3 Dimensional Cell Based Assays in Hollow Fiber Bioreactors
John J S Cadwell, President and CEO, FiberCell Systems Inc
William J Whitford, Sr Manager, HyClone Cell Culture , GE Healthcare

Hollow fiber based animal cell culture was first developed by Richard Knazek at the NIH in 1972 (Knazek et al, 1972). Hollow fiber bioreactors (HFBR) offer a method by which cells can be cultured at tissue-like densities over long periods of time. Hollow fibers act as "artificial capillaries" and perform much as capillaries do in the human body. These bioreactors hit their peak of popularity in the late 1970's to 1980's where they were employed in the bio-manufacturing of monoclonal antibodies (Tharakan and Chau, 1986).

The biomimetic HFBR system is a high-density continuous perfusion culture system. It presents many unique distinctions from the commonly employed non -porous plastic surfaces of eg, flasks, microcarrier beads and discs or roller bottles. An HFBR includes a cartridge containing thousands of semi -permeable hollow fibers in a parallel array within a tubular housing fitted with inlet and outlet ports. These fiber bundles are potted at each end so that any liquid entering the ends of the cartridge will necessarily flow through the interior of the fibers. HFBRs present a 3-D environment similar to the conditions found in vivo, and support the continuous control of such parameters as oxygenation levels, medium composition, drug concentration and shear stress. HFBRs are an effective means for in vitro assays and the generation of a number of products, from secreted proteins or viruses to cells or conditioned medium

The fact that in a HFBR the cells are bound to a porous support provides a number of distinct features. Cultures in this system can maintain viability and production-relevant metabolism in a post-confluent manner for extended periods of time-months or longer. Another advantage is that due to the extremely low shear generated with the cartridge when cells become necrotic they will not release significant cytoplasmic proteins or DNA into the culture medium. Through the selection of fiber porosity, desired products can be retained to significantly higher concentrations and the location /effects of cytokines can also be controlled. Recombinant proteins can be selectively retained and concentrated and cytokines and other factors that facilitate cell-to-cell interactions can be concentrated as well. Small molecule drugs can easily exchange across the fiber and rapidly reach equilibrium.

The small diameter of the fibers (in the order of 200 microns) generates an extremely high surface-area-to-cartridge volume ratio in the range of 100 -200cm2/ ml. Coupled with the high gross filtration rate of the more optimised polysulfone fibers, the rate of exchange of primary and secondary metabolites appears high enough to support any practical purpose. Productive cell densities of 1-2x108 or more have been reported approaching in vivo tissue-like densities (Pera 2014).

Operation of a HFBR, in its most simplified form, begins by seeding a prepared cartridge with either suspension or harvested adherent cells. The reactor cartridge is connected to an external reservoir and the medium recirculated from the reservoir through the cartridge. Mass transfer of gasses can be accomplished in a variety of ways, with one being diffusive exchange through a loop of gas permeable silicone tubing prior to the medium entering the bioreactor itself. Medium recirculation rate and culture feeding can be linked to any number of culture parameters, and a number of control and automation options have been explored. During this renaissance of perfusion employment in general, and of HFBR in particular, several characteristics of hollow fiber cell culture have recently been identified:
  1. Reduction in apoptosis (Weeraphan, et al, 2012)
  2. Consistency of culture over long periods of time (Pavlakis, 2013 )
  3. More in-vivo like growth conditions resulting in improved cell function (Bennet et al, 2007)
  4. Facilitation of the use of serum free, protein free and chemically defined media formulations (Whitford and Cadwell, 2011)

HFBR provides many particular (and some unique) culture characteristics due to a number of physical and ambient chemical conditions provided by the system, including;
  1. Perfused medium flow and porous support permits long-term culture
  2. High cell density culture increases cell-to-cell contact
  3. Selectable MWCO of fibers concentrates interactive cytokines
  4. Selectable MWCO of fibers segregates cells, metabolites and products .
  5. Directional flow establishes gentle interstitial gradients within the cell mass.
  6. Hydrodynamic (shear) force on endothelial cells required for proper physiology.
  7. Long-term high-density culture on porous support facilitates cell-to -cell interactions.
P.C. Liao et. al. have identified a potential mechanism for the greatly reduced degradation of product during HFBRs highly extended culture duration (Wu et al, 2009). They found a significant reduction in apoptosis of up to 90 percent. This, and the absence of cell-disruptive shear forces inside the HF cartridge, results in dramatically reduced post-secretion product alterations and cleaner harvests (Srisomsap et al, 2010).

