To better understand the state of immune cells during the process of CAR T cell therapy, it is important to note that at the time of leukapheresis, the patient’s T lymphocytes have usually already undergone at least two lines of other cancer therapies (9). These cells, influenced by the previous lines of immunosuppressive medication and the inhibitory tumor microenvironment (persistent antigen stimulation, inhibitory signaling, hypoxia, acidosis, etc.), often enter the CAR T production process already exhausted, terminally differentiated, and with impaired mitochondrial function (54, 55). During the production process, the cells are activated, genetically modified, proliferated, kept, and stored in in vitro conditions (56, 57). Upon infusion into the patient, it is desired that the CAR T cells further clonally expand, migrate rapidly to the tumor site, recognize, and efficiently kill tumor cells, with each CAR T cell eliminating as many tumor cells as possible (58, 59). Since all of these processes are extremely energy consuming, adequate energy production and cellular energy metabolism are crucial for an effective and successful therapy course. In this context, mitochondria play a key role as cellular organelles, responsible for energy production and metabolism (60, 61), constantly adapting to environmental stimuli and the energy demands of the cell. A simplified schematization of mitochondrial dynamics during different phases of CAR T cell therapy is presented in Figure 1 .

For successful therapeutic outcomes at each phase of the process, it is imperative that mitochondrial function remains robust and demonstrates rapid adaptability to alterations in the cellular milieu and metabolic demands. Five main groups of mitochondrial processes and their potential impact on CAR T cell therapy are further discussed. These are metabolic reprogramming, mitochondrial mass and biogenesis, mitochondrial membrane potential, production and neutralization of reactive oxygen species (ROS), and mitophagy.

Metabolic reprogramming in T lymphocytes refers to the shift in cellular metabolic pathways in response to changes in cellular energy requirements. The primary cellular metabolism in naive, non-activated T cells is oxidative phosphorylation, in which ATP is generated by the transfer of electrons through the electron transport chain at the inner mitochondrial membrane, producing few toxic byproducts and efficiently utilizing glucose (62, 63). However, when cells’ energy demands increase (e.g., during activation, proliferation, cytotoxic activity, or other complex cellular processes), cells shift their metabolism toward glycolysis (64). The latter produces ATP molecules faster, but less efficiently and with the production of toxic byproducts, such as excessive lactate, which can lead to acidification of the cellular environment and loss of cellular functions (65). In addition to glucose metabolism, other catabolic pathways, such as efficient fatty acid oxidation, play critical roles in T cell development, central memory differentiation, cell survival, and long-term remission (66). While shifts in metabolic pathways in healthy cells occur regularly in response to stimuli for altered energy demands, it has been shown that the most effective CAR T cells possess a balanced metabolic profile and are characterized by the ability to quickly shift from one metabolic type to another and vice versa. Inefficient metabolic shifts can result in prolonged glycolysis, inefficient energy production, and consequently ineffective and short-lived CAR T cells (60, 67–69).

Adequate mitochondrial mass is another critical factor defining cellular energy production capacity (70). Along with the increased energy demands and metabolic switch to glycolysis in T lymphocytes or CAR T cells, these cells enhance their mitochondrial biogenesis, resulting in elevated number of mitochondria per cell, and increased mitochondrial size and mass to increase the energy production capacity (71). For the CAR T production process, it is desired that the input T cells have intact mitochondrial function and high mitochondrial biomass (67). After T cell selection, the cells first undergo activation characterized by mitochondrial fission and multiplication. This leads to the formation of punctate mitochondria with loose cristae, reducing the efficiency of oxidative phosphorylation and triggering the initiation of glycolytic metabolism characteristic of effector T cells (60, 71). After effector function, a small proportion of T lymphocytes transform into a memory phenotype with large, elongated mitochondria. These mitochondria possess a high capacity for energy production, which enables them to maintain oxidative phosphorylation and allows the cells to persist in the organism for prolonged time periods (60, 70). However, most effector T lymphocytes become exhausted, the mitochondria disintegrate, shrink, and all metabolic types vanish, leading to cell apoptosis (71). In the context of predicting CAR T therapy progression, Rostamian et al. (60) found that impaired mitochondrial function with low mitochondrial biomass prior to infusion of CAR T cell product leads to poor therapeutic outcomes.

