Chimeric antigen receptor (CAR) T cells emerged as a new way of treating leukemia, and they are now being evaluated in at least 500 clinical trials for their efficacy against a variety of liquid and solid cancers. Although CAR T cells show promise as cancer therapeutics, they still pose clinical challenges. Perhaps the greatest challenge is maintaining effective control of the potency of CAR T cells after they have been introduced to patients.

A safer CAR T-cell therapy

One of the biggest concerns of CAR T-cell therapy is safety. The introduction of hyperactivated CAR T cells significantly increases the levels of inflammatory cytokines, with IL-6 being one that has been most extensively studied due to its strong correlation with the deadly cytokine release syndrome (CRS) and neurotoxicity.1 It was once assumed that CAR T cells are responsible for spike in IL-6 concentration, but recent literature describing results from in vitro models and mouse models suggest that increased levels of IL-6 may be a result of cross talk between hyperactivated CAR T cells and myeloid cells such as macrophages.2

To exploit the therapeutic efficacy of CAR T cells while controlling their toxicity, drugs such as siltuximab, a monoclonal antibody against IL-6, and toxilizumab, an inhibitor of IL-6 cell membrane receptor, are typically administered to patients for relieving CRS symptoms.3 As many of these drugs are unable to cross the blood-brain barrier, corticosteroids are used instead as an alternative to reduce neurotoxicity or CAR T cell–related encephalopathy syndrome (CRES).

CAR T-cell therapy also suffers from “on-target/off-tumor” toxicity as nonpathogenic healthy cells and tissues tend to share similar, albeit lower, antigen expressions as cancer cell and tissues. To regulate the biological activities of CAR T cells, researchers have programmed CAR T cells to express known cell surface antigens such as CD20 and epidermal growth factor receptor (EGFR) with subsequent cell death triggered by binding to monoclonal antibodies such as rituximab for CD20 and cetuximab for EGFR.4 This approach, which is known as the “suicide gene” strategy, has achieved notable clinical successes.

Noninvasive modulation of CAR T cells

Although drugs to control CRS, CRES, and off-target toxicity have proven clinically useful, they need to be improved. The drugs that are currently administered systematically offer poor temporal and spatial control, and they may not confer maximal protection at the right dosages, administration times, and target sites.

“CAR T-cell therapy is becoming a paradigm-shifting therapeutic approach for cancer treatment,” says Yingxiao Wang, PhD, professor of bioengineering, University of California, San Diego. “However, major challenges remain before CAR-based immunotherapy can become widely adopted.

“Given the complicated physiological system and diversity of individual tumors, it remains a main difficulty to identify ideal antigens for currently available approaches. Therefore, there is an urgent need for high-precision control of CAR T cells to confine their activation to local tissue space at specific time periods.”

Capitalizing on advances in remote cell control and gene delivery technologies, scientists are developing techniques to remotely regulate the biological functions of CAR T cells.5 Several of these techniques make use of genetic switches that may be induced with light, ultrasound, or heat.

Optogenetic control

A team led by Shu Chien, PhD, and Yingxiao Wang, PhD, both from the University of California, San Diego, engineered light-controllable CAR T cells for cancer immunotherapy.6 Their system has a distinct advantage over the suicide gene approach. Unlike the suicide gene approach, which deactivates CAR T cells permanently, their system can activate CAR T cell genes reversibly, enhancing the CAR T cells’ therapeutic efficacy and safety profiles.

Their technique, which is named light-inducible nuclear translocation and dimerization (LINTAD), makes use of blue light to induce the translocation of LexA DNA binding domain-CRY2-Bipartite light-inducible nuclear localization signal (LCB) into the nucleus. Blue light stimulation also induces binding of CIB1-VPR transcriptional activator (CV) with LCB, with subsequent targeting of VPR to the promoter to trigger reporter gene expression.

