Serum-Free Media—Optimizing CAR T-Cell Therapy Production

Lonza describes how serum-free media can help manufacturers realize time, cost, and consistency benefits while bringing T-cell therapies to more patient.


By Amber Jones and Kayla DeOca, PhD

Since its inception, chimeric antigen-receptor (CART) T-cell therapy has relied on human serum to drive the T-cell expansion stage. The first published protocols used serum,1 and other researchers quickly followed. As CAR T-cell therapies have evolved, so have the culture systems that enable the best expansion and control over the manufacturing process.

Human serum is costly and troublesome, with wide batch-to-batch variability that requires laborious validation tests. Huge operational and labor costs are incurred for this stringent, lengthy testing, and yet CAR T-cell therapy production inconsistencies still occur. The result: expensive therapies that take considerable time to reach patients. This article discusses how using media that are completely free of serum or serum substitutes could help manufacturers attain higher degrees of control and efficiency in the production of CAR T-cell therapies. By taking advantage of the quality and consistency of serum-free media, manufacturers could deliver highly effective CAR T-cell therapies quickly and economically.

Opening serum’s black box for more effective cell growth

Human serum has been the main driver for T-cell expansion, primarily as it contains the critical growth factors needed to accelerate proliferation. Early in CAR T-cell research, there was a conscious shift from animal-derived media, such as fetal bovine serum, to human forms to mitigate adverse immune reactions. For example, in one study where donor-derived T cells were isolated and expanded in media containing fetal calf serum, some patients developed antibodies to the serum during their T-cell immunotherapy.2,3

Serum is a black box full of many elements, including cytokines, agonistic and antagonistic regulators, undefined antigens, proteases, and even degradation factors. To optimize T-cell expansion and minimize immune responses, we must fully understand serum’s components, and for future processes, we must isolate and include only those factors that are beneficial.

In culture systems, T-cell activation typically involves the use of antibodies. However, normal human serum contains functionally active components of complement which raise the risk that antibody-dependent cellular cytotoxicity will result in the depletion of T cells.4 Cytokines also display varying stability when stored in serum,5 which could cause an issue when culturing cells from stock media that contain serum and cytokines. Since the amount of cytokine can gradually destabilize and decay over time, cytokine stimulation given to the cells can be suboptimal.

The current, serum-based protocols that are used to produce CAR T-cell immunotherapy need to cross many technical barriers during the expansion and culturing phases. For example, most individuals have peripheral blood mononuclear cell (PBMC)-derived T cells that are autoreactive and capable of driving autoimmunity.6

This is of importance because human serum is composed almost entirely of self-proteins, self-peptides, and self-antigens. In this culture system, T cells are consistently exposed to self-antigens, thereby increasing the chance of activation and outgrowth of autoreactive T cells, rather than the antigen-specific cells of interest.

By being constantly presented with serum-based antigens, T cells are also liable to be in a permanently activated state that could lead to T-cell anergy. In contrast, T cells expanded in serum-free media come to rest quicker than cells expanded in serum-containing media.7 This resting state is important because, once rested, T cells can be reactivated more effectively once they are transferred to the patient. This is also true in culture, where rested T cells generate greater populations during expansion.

Serum—An inconsistent and costly component

The undefined nature of human serum is not the only reason that developers are looking to remove it from the CAR T-cell therapy manufacturing process. Obtaining high-quality lots of serum is difficult, and the prices (typically $800–1,000/L) are high.

Since no two human serum samples are the same, huge batch-to-batch variability can occur due to donor differences. This variability affects not only the quality of the expansion stage but also the patient’s response to a specific lot, directly impacting the effectiveness of the therapy.

To mitigate the risks associated with lot variability and differing patient responses, each lot undergoes weeks of testing against comparison data to prove its likely effectiveness. Growth promotion tests, viability analysis, characterization of cell surface markers, phenotyping, and cytotoxicity profiles are all common analyses, which can take upward of four to six weeks (with one to two weeks needed just to evaluate sufficient growth). Each test aims to establish a like-for-like comparison with previous expansions to increase the likelihood of delivering a treatment that performs consistently.

Although testing is critical, it requires extensive time from specialists, adding to the burden of human serum use. The time, resource, and material costs accrued during this stage all add to the overall cost of CAR T-cell therapy.

Defining and controlling expansion to accelerate time to market

Truly effective cell media should contain all elements needed to support and enhance effective T-cell expansion, without detrimental elements. Early substitutes for whole serum still contained human blood components and, therefore, presented the same issues around batch-to-batch variability, the need for rigorous risk assessments, and slower regulatory approval.

By creating a fully defined growth environment completely free of human serum or its substitutes, scientists can access predictable and consistent T-cell expansion, and gain substantial rewards. First, when the high variability between donor responses is eliminated, researchers and clinical teams can better understand the impact of the therapy being used. Second, when the labor demands and costs associated with testing and enabling smoother regulatory approval become more manageable, lower-cost therapies can be brought to market faster. Third, when human serum is removed, more patients can access life-altering treatments without the risk of batch-to-batch variability. It is only a matter of time before serum-free media will bring this scenario into the mainstream.