The mitochondrial membrane potential (ΔΨm), generated by pumping protons from the mitochondrial matrix into the intermembrane space, is another indicator of mitochondrial function and the antitumor efficacy of CAR T cells and is crucial for efficient ATP synthesis (72). High mitochondrial membrane potential characterizes the effector phenotype of T lymphocytes, along with increased glycolysis, ROS production, and cellular impairment. In contrast, low mitochondrial membrane potential is characteristic of naive and memory T lymphocytes and is favored in input cells in the CAR T production process for better energy production capacity of the final product (60, 68). In terms of cytotoxic T lymphocytes, lower ΔΨm levels are desirable, as they indicate a better metabolic capacity of the cells, less exhaustion, correspondingly low glycolysis levels, better persistence in vivo, better migratory capacity, and antitumor efficacy (68).

Reactive oxygen species (ROS) are oxygen-containing molecules that are mainly generated in the mitochondria (73). Due to their instability, they react rapidly, causing cellular defects at the DNA, RNA, or cellular structure levels, and can even induce cell death (61). Small amounts of ROS are continuously produced and act as signaling molecules, which are then neutralized by cellular antioxidant mechanisms (74). However, under pathological conditions (e.g., cancer) and in exhausted cells, ROS concentrations can greatly increase (71, 75) and damage cellular structures to the point of irreparability (71), impair T cells function (76), and induce T cells senescence (77). Elevated ROS concentrations and impaired antioxidant mechanisms for ROS neutralization in T lymphocytes and CAR T cells prior to infusion of CAR T cell product are indicative of a poor prognosis for effector cell function upon infusion into the patient (60).

Mitophagy is a multistep process that involves recognition of damaged or dysfunctional mitochondria, their uptake into autophagosomes, and subsequent degradation by fusion with lysosomes (78). The process is tightly regulated at multiple levels, including activation of specific mitophagy receptors, recruitment of autophagic machinery components, and coordination of autophagosome-lysosome fusion (79). Mitophagy is essential for proper mitochondrial function in CAR T cells, as it helps to prevent the accumulation of damaged mitochondria that otherwise accumulate excessive amounts of ROS and impair energy production throughout the CAR T production process as well as the therapy course (67, 80).

Mitochondria therefore hold great potential as therapeutic targets to aid the antitumor therapies and as predictive biomarkers for assessing the therapy course prior to CAR T cell infusion (60, 61, 64, 67, 75, 80).

As research continues to illuminate the dynamic role of mitochondria in CAR T cell therapy, understanding and monitoring mitochondrial processes may lead to more effective therapeutic outcomes. Methods to assess mitochondrial function can be categorized at the genomic, transcriptomic, proteomic, and metabolomic levels (81). Primary mitochondrial genetic disorders arise from cellular or mitochondrial pathological mutations (81) that can be identified by genome sequencing analyses (82). At the transcriptomic level, gene expression can be assessed using techniques such as RNA sequencing, polymerase chain reaction (PCR), Northern blotting, microarrays, and many others (83). Epigenetic regulation and post-translational modifications also play an important role in modulating mitochondrial dynamics (84). A variety of techniques are available for proteomic analysis. For example, fluorescently labeled dyes can be used to stain target molecules, allowing determination of their concentration, localization, and dynamics. Such measurements can be performed with fluorescence microscopy (i.e., flow cytometry) and allow visualization and quantification of mitochondrial membrane potential, mass, and other parameters (85). Other common methods for analyzing protein content include Western blotting, electrophoresis, ELISA, chromatography, mass spectrometry, protein microarrays, etc. (81, 83, 84) The Seahorse analyzer is an excellent tool for determining the metabolic status of target cells (86). In addition, high-resolution respirometry, isotope tracking, and other methods have proven useful in this field (87). Other microscopy techniques, such as transmission electron microscopy (TEM), provide high-resolution images of mitochondria that allow direct observation of changes in mitochondrial morphology and structure (88). The analysis of mitochondrial characteristics offers an insight into the cell’s functional state, potentially serving as a biomarker for predicting cell behavior and progression during the CAR T production process. Since mitochondrial characteristics are indicative of cells’ energy capacity, apoptotic susceptibility, and cytotoxic functionality, they could be utilized to forecast the anti-tumor cytotoxicity of CAR T cells before therapy initiation. This prospective approach may allow for the early identification of therapeutic potential of CAR T cells, enhancing patient-specific treatment strategy.

