiple batches cannot all be maintained by the same protocol on the same time schedule, which currently precludes automation of that process. Relevant process variables will need to be defined and measured to take the 'art' out of producing such cultures before this system can become amenable to automation; therefore being safer and, over time, more efficient and cost effective. The development and integration of additional PAT into bioreactors will enable automation in the future. At the moment, many assays indicating the culture's growth and differentiation status are done manually. Development of online, singleuse sensors with feedback loops will permit automation. Additionally, the FDA guidance on manufacturing autologous somatic cell therapies indicates that acceptable ranges for operating and control parameters be defined for a cell growth process — typically these would include cell doubling time, cell purity, viability and culture time. These native differences in the donor cell populations may necessitate the use of alternative parameters or validation of a wider acceptable range than the manufacturer would prefer.  

Scaleup of adherent cell culture has always been a difficult challenge in biological manufacturing. Some manufacturers of laboratoryscale tissue culture plasticware make 'larger scale' units. While these ideally mimic the smallscale cultivation situation, which facilitates development of the growth process, they do become unwieldy on a larger scale and still have an upper limit on surface area. Therefore, manufacturers are forced to multiply the numbers of bioreactors they run, which mitigates risk and also creates more labour and batch size limitations, limiting economies of scale and increasing the quantity of testing, verification and documentation.

Limited purposedesigned, singleuse equipment is available for the cell harvest, wash and concentration part of the process. This is one part of the process where adaptation from a supernatantfocused process is of limited utility; those processes often use continuous (reusable) stackeddisc centrifuges that are efficient for processing volumes over several hundred litres and which pellet and discharge cells. For maintenance and recovery of a viable cell population, however, lower shear methods that don't deprive cells of oxygen and nutrients would be more ideal. 

Already, new embryonic stem (ES) cell line derivations can be made much more stable and robust than earlier derivations; various 'adult' stem cells can be grown reliably and defined, whilst animal-free media have been developed for cultivating such cells. This evolution mirrors that of the development of Chinese hamster ovary (CHO) and similar cell lines and their media for monoclonal antibody production more than 20 years ago. The sheer volume of current research detailing biochemical pathway inhibitors, surfaces for preferential growth or directed differentiation of many different stem cell lines and moves from 2D adherent culture to 3D culture (whether on microcarriers or using other suspension techniques) is indicative that production of sufficient cell numbers for allogeneic therapy will quickly become more easily achievable. 

For autologous therapies where the donation of cells is small, this may be more difficult. Yield needs to be sufficient for both the therapy and safety/efficacy release testing. In this situation, guidelines may need to be more flexible for testing as there are relatively few tools available to increase yield for lowdividing cell types.  

Relatively few biological processes have absolutely required top-to-tail sterility assurance, even if that situation has been ideal. Most processes have had some option for terminal sterilisation whether by heat or filtration — a situation that can't work with cells. Most largescale cell manufacturing processes use singleuse bags, tubing, aseptic connectors or tubing welders and sealers in a clean room to create a continuous aseptic process. For smaller scale processes — particularly patient-specific autologous therapies — this becomes problematic and the costs rise. Some of these therapies are made using multiple open manipulations in a laminar flow hood; robotics and new equipment development from manufacturers with a cell therapy focus are targeting this area. 

Of those challenges discussed above, could you please explain how they can be overcome?

Vendors will need to overcome shortcomings in their equipment and disposables that limit their use in this nascent bioprocessing market. These may be 'small' issues (such as temperature limitations on transfusion bag film or having bioprocessing singleuse disposables in the right size) or larger issues (such as the lack of critical harvesting equipment or small-batch final bag fill equipment). Establishing cooperations and information sharing between researchers, manufacturers and vendors will assist the rapid development of appropriate processing equipment. Likewise, involving the regulators in this process will optimise the concurrent workup of manufacturing and compliance with existing regulations, as well as those that will inevitably be added as this field matures.  

Manufacturing models will need to be assessed on a productbyproduct basis. For example, having small scale manufacturing close to the point of harvest and use makes sense for most autologous therapies; indeed, several vendors have been developing selfcontained, easily transportable 'manufacturing suites' to serve this model. Allogeneic therapies that can be manufactured on a large scale will most likely utilise centralised facilities and have to put more effort into validating the transport and storage chain. This is something that will have to be assessed on a perproduct basis. 

