The use of chimeric antigen receptor (CAR) technology has contributed towards significant advances in the treatment of certain types of cancer. This technology harnesses the immune defenses (e.g., T cells) to specifically target a patient’s cancerous cells with modified immune cells carrying a CAR “payload”. As with many new technologies, rapid progress is being made that overcomes the barriers and hurdles associated with earlier generations of the CAR T cell technology.

These next sections will discuss some of the more recent improvements to the development and manufacturing of CAR T cell therapies, including approaches to T cell isolation, engineering steps to produce CAR T cells, and strategies for the expansion of engineered cells for subsequent patient treatment (Figure 1). This section will also introduce some newer approaches that use natural killer (NK) cells as immunological weapons for cancer treatment. A short introduction to the biology behind CAR T cell therapies is provided here as a platform to discuss the manufacturing process.

Figure 1. Similarities and differences associated with autologous vs allogeneic approaches to CAR T cell therapy workflow.

What is CAR?

A chimeric antigen receptor (CAR) is an artificial receptor that is engineered to be expressed on immune cells such as T cells and NK cells. For CAR T cell therapies, the T cells are engineered to express a CAR protein that recognizes unique tumor antigens. The CAR protein is composed of an extracellular domain derived from a monoclonal antibody and an intracellular domain derived from T cells. The design and construction of these components is what make CAR T therapy one of the most advanced forms of adoptive cell therapies. The CAR protein is composed of three parts (Figure 2):

  1. The extracellular domain—a single-chain fragment variant (scFv), which is derived from a monoclonal antibody molecule specific to a unique tumor antigen (e.g., CD19 on leukemia cells)
  2. A transmembrane domain—to serve as an anchor
  3. Anintracytoplasmic domain—the “functional” component derived from T cells
Diagram of a chimeric antigen receptor, or CAR, attached to a T cell, showing the various structural components in greater detail.
Figure 2. Anatomy of a chimeric antigen receptor.

Extracellular domain (scFv)

The scFV or extracellular domain is the tumor antigen-binding domain, which specifies the CAR T target. It is located on the T cell membrane and is a single-chain antibody fragment derived from a monoclonal antibody. The scFv is made up of the variable region of a light and heavy (VL and VH) chain and is fused to the transmembrane domain with a short linker. Like any antibody, these single-chain antibodies can bind to protein, carbohydrate, and glycolipids [1].

Transmembrane domain

The transmembrane domain functions solely to stabilize the scFv portion of CAR on the T cell surface. The transmembrane domain is usually derived from CD8α but can also be based on CD4 or CD28 [2].

Intracytoplasmic domain

The intracytoplasmic domain, which is derived from the CD3 Ϛ chain, is the functional (or signaling) end of the CAR. After the binding of the CAR scFv to the tumor antigen, the CAR intracytoplasmic (CD3 Ϛ chain) forms a cluster, which will initiate activation signaling, ultimately leading to cytotoxicity of the tumor cells.

The design of each of these parts of the CAR is critical for the success of the anti-tumor response. As expected, there have been several improvements to make CAR T cells kill more efficiently, persist longer in vivo, and be less toxic. For example, a second-generation CAR added an immunomodulator at the intracytoplasmic domain (e.g., CD28 or CD137 (4-1BB)) and improved the killing machinery. When both immunomodulators CD28 and CD137 were added (third-generation CAR), persistency improved [3,4].

Autologous versus allogeneic CAR T therapies

Early CAR T cell therapy work relied on harnessing the power of T cells isolated from the cancer patient. The patient’s T cells were then modified to target cancer cells and infused back into the patient. This process, known as autologous CAR T, typically took 3–4 weeks and had an approximate failure rate of 7–10 percent [5]. While great success was achieved with this approach, autologous CAR T therapy manufacturing is a lengthy process that extends treatment timelines and is not scalable.

The limitations of autologous CAR T therapy can be overcome by engineering third-party T cells derived from healthy donors. These so called “off-the shelf” or allogeneic CAR T cells can be made in advance and released for immediate use when needed by the patient. Unlike autologous therapies that directly treat one patient, allogeneic therapies can treat multiple patients. Table 1 summarizes the many benefits to using an allogeneic approach compared to an autologous approach. In addition, the allogeneic approach provides a standard drug product produced from donor sources displaying an optimal immunological profile enriched with stem cell memory T cells (TSCM). This could make allogeneic CAR T cell products a first line therapy for B cell malignancy.

Table 1. Differences between autologous and allogeneic CAR T therapies.

Autologous CAR T productAllogeneic CAR T product
One product for one patientOne product for multiple patients
Patient donor—high variability in quality/quantity (low TSCM number)Consistent quantity/quality from selected healthy donor with high TSCM number
Cannot select desired T-cell phenotypes/functionsCan optimize T-cell phenotypes/functions (e.g., edit in homing and growth genes)
Urgent timelines to be met for product generation to meet the need of the individual patientPrepared and ready for patients when needed
Limited in scalabilityEase of scalability
Single cancer targetMultiple cancer targets (multiple genes edited)
Increased treatment costs (quality testing and regulatory costs specific to single patient)Decreased treatment costs (quality testing and regulatory costs spread over many patients)

Issues with allogeneic CAR T therapy: overcoming patient rejection

While allogeneic CAR T therapy helps to address some of the issues encountered with autologous approaches, it still faces serious hurdles. Most importantly, allogeneic CAR T therapies can cause serious life-threatening reactions arising from patient rejection—where the patient’s own immune system recognizes the donor cells as foreign [6,7]. Figure 3 illustrates the basic biology behind patient rejection of allogeneic T cells. This rejection is driven by the interaction of human leukocyte antigen class I (HLA I) and T cell receptors (TCR) that are expressed on both the donor’s and patient’s T cells and can lead to three rejection scenarios:

  • Graft-versus-host disease (GvHD)
  • Host-versus-graft-disease (HvGD)
  • A patient’s NK cells attacking allogeneic CAR+ T cells, with masked HLA (a modification strategy used to avoid the first two scenarios)
Human leukocyte antigen class I (HLA I) and T cell receptor (TCR) on donor and recipient cells can lead to causes of patient rejection in cell therapies.