The endothelium is a layer of endothelial cells that form the inner lining of blood and lymphatic vessels and play a crucial role in many bodily functions, including the regulation of inflammation, blood clotting, and the formation of new blood vessels (angiogenesis) (89). Upon infusion of CAR T cells, the infused cells migrate to the tumor site and induce apoptosis of tumor cells. In addition to the cytotoxic effect, they secrete cytokines that trigger inflammation and activation of endogenous immune cells (such as macrophages, dendritic cells, natural killer cells, B cells, etc.), fibroblasts, and endothelial cells (90–94). Activated endogenous cells also secrete proinflammatory cytokines and chemokines, which can lead to overactivation of the immune system, endothelial damage, and increased vascular permeability (5, 95). Among these inflammatory cytokines and chemokines, IL-6 is considered the critical cytokine involved in endothelial permeabilization and induction of CRS (5, 96, 97). After endothelial activation and permeabilization, activated endothelial cells also begin to secrete inflammatory signals (such as IL-6). This further leads to increased permeability of the blood-brain barrier, infiltration of inflammatory molecules and immune cells into the central nervous system, and onset of ICANS symptoms such as headache, nausea, confusion, blurred vision, delirium, coma, or even death (98, 99). IL-6 antagonists (such as tocilizumab) are used as intervention drugs to treat severe CRS and ICANS symptoms (96, 97, 100). The stages of endothelial activation and blood-brain barrier permeabilization in CAR T cell therapy leading to the occurrence of CRS and ICANS are shown in Figure 2 .

In CAR T cell therapy, endothelial activation plays a crucial role in inflammation, regulation of the immune response, and development of side effects (CRS, ICANS, etc.) (5, 101). After administration of the cell product and migration of CAR T cells to the tumor site, the antitumor immune response is initiated. At this time, it is desired that endothelial activation and vascular permeability remain low to moderate to allow for an effective immune response and inflammatory signaling without causing severe inflammation or vascular injury (5). It is important to note that increased endothelial activation and vascular permeability can lead to severe inflammation and high-grade side effects, resulting in less efficient tumor cell killing, unsuccessful therapeutic outcomes, and unmanageable development of side effects that can result in long-lasting consequences and even death (101). Not only does the endothelial activation play a crucial role after administration of the cell product and therapy progression, but studies have also shown that endothelial activation prior to CAR T cell infusion may also contribute to therapy progression and have a negative prognostic effect on CAR T therapy outcome and the occurrence of CRS and ICANS (6). There are many reasons for endothelial activation prior to CAR T cell infusion. On the one hand, it may be a consequence of previous cancer therapies (chemo- or immunotherapy) and the lymphodepleting regimen. Tumor burden with inhibitory tumor microenvironment (TME), hypoxia, and permanent antigen stimulation may also trigger endothelial activation. On the other hand, factors may be un-related to the tumor, such as other medical conditions (diabetes, hypertension, etc.), infections and inflammations, or the physiological state of the patient (obesity, physical performance, age, stress, injuries, etc.) (55, 98, 102). It is usually impossible to select a single factor, but a combination of the listed reasons typically results in excessive activation of the endothelium.

In the context of predicting the outcome of CAR T cell therapy, the Endothelial Activation and Stress Index (EASIX) score has been proposed. It is defined as (creatinine [mg/dL] × lactate dehydrogenase [LDH; U/L])/platelets [10⁹ cells/L] or modified EASIX score combined with CRP × ferritin (EASIX-FC) (42, 43, 103). The correlation between the EASIX score prior to CAR T cell infusion and the occurrence of CRS and ICANS was confirmed (44, 103). However, while EASIX can be a useful predictive tool, it does not directly measure endothelial activation. Instead, it uses surrogate biomarkers that may be influenced by various other factors and accompanying pathological conditions.