What are the main obstacles to implementing QbD principles to cell-based bioprocessing? 

One of most conducive factors for implementation of QbD in cell therapy processes is that much cell bioprocessing is new, i.e., without legacy plants and protocols. With cell therapies arising from basic biological research, however, early development work may have been done in a laboratory without an understanding of GMP and with insufficient process or development documentation. Many cell biologists have a 'feel' for growing cells, but this doesn't translate into a reproducible, documented process. The aim is to design the process, rather than the personnel, to be reproducible and, depending on the origin and training of the personnel who invented or developed the therapy, the culture of the workplace may need to shift to embrace the principles of QbD. 

Improvements in purification techniques — their specificity, scalability and options for single-use —will help the industry progress to make a more reproducible, safer cell product. 

The innate variability of a cell type derived from different sources is also a factor for autologous therapies. If a process is workedup on mesenchymal stem cells (MSCs) from a selection of sources; the parameters for growth, viability, etc. could vary depending on the site the MSCs were harvested from, the age of the donor if the patients to be treated are dissimilar from those donors used for the process validation, genetics and general health. It is unlikely that it will be possible to validate a process with just one cell source or to draw from master or working cell banks that can be used for allogeneic therapy production.  

The cell processing industry has learned lessons from the existing bioprocess industry, from where many key personnel have transferred, and this manufacturing experience is essential to balance out the basic R&D origin of many therapies currently in development. 

What equipment types are most widely in use in the bioprocessing of cell-based therapies today? Where is there room for improvement in current practices?

Adoption of singleuse processing tools is almost at saturation level within cellbased bioprocessing. While singleuse bioreactors and fluid handling systems are still on the upswing within traditional bioprocessing, they've proven to be absolutely essential for cell production due to the sterility assurance, elimination of crosscontamination risk and negligible contamination risk from environmental toxins (residues, endotoxin etc). Cell processing does not allow for the removal of such incidentals, so manufacturers have adopted this equipment to prevent risk of introduction during processing. 

We have already discussed challenges in scaleup of product filling and cell harvesting equipment/techniques, but bioreactor design is another area where some improvements will be made over time. Various groups are working on stem cell cultivation techniques to enable less spaceintensive scaleup, an issue that other bioprocessing industries haven't faced directly; CHO and other cell lines used for protein production are typically adapted for suspension culture. Most stem cell types are unconditionally adherent. Cultivation as aggregates, in alginate encapsulation, on 3D matrices of various sorts and on microcarriers are all works in progress. Improvements in design, handling and automation of adherent cell bioreactors is possible within a limited scope, but overall the industry may move to using adherent surfaces in suspension reactors (e.g., Cultibag RM; Sartorius Stedim/wave reactors or stirred tanks) if technical problems with adapting to microcarriers or suspended 3D matrices can be solved. This would also enable further process automation, including use of singleuse pH and dissolved oxygen sensors to monitor and control cultivation parameters, reducing handson time. 

What are your predictions for the future of cell-based bioprocessing? 

Cell purity and sorting requires innovation in techniques and equipment from industry. Stem cell populations are rarely, if ever homogeneous, and most therapies require expansion of the cells and differentiation. Currently, this is not an 'allornothing' phenomenon — some cells will differentiate down other lineages than desired. For safety and efficacy reasons, cell products will need to be defined as part of their release testing. Depending on the derivation of the product, some contaminating cell types may be dangerous to the patient (e.g., undifferentiated ES cells) or simply have an unknown effect on the 

 culture. Sorting techniques, such as fluorescenceactivated cell sorting, can produce a pure population, but most (with the exception of the BD Influx single-use fluidics option; BD Sciences) have reusable product contact components that are difficult to sanitise confidently. Positive and negative bead selection and centrifugal elutriation provide some selection, but only by a limited number of parameters. The remaining cell populations and any carryover from the purification process needs to be tested for nontoxicity and stability. Improvements in purification techniques — their specificity, scalability and options for single-use —will help the industry progress to make a more reproducible, safer cell product.

 Product Manager, Laboratory Technologies, Stem Cells & Tissue Engineering at Sartorius Stedim Biotech

Articles

Discussing the key challenges that face cell-based bioprocess manufacturers today, how these challenges might be addressed and the innovations that could facilitate the process in the future.