In a HFBR, the cells are bound to a porous support, cell division rate and generation number is reduced, and cultures do not require splitting. Passage number becomes irrelevant and cells grow in multiple layers in a "post-confluent" fashion as exemplified by Dr George Pavlakis (NCI, Frederick, Md). Here, 293 T cells were transformed to produce a very complex, problematic protein. While attempts at producing this protein in standard cultures modes were unsuccessful, HFBR provided properly dimerised product with complete and consistent post-translational modifications through over 140 days of production.

In-vivo, most animal cells grow in 3-D at very high density under tightly defined and highly controlled conditions. There is very little variation in oxygen tension, pH, glucose levels etc. It has been recently observed "The 2D cell culture systems used so far have several drawbacks: the morphology, proliferation, metabolism and expression profiles of cells grown in 2D systems are very different to cells in living tissues" (3D Cell Culture 2012 ). The consistency of HFBR culture conditions has a direct effect on cell physiology and product generation.

The culture-contact surfaces of many systems are composed of single-use (SU) materials and provide a number of benefits in manufacturing and assays systems. HFBR culture characteristics, such as very high cell density, allow for a reduction in serum concentration and facilitate adaptation to commercially available serum free media. This has been taken one step further with the introduction of commercially available perfusion optimised serum replacement (Whitford and Cadwell).

Applications for Cell Based Assays
Cell based assays and in vitro testing methods are a useful, time and cost effective tool for drug discovery. However, it is generally accepted that many of the available assays are not effective for examining the effects of both time and concentration, and do not closely mimic physiologic kinetics. More specifically, that they do not report pharmacodynamic actions (what a drug does to the body) and pharmacokinetic actions (what a body does to the drug). Static cell based assays in plates, flasks or other formats do not readily permit changes in drug concentration as would be seen in humans from administration, uptake, distribution, distal metabolism and elimination effects. Animal models generally do not provide the same drug kinetics as would be found in humans, many infections cannot be supported with an animal model and many times the bacterial load is not high enough to reveal the emergence of resistance. HFBR cartridges have continuous medium circulation supporting dynamic control of drug concentration over time and resulting in the mimicking of naturally occur gradients in tissue drug concentration. A high surface-area-to-volume ratio permits extremely rapid exchange of metabolites and pharmacoactive molecules between the central reservoir and cells growing in the relatively small ECS of the cartridge. The volume of this central reservoir can be easily adjusted to permit rapid and reproducible changes in drug concentration.

Simulation of the kinetics of multiple drugs can also be accomplished so drug/drug interactions and combination therapies, as well as transport and efflux, can readily be modeled. The system is compact enough that multiple cartridges can be conveniently manipulated in a relatively small space, providing multiplexed or parallel and higher throughput type activities. Such systems can be configured for cell based assays employing either a single-cell type or multi-cell in co-cultivation.

Examples of HF-Based Cell Assays
HF based assays are inherently more complex and costly to design and set up than conventional cell-based assays. However, these assays can generate data that is not available in any other manner, and can bridge an important gap between animal studies and phase I clinical trials. Large numbers of cells can be assayed and over a period of time. Drug concentrations can be controlled in a dynamic fashion and both adsorption and elimination curves can be modeled. Multiple tests can be performed on the same cell populatio. Three dimensional cultures of multiple cell types can model complex processes such as virus infections in tissues, hematopoiesis, cancer cell propagation, cancer cell metastasis, and the blood brain barrier (Mancuso et al, 1990).

Assays Using Only One Cell Type
The simplest type of 3-dimensional cell based assays performed in hollow fiber bioreactors consist of only one cell type, seeded in the ECS. HF bioreactor culture is the only cell culture method that can support cells at physiologic cell densities and provide for in vivo-like viral infection- and pharmaco-kinetics at these densities. Both adherent and suspension cells can be cultured in a HFBR. Adherent cells bound to a porous support do not require periodic splitting and can be maintained for extended periods of time. Suspension cells can also be supported for extended periods of time due to the same constant feeding and removal of metabolites.