In the search for better biomarkers of endothelial activation, the candidates can be classified into three groups based on their effect on the endothelium. The first are endothelial stabilizers, which are mainly synthesized by the endothelium and released into the bloodstream. They are responsible for maintaining endothelial homeostasis and stability and are absent or under-expressed in pathological conditions with endothelial overactivation. Common examples of endothelial stabilizers are nitric oxide (NO) (104), VE-cadherin (105), antioxidant compounds such as superoxide dismutase (SOD) and catalase to combat oxidative stress and maintain endothelial stability (106), extracellular matrix (ECM) components such as collagens, laminins, fibronectins, etc., that structurally support the endothelium and are responsible for maintaining endothelial barrier function (107), and many others. The next group of biomarkers for endothelial activation are endothelial destabilizers, which can be secreted from various cell types and are typically elevated in pathological conditions such as inflammation, stress, cancer, injury, and others. Some common examples of endothelial destabilizers are inflammatory cytokines, such as tumor necrosis factor-alpha (TNF-α) and IL-6 (108, 109), ROS that can cause oxidative stress and damage to the endothelium (110, 111), matrix metalloproteinases (MMPs) that can degrade the ECM and impair endothelial structural support and barrier function (107, 108), and many others. The next important group of endothelial activation biomarkers are endothelial adhesion molecules, which are expressed on the surface of endothelial and other cells and play a crucial role in the interaction and adhesion of leukocytes, other cells, and the ECM to the endothelium (112, 113). Endothelial adhesion molecules play dual roles in stabilizing and destabilizing the endothelium and also in controlling the contact between CAR T cells and their targets (114). Under normal physiological conditions, they contribute to the maintenance of vascular integrity and homeostasis by regulating leukocyte recruitment and transendothelial migration. However, in pathological conditions such as inflammation, infection, or cancer, excessive or prolonged expression of adhesion molecules can lead to endothelial destabilization, increased vascular permeability, and leukocyte infiltration (114, 115). To highlight only a few of the important examples of adhesion molecules, Intercellular Adhesion Molecule-1 (ICAM-1), angiopoietin-2 (Ang-2), Vascular Cell Adhesion Molecule-1 (VCAM-1), and others contribute to endothelial permeabilization. In elevated concentrations, they exhibit a poor prognostic effect and may immunosuppress CAR T cells (116, 117).

It is important to note that to maintain endothelial homeostasis, a precise balance between stabilizing and destabilizing signals must be maintained. Many of the endothelium-related molecules are released into the circulation and can be easily measured from blood samples. They therefore represent a great source of potential biomarkers for predicting CAR T cell therapy outcomes and side effects susceptibility. Moreover, endothelial markers, reflecting the state of vascular health, could serve as valuable tools for predicting therapeutic outcomes even before the initiation of CAR T cell therapy process. Their role in vascular integrity and reaction to inflammatory stimuli makes them promising indicators for assessing the efficacy and potential side effects of treatments in advance.

The term central nervous system (CNS) includes the brain, spinal cord, nerves, and associated cells. The causes of CNS impairment and injury may be due to concomitant diseases and disorders (autoimmune diseases, neurodegenerative diseases, stroke, etc.), neurological diseases of exogenous origin (toxins, inflammation, infection, injury, etc.), or tumor burden and the TME. The malignancy itself can promote inflammation, tissue damage, and CNS impairment, but prior cancer therapies (chemo- or immunotherapy) and the lymphodepleting regimen may also have an impact (6, 9, 118, 119).

The connection between CNS impairment and the outcome of CAR T cell therapy is an emerging area of research. Focusing on even the smallest changes in markers of blood-brain barrier disruption and markers of neuronal and glial injury could help in predicting and monitoring the progression of ICANS (6, 119–121). Schoeberl et al. (122) observed that efficient ICANS prediction could be achieved in patients without a history of neurological disorders, while patients with accompanying neurological disorders and diseases show signs of previous and/or chronic neuronal damage and respond very heterogeneously to the treatment. Therefore, the predictive accuracy for therapy outcomes and ICANS is limited to individuals without prior neuronal injuries (122).