Figure 3. Biology of HLA I and TCR interaction and patient rejection. The interaction between the TCR and HLA I on both donor and patient T-cells drives two rejection pathways. In graft vs host disease, the donor TCR recognizes the allogeneic peptide/HLA I of the host T cell, resulting in rejection (i.e., killing) of the host cell. In host vs graft disease, the opposite occurs—the host TCR recognizes the donor HLA I as foreign and targets it for killing.

TCR is a membrane-bound protein consisting of α and β chains and is expressed as part of the CD3 complex molecule on the surface of all T cells (Figure 3). The surface-displayed HLA class I molecules appear ubiquitously on cells throughout the body and consist of an α chain that is stabilized by β2-microglobulin (β2M, Figure 3).

HLA I molecules are made up of a group of six genes which are designated as A, B, C, E, F, and G. These genes are further divided based on their polymorphism. The genes A, B, and C are highly polymorphic, with over 6,000 alleles represented in each one of them. The non-polymorphic genes are E, F, and G, with allelic variants of less than 300 [8,9].

The mechanism of rejection is initiated by the recognition and interaction of the TCR αβ chains of T cells with the HLA class I molecules (Figure 4). In graft-versus-host disease, the allogeneic rejection arises when the TCR on donor T cells regards the HLA I complex on the recipient’s cells/tissues as foreign and attacks them [8]. Similarly, in host-versus-graft disease, the TCR of the recipient’s T cells recognizes the HLA complex on donor T cells as foreign and attack them.

To remove these rejection barriers, scientists can exploit the fundamental biology of the TCR and the HLA class I complex. More specifically, the disruption of β2M through gene editing can be used to prevent mature donor HLA class I molecules from reaching the cell surface, essentially shielding the donor T cells from recipient T cell recognition and elimination. Similarly, disruption of the donor cell TCR α or/and β chains through gene editing can prevent the recognition and attack of donor T cells by the recipient T cells.

Figure 4. Off-the shelf strategy to prevent allogeneic rejection responses. Right side: Elimination of the TCRα chain on the grafted allogeneic CAR T-cells prevents graft vs host disease (GvHD). Left side: Elimination of HLA I through disruption of β2M on the grafted allogeneic CAR T-cells prevent host vs graft disease (HvGD).

These approaches, however, can lead to a third rejection barrier caused by the loss of the HLA class I on the donor cells, rendering the donor cells susceptible to targeting by the recipient’s own NK cells (Figure 5), also known as the “missing self signal” [10,11]. To overcome recipient NK cell-mediated elimination of HLA/TCR donor cells, researchers can genetically modify the donor cell to express an inhibitory molecule such as non-polymorphic HLA-E (Figure 5) [12,13]. This modification can be performed by inserting or “knocking-in” the sequence to HLA-E fused with β2M. This step leads to the stable expression of a type of HLA class I molecule on the donor cells and prevents recipient NK-mediated killing of those cells (Figure 5).

Figure 5. Addition of HLA-E on donor T cells will prevent host NK killing. An allogeneic CAR T cell that lacks the HLA I protein would be a target for the natural killer (NK) cells (also known as the “missing self signal”). This is a normal immune reaction to cells that do not express HLA I. Engineering cells with a knock-in HLA-E gene will prevent the killing of the allogeneic CAR T cells by the host NK cells.


The choice of allogeneic T cells sources allows for the use of material that features a higher quality starting blood from healthy third-party donors and improved immunological cell makeup which can improve scalability of the final product. It also provides an approach to treat multiple patients unlike an autologous approach. The use of allogeneic T cell sources can lead to patient rejection outcomes (e.g., GvHD and HvGD), and technical advances are being utilized to mitigate some of these issues through the use of gene editing of HLA class I and/or TCR genes to overcome the “foreignness” seen in the host system. Clinical successes provide evidence that allogeneic CAR T cell therapies employing some of these masking techniques enable use of this approach for a wider number of patients and fueling the growth in this space with multiple allogeneic CAR focused companies already testing allogeneic CAR T therapy in the clinic (Table 2).

Table 2. Allogeneic CAR T companies with clinical trials.

CompaniesProductsCAR targetsAllogeneic cell sources
Allogene and PfizerUCART19CD19T cells
Kuur TherapeuticsKUR-502CD19NK, T cells
Cellectis and PfizerUCART19, UCART123CD19, CD123T cells
CelyadCYAD-211BCMAT cells
CRISPR TherapeuticsCTX110CD19T cells
Fate TherapeuticsFT819CD19iPSC-derived T cells
Poseida TherapeuticsP-BCMA-ALL01, P-MUC1-ALLO1BCMA, MUC1T cells
Precision BiosciencesPBCAR269ACD19T cells
Tessa TherapeuticsCD30.CAR-EBVSTCD30EBV T cells