The ability to add medium with or without drugs is particularly important for dose fractionation pharmacodynamic studies where compounds are added to the system over a short period of time and then removed by dilution with drug free medium without disturbing the cells or their environment. Both absorption and elimination can be simulated. The small volume of the ECS contributes to the rapidity and economy with which this equilibrium takes place. Lastly, and perhaps most importantly because of their large size viruses and virus infected cells are retained in the small volume of the extra-capillary space. These infectious agents cannot cross the fibers into the medium. The system is completely closed to the external environment and provides an added biosafety component protecting laboratory personnel from exposure.

Antiviral Pharmacodynamics
HFBR support the establishment of a pharmacodynamic index (dose and schedule ) of a particular entity for a particular virus:

The high density culture within HF reactors supports the efficient (biomimetic) cell-to-cell spread of virus is very for either matrixed or suspension cells. In either case, released virus and virus-infected cells accumulate in the ECS over time. Computer controlled pumps administer drug through a port in the central reservoir to model any schedule of ambient drug exposure.

The concentration of antiviral drug in both the reservoir and the ECS itself can also be monitored by regular LC/MSn (or equivalent) assay of representative samples. The dosing regiment providing inhibition of viral replication and/or cell-to-cell spread of virus can be determined by sequential analysis of cell and ambient media viral titer and drug concentration from identified points later validated by, eg, LC/MSn. (McSherry, et al 2011).

Recently, the nucleoside analogue, 2', 3'-didehydro-3'deoxythymidine (d4T) was examined using hollow fiber infection models (HFIM) (Drusano et al, 2002 ). In separate experiments hollow fiber units infected with the same amount of virus and treated in the same way, but with d4T at half these doses failed to completely inhibit virus replication. The HFIM system predicted that the minimum effective dose of d4T to treat patients infected with HIV was approximately 0.5 mg/kg/day administered twice a day. This prediction was confirmed in a clinical study (Anderson et al, 1992).

Protease inhibitors: A hollow fiber system was used to determine the minimum concentration of the protease inhibitor A-77003 that would inhibits the replication of HIV in CEM cells (Preston et al, 2003).

Vaccinia Virus
For improved pharmacodynamic studies on the smallpox (varola) model virus, vaccinia, some researchers are employed hollow fiber-based models. One group looked at the effect of cidofovir on vaccinia virus replication in the HeLa-S3 cells monitored by FACS analysis of virus-infected cells and by the production of infectious virus using a plaque assay (McSharry et al, 2009a ).

Influenza Virus
The current recommendation for treatment of influenza with oseltamivir is to take two 75 mg tablets twice a day. Recently researchers employed hollow fiber-based models to performed dose range and dose fractionation experiments in MDCK (McSharry et al, 2009b). The data showed that in the absence of drug the virus grew well in the HFIM system and that the pharmacodynamically-linked index for oseltamivir for the R292 strain of influenza A virus is the AUC/EC50/95 ratio. This means that the model indicates that at the appropriate dose, oseltamivir could be given once a day. The demonstration that adherent cells can be used to grow virus in the HFIM system opens this system up to the pharmacodynamic analysis of antiviral compounds for a wide variety of viruses (Brown et al, 2011).

Anticancer Agents
Another example of a 3-D cell based assay using only one cell type in the ECS of the hollow fiber cartridge is for the analysis and characterisation of anticancer agents. Anti-cancer agents also exhibit both time and concentration dependent efficacy.

Mark Kirstein reported on the use of the hollow fiber model for anticancer drug evaluations. In this case gemcitabine was examined in the anchorage dependent MDA-MB-231 breast cancer cell line (Kirstein et al, 2006). More accurate results are obtained for a few reasons. One, because the multi -layer, 3-dimensional organisation of these continually perfused cells more closely reflects the in vivo structure of the tumor, they have an increased relevance for assays of anti-cancer agents. Another is that the cell division rates in static cultures are commonly artificially high, and this can render them more sensitive to some chemotherapeutic agents then their natural counterpart (Kirstein et al, 2008).