Biomarkers for determining CNS impairment can be monitored after cell infusion to observe disruptions in CNS homeostasis. The measured values can serve as early indicators of ICANS. However, an emerging field is the use of biomarkers of CNS impairment prior to infusion of CAR T cells. These markers reflect impaired CNS homeostasis and possible CNS injury that may later lead to the development of high-grade ICANS (118). Recent studies have shown that levels of CNS impairment markers prior to CAR T cell infusion correlate with the development of ICANS after CAR T cells administration (121, 123, 124). Several notable biomarkers of neuronal or glial injury have been identified that show considerable promise for predicting the occurrence of ICANS with CAR T cell therapy. Such examples include neurofilament light chain (NfL), a protein originally located in neurons and released into the cerebrospinal fluid (CSF) and into the bloodstream during neuronal injury (118, 122); glial fibrillary acidic protein (GFAP) which indicates astrocyte activation and astrogliosis, often associated with neuroinflammation (123, 124); S100 calcium-binding protein B (S100B), which is released by activated astrocytes and indicates CNS injury (124), and many others. These markers are secreted into the CSF upon CNS injury, but their concentrations in the CSF correlate directly with their concentrations in the blood and can therefore be easily measured from a blood sample (122, 125, 126). Furthermore, the predictive value of CNS impairment markers prior to therapy initiation is gaining attention. By assessing these markers before starting CAR T cell therapy, clinicians might better anticipate therapeutic outcomes and the risk of ICANS, enabling more tailored and proactive management strategies. This approach leverages the correlation between pre-treatment levels of CNS markers and the likelihood of subsequent ICANS, highlighting their utility in enhancing patient-specific therapeutic strategies.

The concept of “immune system markers” encompasses diverse facets of a patient’s heterogeneous immune system. Such aspects include the patient’s baseline characteristics such as age, performance status, organ function, comorbidities, immune system characteristics, immune cells function and exhaustion, and other factors. These characteristics may influence the course and outcome of CAR T cell therapy (37, 45, 47, 55, 102, 127, 128). Moreover, immune system function markers may denote cell markers that differentiate between subpopulations of immune cells and define their phenotypic characteristics (47, 129). This has notable implications for the production process of CAR T cells and their subsequent antitumor efficacy post-administration (47, 55). Therefore, an in-depth understanding of individual immune systems and immune cells characteristics could potentially pave the way for improved prediction of response to CAR T cell therapy and resulting therapeutic outcomes.

Various baseline characteristics and blood biomarkers have been identified that might predict the outcomes of CAR T cell therapy, thus highlighting their importance for therapy selection and management. Among physiological measures, parameters such as age, heart rate, body temperature, comorbidities, and blood pressure have displayed the highest predictive values (47, 128, 130). Among blood biomarkers, leukocyte count, inflammatory cytokines, hemoglobin, creatinine, CRP, ferritin, fibrinogen, and platelets have been shown to predict the development of severe CRS (29, 52, 131, 132). However, consideration of these patient characteristics alone does not provide a sufficiently specific and robust predictive model for application in a broader population. Given their high variability, which may be influenced by previous therapies, patient lifestyle, and disease burden, baseline patient characteristics should be used in conjunction with more robust biomarker systems (47, 55).

Given the nature of CAR T as a T cell therapy, T cell biomarkers are frequently being monitored throughout the process. For instance, studies have shown that a defined CD4:CD8 ratio of T lymphocytes at the time of leukapheresis (ranging from 1:1 to 3:1) is associated with a better proliferative capacity for the CAR T production process (129, 133, 134). A higher CD45RA : CD45RO ratio at the time of leukapheresis indicates an increased proportion of naive, less differentiated T cells correlated with improved proliferative capacity and therapeutic outcome (67). CAR T cell subsets can be distinguished as naive T cells (CD45RO−/CD62L+/CD27+), central memory T cells (CD45RO+/CD62L+/CD27+), effector memory T cells (CD45RO+/CD62L−/CD27−), and effector T cells (CD45RO+/CD62L−/CD27−). Activated CAR T cells express activation markers such as CD25, CD69, and CD137 (94, 135, 136). Furthermore, studies indicate that higher levels of central memory T cells (Tcm) and lower levels of effector T cells (Tef) in the infusion product are associated with improved therapeutic outcomes (50, 94, 129, 136, 137). Elevated levels of exhausted and senescent (CD57+) T cells in the infusion product correlate with poor therapy progression (94, 138). To achieve long-term remission, it is therefore advantageous to have memory CAR T cells that persist over time and provide an efficient antitumor response in the event of relapse (139). New phenotyping biomarkers with higher predictive capacity for CAR T therapy progression are being extensively investigated.