When primary T cells were transfected with LINTAD regulators and the CD19 CAR reporter, CAR T cells demonstrated 7.3-fold higher in vitro cytotoxicity against target leukemic B cells than control cells. Next, the team demonstrated that blue light could penetrate skin tissues and activate LINTAD gene induction in vivo. However, this system was effective only when the concentrations of both LCB and CV peaked, suggesting that in vivo binding between the two complexes is weak. To overcome this challenge, other genetic controls were introduced into the LCB-CV complex to achieve better light-inducible tumor-killing efficacy and safety in vivo.

Acoustogenetic control

Although optogenetics can modulate CAR T-cell functions, this technique is limited if it relies on light that has poor penetration through deep tissues. Recognizing this problem, the technique’s developers decided to develop an alternative technique. Chien and Wang and their colleagues investigated the use of focused ultrasound to generate heat for remote regulation of CAR T cells.7

A 15-min application of ultrasound stimulation, the scientists found, could flip a genetic switch involving a heat shock promoter that is activated at 43°C. In the “on” position, the switch induced expression of CD19 CAR in 29% of T cells but did so in only 1.9% control cells. When cultured with target leukemic B cells, the ultrasound-stimulated CAR T cells eliminated up to 83% of targeted cancer cells.

For in vivo testing, magnetic resonance imaging (MRI) was used to guide focused ultrasound waves and deliver thermal energy to confined spaces. A pulsed stimulation was used as it was less toxic than a continuous stimulation for the same duration. Human PC3 cells expressing prostate-membrane-specific antigens (PMSAs) were first injected at both prostates of mice.

Five days later after tumors developed, engineered CAR T cells targeting PMSAs were introduced at both tumors, but only the left prostate was stimulated with ultrasound. As expected, there was more significant inhibition of tumor growth at the left prostate, indicating the promise of acoustogenetics for remote control of CAR T-cell therapy to treat solid tumors safely and locally.

Thermogenetic control

At the Georgia Institute of Technology, researchers led by Gabriel Kwong, PhD, similarly made use of heat shock elements to genetically regulate the expressions of CAR proteins.8 However, as heat shock proteins can be activated not just by heat but also nonspecific stimuli such as hypoxia and mechanical stresses, the researchers carefully tested different plasmid constructs and improved the heat selectivity of their genetic constructs by incorporating more heat shock elements.

Kwong and colleagues also chose to use plasmonic gold nanorods to convert near-infrared (NIR) light into heat because these nanomaterials have been found to preferentially aggregate in tumors. NIR was found to thermally regulate CAR-T cells’ production of IL-15 superagonist, which is known to enhanced activation and cytotoxicity of CD8+ T cells.9

Although IL-15-producing CAR T cells reduced tumor burden and improved survival, more than 70% of the animals had to be euthanized after a month. In contrast, with the incorporation of NIR stimulation, no animal reached euthanasia criteria within the time window of the study. The authors concluded that thermal regulation of CAR T cell production of an IL-15 superagonist could significantly improve safety and therapeutic outcomes.

“In the future, we envision that thermal control of CAR-T cells may be used to treat primary cancers noninvasively or manage disseminated disease such as liver, lung, or brain metastases locally. This would require integrating our technology with platforms such as high intensity focused ultrasound to heat sites that are deep within the body and out of reach of photothermal targeting,” adds Kwong.

Better thermal regulations

Wanting to improve the sensitivity of the thermal bioswitches used to modulate CAR T cells, a research team led by Mikhail Shapiro, PhD, professor of chemical engineering at the California Institute of Technology, screened a total of 13 heat shock promoters (HSPs), including naturally occurring and rationally modified HSPs from humans, mice, and roundworms (Caenorhabditis elegans).10 The HSPs were found to have different baseline expressions and sensitivities to temperature change.

To facilitate sustained and amplified thermal activation, the team next designed a genetic circuit where a drug, doxycycline, was used to induce stronger transcriptional activation. In a proof-of-concept experiment, the team showed that in the absence of doxycycline, there was low production of IL-21, but that in the presence of doxycycline, there was sustained production of IL-21 for five days. The scientists concluded that their strategy of using doxycycline to chemically modulate CAR-T cells’ activities could enhance the safety of CAR T-cell therapy.