Assays Using More Than One Cell Type
Hollow Fiber and 3-D Co-cultivation for Cell Based Assays

HFBR-based culture is one of the few in vitro techniques providing large numbers of cells in close enough proximity and sufficiently high density to observe a number of tissue-like behaviors. These behaviors include:
  1. Signaling and direct interactions between different cell types
  2. Coordinate (or synergistic) activity upon the ambient media
  3. Coordinate (or synergistic) activity upon a compound or drug
  4. Mixed cell type-specific effect (+/-) upon of viral infections
HF cell culture permits the recapitulation of natural structures containing more than one cell type in a defined, controlled, and more biomimetic environment. Endothelial cells are the only cells that can be easily cultured on the insides of the fibers. There are two types of cell co -cultivations that can be performed in a hollow fiber bioreactor. They are, one or more cell types
  1. on both the inside and on the outside of the fibers (Davis, 2007 )
  2. on only one side of the fibers (typically on the ECS)(FCS Application, 2012)

An example of the first type of cellular co-cultivation is the use of HFBRs to culture endothelial cells on the insides of the fibers while culturing a different cell type on the outside to provide cell signaling (Redmond, Cahill, and Sitzman, 1995).

The concept of "organ recapitulation" in hollow fiber was first applied by Jorg Gerlach using liver tissue (Gridelli et al, 2012). The system used was a complex HFBR with two different fiber types and three separate bundles of fibers. Primary function in a mixed population of liver cells was maintained for 4 weeks.

Bone Marrow Model
Perhaps the most rigorous application of HF systems for single-compartment cell co-cultivation is in the area of stromal cell/suspension cell interactions. In demonstration of the second type of culture, Dr Mayasari Lim at University Hospital, Hong Kong has published an article on the co -cultivation of a human bone marrow stromal cell line with a leukemic T cell line (Usuludin, Cao and Lim, 2012). When cultured in the HF cartridge the T cells underwent a 4000 fold expansion, in line with what occurs in vivo.

Recently, the efficacy of such an HF platform was evaluated in comparison to standard cultures performed on tissue culture polystyrene (TCP). A human stromal cell line (HS-5) was employed as a co-cultured stromal support of lineage-cell depleted human cord blood cells (Xue et al, 2014).

Results showed that the performance of the HFBR in supporting total cell and CD34+ progenitor cell expansion was comparable to that of cultures on TCP, while cells harvested from the HFBR had a higher clonogenic ability. The findings demonstrate the feasibility of utilising an HFBR for creating a complex cell matrix architecture, which may provide good in vitro mimicry of the bone marrow supporting large-scale expansion of HSCs.

Asymmetric Co-Culture using Endothelial Cells
3-Dimensional cell based assays utilising co-culture of multiple cell types in hollow fiber bioreactors can recapitulate more complex structures than those with a single cell type. One type of cellular co-cultivation, as described above, is the use of hollow fiber bioreactors to culture endothelial cells on the insides of the fibers and a different cell type on the outside. In one study, altering the flow rate changed the shear stress, g-protein formation and endothelin receptor expression was directly modulated, even though there was no physical contact between the two cells types (Redmond, Cahill and Sitzman, 1995).

Blood/Brain Barrier Model
There are many in vitro approaches to the modeling BBB physical and biochemical behavior, but most fail to represent its natural three -dimensional nature, and do not support the associated exposure of endothelial cells to such complex influences as exist in vivo. To answer this challenge, Janigro developed a new, dynamic, in vitro BBB model (NDIV -BBB) designed to allow for extensive pharmacological, morphological and physiological studies (Stanness et al, 1999). His new dynamic HF-based model of the BBB allows for longitudinal studies of the effects of flow and co -culture in a controlled and fully recyclable environment.

In perhaps the ultimate embodiment of state of the art hollow fiber based cell assays, Dr Chris Pepper et al have developed a model for primary investigations and assaying for cancer metastasis (Walsby et al, 2014. A dynamic in vitro model was developed in which CLL cells experience shear forces equivalent to those in capillary beds and are made to flow through capillary-like hollow fibers lined with endothelial cells. CLL cells treated in this way increased their expression of CD62L, CXCR4, CD49d and CD5 and migrated through the endothelium into the ‘extravascular’ space’ (EVS). The degree of migration observed strongly correlated with CD49d expression and treatment with the CD49d blocking antibody. Taken together these data provide evidence for a novel, dynamic and reproducible in vitro model of lymphocyte migration and cancer metastasis.

Novel material features supporting a renaissance in the technology include fibers composed of new materials, new surface derivatisations and porosities providing improved binding, flux and flow rates. New and creative application development has kept pace with the availability of these new technologies expanding the scope of applicability of 3-D hollow fiber cell based assays. The result is a robust and flexible technology with diverse applications. Such applications range from protein biological production providing ultra-clean harvest and simplified purification-to improved in vitro viral infection systems providing more accurate drug candidate PK/PD modeling.