Another significant component of biomarkers of the immune system pertains to the antitumor cytotoxic activity of the CAR T cells. The mechanisms entailing migration, tumor cell recognition, cytokine release, and target cell killing are highly complex and play key roles in successful antitumor efficacy of CAR T cells after administration of the cell product into the patient (29, 114, 140, 141). The specific steps of the cytotoxic function are shown in Figure 3 . Any malfunctions within these cytotoxic mechanisms can lead to unsuccessful therapy and severe inflammation. Such malfunctions may stem from relatively rare genetic disorders or from T cells dysfunction, which may be a consequence of disease burden, patient characteristics, and exhausted and senescent T cell phenotypes prior to the CAR T production process (29, 55, 102, 142). The malfunctions may also arise during the production process, as cells respond differently to ex vivo manipulation due to their individual characteristics (55). The vector encoding the CAR receptor is integrated semi-randomly into the genome, leading to variable expression and consequently variable efficacy of the CAR T cells. The receptors can also be expressed constitutively for extended periods or inductively for a brief duration (143, 144).

Furthermore, the integration of transgenes via viral vectors raises concerns about potential risks, including insertional oncogenesis, gene inactivation or dysregulation, and impairment of cell functions (144). Early detection of such integration events is crucial for ensuring the safety of CAR T cell therapies. Potential biomarkers, such as abnormal gene expression levels, novel fusion transcripts, epigenetic changes, etc., and assays, such as linear amplification-mediated PCR (LAM-PCR), high-throughput sequencing (i.e. integration site sequencing), whole-genome sequencing (WGS), etc. could serve to identify transgene integration sites (145). Advanced bioinformatics tools are further used to analyze the data to assess the potential impact of transgene integration on gene expression (145). Monitoring these integration events could provide insights into the safety profile of CAR T cell products and help mitigate risks associated with gene therapy.

After infusion of the CAR T cell product, the modified cells first migrate to the tumor site. The migration and infiltration into tumor tissue are the main obstacles of CAR T efficiency in solid tumors (146) but also play an important role in hematologic malignancies (99, 115). Adequate expression of adhesion molecules (e.g., LFA-1, VCAM-1, ICAM-1, VEGFA, and others), chemokines (e.g., CCL3, CCL4, CXCL9, CXCL10, CXCL11, CXCL12, and others), and other guidance molecules are of paramount importance for effective CAR T cell homing (114, 147–153). Therefore, inadequate expression of these navigation-related molecules could serve as a negative predictive factor for the progression and outcome of CAR T cell therapy (114, 149, 150). For successful effector functions, it is imperative for CAR T cells to rapidly recognize tumor cells and facilitate CAR receptor binding with the CAR antigen (e.g., CD19, CD20, CD22, BCMA, etc.) expressed on tumor cells (154). Recent clinical evidence indicates that antigen downregulation and escape have arisen as major obstacles that affect the overall efficacy, success rate, and long-term remission after CAR T cell therapy (155). Therefore, sufficient antigen expression on tumor cells may serve as a prognostic tool for therapy outcomes and may even influence patient eligibility for CAR T cell therapy. Upon recognition of target cells, CAR T cells trigger a series of cytotoxic reactions aimed at inducing target cells apoptosis.

In the event of abnormalities within cytotoxic mechanisms or prolonged duration of the immunological synapses leading to an extended effector function timeframe, this could escalate the release of cytokines and chemokines, thus increasing inflammation, compromising the efficacy of the therapeutic response, and potentially causing a relapse of antigen-free malignancy (29, 140, 142, 147, 156–158). Methods for ex vivo examination of cytotoxic efficiency of CAR T cells are described in Table 2 . A careful examination of cytotoxic mechanisms of T cells prior to CAR T cell infusion may provide insight into potential defects and serve as an initial indication of therapy prognosis and the probability of severe inflammation occurrence. This information is crucial for predicting the course of CAR T cell therapy prior to cell infusion.