Navigating the transition to a serum-free future

The use of serum-free media has received its fair share of negative feedback from the CAR T-cell therapy community, and rightly so. Many researchers tried earlier serum-free options and found that they didn’t work as well as human serum, so they added serum to achieve the required results. However, the latest advances in serum-free media really do mean serum-free, containing only the elements that benefit the stable and consistent growth and proliferation of T cells.

In a recent study evaluating CAR T-cell growth, T cells grown without serum displayed improved in vitro functions compared to T cells from serum-containing media.7 In fact, the inclusion of human serum hindered the ability of CAR T cells to function, persist, and control tumor growth in vivo. CAR T-cell therapy produced without serum durably controlled leukemia while the T cells grown in serum offered only transient control.7 These data demonstrate the importance of shifting toward a serum-free media future to accelerate 
and promote better, more efficacious CAR T-cell therapy.

Incorporating truly serum-free media in therapy protocols, though, entails modulation to workflows and a slightly altered way of protecting and monitoring cells. Unlike the repetitive validation processes needed to establish lot viability with serum-based media, this adjustment will need to happen only once, with each subsequent expansion using the same simplified and accelerated process.

Key changes will include:

• Waste management. Without serum, waste products (especially lactate and ammonia) will need to be carefully monitored and media replaced when levels reach a predefined limit.
• Cell handling. Cells will need gentler handling without serum as a shear protectant.

• Oxygenation. Cells naturally have a high oxygen demand and, without serum, will require aeration. Static bags and flasks of medium won’t provide the conditions needed.
• T-cell quality. Choosing high-quality, purified T cells will reduce cell death and prevent lysosomal build-up. Lysosomal levels can be further reduced by removing dead cells.

Although adjustments will be needed, these small efforts will be far outweighed by the time, cost, and consistency benefits delivered by serum-free media. Media completely free of serum and its substitutes is already used in the market, and it has demonstrated sufficient T-cell expansion (Figure 1).

Figure 1. Serum-free media products can robustly expand human T cells without added serum. Primary human CD4 and CD8 T cells were mixed at a 1:1 ratio, activated with CD3/28-coated beads, and cultured in two serum-free media and RPMI either without added serum (circles) or with added serum (squares). (Adapted from Medvec et al. Mol. Ther. Methods Clin. Dev. 2017; 8: 65–74.7

Next-generation, serum-free media—A new standard in CAR T-cell therapy production

The use of serum-based media has a strong legacy but is not without its flaws. Human serum is difficult to work with and expensive. It involves laborious validation processes, generates large volumes of waste, and is prone to inconsistencies.

Keeping only the elements that support consistent T-cell expansion and none of the downregulators that drive unpredictable donor variability, truly serum-free media will soon become the go-to solution for CAR T-cell development. Such media are chemically defined and completely devoid of native human proteins—an important consideration for regulators who spend extensive time understanding the traceability and viral testing of these human components. It’s only a matter of time before these protocols become the standard for regulatory submission.

Early adopters of serum-free methods will reduce the time and cost needed to bring CAR T-cell therapies to market. Accordingly, the early adopters will have opportunities to increase their marketshare. Although human serum use will likely continue for historical reasons, adopters of serum-free media will change the future for patients, allowing greater numbers to benefit from lifesaving therapies


Amber Jones is a senior product manager at Lonza, and Kayla DeOca, PhD, is an R&D scientist.


1. Hollyman D, Stefanski J, Przybylowski M, et al. Manufacturing Validation of Biologically Functional T Cells Targeted to CD19 Antigen for Autologous Adoptive Cell Therapy. J. Immunother. 2009; 32(2): 169–180. DOI: 10.1097/CJI.0b013e318194a6e8.
2. Sundin M, Ringdén O, Sundberg B, et al. No alloantibodies against mesenchymal stromal cells, but presence of anti-fetal calf serum antibodies, after transplantation in allogeneic hematopoietic stem cell recipients. Haematologica 2007; 92(9): 1208–1215. DOI: 10.3324/haematol.11446.
3. Tuschong L, Soenen SL, Blaese RM, et al. Immune response to fetal calf serum by two adenosine deaminase-deficient patients after T cell gene therapy. Hum. Gene Ther. 2002; 13(13): 1605–1610. DOI: 10.1089/10430340260201699.
4. Fante MA, Decking SM, Bruss C, et al. Heat-Inactivation of Human Serum Destroys C1 Inhibitor, Promotes Immune Complex Formation, and Improves Human T Cell Function. Int. J. Mol. Sci. 2021; 22(5). 2646. DOI: 10.3390/ijms22052646.
5. Simpson S, Kaislasuo J, Guller, Pal L. Thermal stability of cytokines: A review. Cytokine 2020; 125: 154829. DOI: 10.1016/j.cyto.2019.154829.
6. Danke NA, Koelle DM, Yee C, et al. Autoreactive T cells in healthy individuals. J. Immunol. 2004; 172(10): 5967–5972. DOI: 10.4049/jimmunol.172.10.5967.
7. Medvec AR, Ecker C, Kong H, et al. Improved Expansion and In Vivo Function of Patient T Cells by a Serum-Free Medium. Mol. Ther. Methods Clin. Dev. 2017; 8: 65–74. DOI: 10.1016/j.omtm.2017.11.001.

Previous articleOne-Step ADC Linker Approach Bypasses Antibody Engineering
Next articlePrecision Control of Culture Conditions for Induced Pluripotent Stem Cells