“With HSP-based gene circuits, a brief [1 hour] thermal stimulus can provide sustained [over several days] T-cell activation, leading to expression of CAR proteins and cytokines to promote local immunotherapeutic activity of T cells,” says Shapiro. “However, CAR T cells can also become activated by engaging their target, turning on HSPs even in the absence of heat shock and compromising the specificity of thermal remote control. In our study, we turned this it into an advantage by building it into a positive feedback loop. We plan to develop thermal control mechanism independent of HSPs.”

Open questions and outlook

Adverse physiological reactions attributed to CAR T-cell therapy are undesirable. Remote functional control of CAR T cells to enhance therapeutic and safety outcomes is therefore crucial for more patients to benefit from this promising therapy. Nevertheless, there are a few challenges that these emerging techniques must resolve before they become clinically useful.

One challenge is to establish practicality. For example, genetic switches regulated by noninvasive stimuli, such as light and heat, have been shown to work in preclinical research, in small animals, but they may be difficult to implement in humans. Genetic switches that respond to light do so via optoelectronic devices, the implantation of which requires invasive surgery. Genetic switches that respond to heat do so via focused ultrasound or nanomaterials sensitive to tissue-penetrating light. The delivery of thermal energy, however, requires expensive and customized equipment to provide strong enough stimuli for an adult patient.

“At the current stage, the efficiency of focused ultrasound (FUS) CAR T cells in targeting and penetrating solid tumors after intravenous injection is relatively low,” Wang acknowledges. “As such, intratumor injection of these FUS CAR cells is preferred but not always possible, such as when deep tissues and organs are targeted.”

Another challenge is that the papers cited in this article do not provide sufficient evidence that their techniques offer a safer CAR T-cell therapy. There was scant data on nonspecific CAR T-cell infiltration into organs such as the brain and liver, which are vulnerable to CRES and off-target toxicity, respectively. There was also no data on the blood levels of inflammatory cytokines that are associated with CRS. More comprehensive safety characterization is hence necessary.



  1. Keshavan M. Experimental cancer therapy holds great promise—but at great cost. STAT August 23, 2016.
  2. Murthy H, Iqbal M, Chavez JC, Kharfan-Dabaja. Cytokine Release Syndrome: Current Perspectives. Immunotargets Ther. 2019; 8: 43–52.
  3. Shimabukuro-Vornhagen A, Gödel P, Subklewe M, et al. Cytokine release syndrome. J Immunother. Cancer. 2018; 6: 56.
  4. Bonifanta CL, Jackson HJ, Brentjens RJ, Curran KJ. Toxicity and management in CAR T-cell therapy. Mol. Ther. Oncolytics 2016;3: 16011.
  5. Tay A. Nonviral Genetic Engineering of CAR T Cells. GEN June 23, 2020.
  6. Huang Z, Wu Y, Allen ME, et al. Engineering light-controllable CAR T cells for cancer immunotherapy. Sci. Adv. 2020; 6(8): eaay9209.
  7. Wu Y, Liu Y, Huang Z, et al. Acoustogenetic Control of CAR T Cells via Focused Ultrasound. bioRxiv February 19, 2020.
  8. Miller IC, Sun L-K, Harris AM, et al. Remote control of CAR T cell therapies by thermal targeting. bioRxiv May 7, 2020.
  9. Knudson KM, Hodge JW, Schlom J, Gameiro SR. Rationale for IL-15 Superagonists in Cancer Immunotherapy. Expert Opin. Biol. Ther. 2020; 20(7): 705–709.
  10. Abedi MH, Lee J, Piraner DI, Shapiro MG. Thermal Control of T-cell Immunotherapy. bioRxiv April 20, 2020.
Previous articleAutistic Men Have Enhanced Excitability in Social Brain Regions
Next articleNew Model May Help in Successfully Translating Data from Animal Studies to Humans