Extracellular vesicles (EVs) are small membrane-derived particles that are released by cells into the extracellular space and can be transported throughout the body. These vesicles play an important role in cell-to-cell communication and transport a variety of biological molecules from their cell of origin to target cells (160, 161). Because they are derived from parent cells, the EVs carry markers of parent cell that allow the origin of the vesicles and their contents to be determined (162). By analyzing the vesicles content, cellular signaling can be monitored, providing insight into cell-to-cell communications (163). Their usual cargo is proteins, lipids, DNA, messenger RNAs (mRNAs), microRNAs (miRNAs), and other molecules (164).

Because EVs are involved in many physiological and pathological processes, their content provides valuable insights into the signaling of specific cell populations. For example, they can mediate immune responses, facilitate blood clotting, and contribute to the spread of cancer (160, 165, 166). In the case of CAR T immunotherapy, this could prove useful in assessing the immune system status, immune cell exhaustion and functionality, and antitumor response. On the other hand, by studying tumor cell-derived extracellular vesicles (oncosomes), the information on tumor invasiveness, antigen escape, and inhibitory signaling toward cells of the immune system, including CAR T cells, could be better understood (163, 167). EVs were shown to exhibit an effect on CAR T cells (168, 169). Due to their ability to transport molecules from one cell to another, EVs are being extensively studied for their potential use as drug delivery systems and as biomarkers for disease prognosis and immunotherapy progression (170–172).

To discuss some examples of EVs that could potentially predict response to CAR T cell therapy and the development of high-grade CRS and ICANS, the origin of the vesicles must be taken into consideration. First, the vesicles can be derived from endogenous immune cells. They can exhibit stimulatory or inhibitory functions toward CAR T cells and therapy response (173, 174). For example, an increased number of CD69 positive T cell vesicles can indicate increased T cell activation and act as a negative feedback loop that inhibits further T cell activation (175). Increased numbers of T cell EVs expressing inhibitory molecules such as PD-1, CTLA-4, TIM-3, LAG-3, and others reflect an ineffective and exhausted immune system and could consequently be used to predict poor response to therapy (169, 174, 176).

Possible sources of EVs are also CAR T cells. Studies have shown that persistent concentrations of CAR-positive EVs in the bloodstream of patients after CAR T cell infusion exhibit predictive impact on long-lasting remission (177). Evidence also suggests that CAR-positive EVs assist the antitumor function of CAR T cells by overcoming obstacles and barriers that otherwise limit the effect of the immunotherapy (178–180). The next example is increased levels of endothelial vesicles and apoptotic bodies, which indicate excessive endothelial activation and damage, which may predict the development of severe CRS and ICANS even before infusion of the cell product (170). Tumor-derived vesicles often express inhibitory molecules and reduce the antitumor effect of CAR T cells (169). Elevated levels of circulating tumor DNA correlate with poor therapeutic efficacy and higher CRS levels (181), which can also be applied to circulating oncosomes containing tumor DNA. CD19+ vesicles were shown to cause activation and exhaustion of CAR T cells with decreased antitumor activity and trigger CRS (175).

Stated here are just some common examples of EVs and their potential impact on immunotherapy. The EVs show great potential for predicting the immune response to CAR T cell therapy. However, this field is relatively young and poses many challenges. The first is the development of standardized and optimized extraction procedures for isolation of heterogeneous vesicles from cancer patients’ samples. Because vesicles vary in density, structure, and size, robust isolation techniques with minimal sample loss need to be established (165). Another challenge currently being investigated by many research groups is the development of biomarkers to efficiently differentiate B-cell leukemia or lymphoma from other types of vesicles. Some studies suggest examples such as CD5, CD19, CD31, CD44, CD55, CD62L, CD82, and CD123 (169, 182, 183). Further research on this topic is needed to develop efficient biomarkers and predictive models. EVs could serve as predictors before CAR T cell infusion to provide an impression of cellular signaling and information circulating in the patient’s bloodstream.

The tumor microenvironment (TME) is a complex mixture of various components, including different cell types, signaling molecules, and extracellular matrix components. The TME can contribute to tumor growth, progression, and inhibition of the antitumor immune responses (184, 185). Consequently, the TME components may undermine the cytotoxic potency of CAR T cells, thereby limiting the efficacy of CAR T cell therapy. Accordingly, assessment of individual patient TME characteristics and the inhibitory properties of the TME components on CAR T cells prior to CAR T cell infusion may serve as predictive parameters for determining the potential extent of CAR T cell effector function inhibition after infusion. This could allow the prediction of inflammation development associated with immune cell inhibition.

The TME constituents are divided into six main categories based on their composition and function. First are immunosuppressive cells, which include regulatory T cells (Tregs), myeloid-derived suppressor cells (MDSCs), tumor-associated macrophages (TAMs), and other (113, 186–189). The TME immunosuppressive cells can inhibit the antitumor response in several ways: by inducing anergy, exhaustion, or even apoptosis of T cells (190, 191), by inducing expression of immunosuppressive cell markers (such as immune checkpoint molecules) (187, 189, 192), by signaling proliferation and recruitment of other immunosuppressive cells (189, 192), by altering antigen presentation, which impairs recognition of the tumor by the immune system (187), by altering metabolic pathways to deplete energy sources and produce toxic metabolites (190), and by secretion of immunosuppressive molecules such as cytokines, chemokines, and others (187, 188, 192). The TME can induce expression of inhibitory immune checkpoints, typically expressed on various immune cell types. Normally, these molecules regulate and control the immune response to prevent over-activation. However, in the context of cancer, tumor cells can exploit their mechanisms to downregulate the immune system response, thereby facilitating evasion of the immune system. Prominent examples of inhibitory immune checkpoints include PD-1, CTLA-4, TIM-3, LAG-3, etc. (193, 194) Moreover, these inhibitory checkpoint molecules can hinder the activation and functionality of CAR T cells after infusion (193), and their quantitative expression serves as an estimate of their inhibitory effect on CAR T cells (195). Albeit more common in solid tumors, hypoxia may also manifest in the bone marrow microenvironment and contribute to immunosuppression in hematologic malignancies (187, 190). Hypoxic conditions can induce accumulation of immunosuppressive cells and molecules, inhibit the effector function of T cells, and promote immune evasion by tumor cells (113, 196). Tumor cells also expedite metabolic pathways to produce sufficient energy for tumor growth, depriving tissues of nutrients such as glucose, glutamine, amino acids, O₂, etc. (113, 190, 197) This increased metabolic activity generates toxic or acidic metabolic byproducts (such as lactate) (112, 190) that contribute to acidification of the tumor microenvironment and subsequently suppress the immune response and CAR T cell function (112, 113, 196). The dysregulation of metabolic pathways and the imbalance of metabolites can result in the production of ROS, causing further damage to immune cells and tissues, inhibiting the antitumor effect of T lymphocytes and CAR T cells, and promoting tumor growth (113, 191). Another significant impact of the TME is the degradation and alteration of the ECM by degradative enzymes secreted by tumor cells (e.g., metalloproteinases, collagenases, oxidases) (192, 198). Degradation and alteration of the ECM can lead to impaired tissue integrity, accumulation of metabolic byproducts, and promotion of tumor spread and growth (112, 199). The main components of the inhibitory TME associated with hematologic malignancies are shown in Figure 4 .

Considering the factors described above, the characteristics of the TME are increasingly recognized as potential predictive biomarkers for CAR T cell therapy progression even before infusion of the cell product into the patient. For instance, the presence of specific immunosuppressive cell types (such as CD4+/CD25+/FOXP3+ Tregs) (99, 195), the expression of certain inhibitory molecules (e.g., PDL-1, TGF-β, IL-10, ROS, etc.) (55, 112, 200), and the overall metabolic state of the TME (e.g., lactate, LDH, etc.) (102, 112, 195) could provide insight into the ability of CAR T cells to function effectively after infusion. A thorough understanding of the interplay between CAR T cells and the TME will also aid in development of strategies to overcome the inhibitory environment. Approaches such as co-administration of immune checkpoint inhibitors, supplementation of cytokines, or genetic modification of CAR T cells to resist the immunosuppressive environment are currently being investigated to increase the efficacy of CAR T cell therapy (112, 174, 195, 196). Given the dynamic nature of the tumor immunological microenvironment, which exhibits variations over time, it is of paramount importance to personalize and monitor immunotherapies to maximize the therapeutic efficacy (196). An overview of common techniques for ex vivo biomarker analysis of the discussed biomarkers for predicting treatment response and side effects in CAR T cell therapy is presented in Table 3 .