CD26 CAR-T cells have attenuated mitochondrial and glycolytic metabolic profiling

Abstract Background Multiple targets of chimeric antigen receptor T cells (CAR-T cells) are shared expressed by tumor cells and T cells, these self-antigens may stimulate CAR-T cells continuously during the expansion. Persistent exposure to antigens is considered to cause metabolic reprogramming of T cells and the metabolic profiling is critical in determining the cell fate and effector function of CAR-T cells. However, whether the stimulation of self-antigens during CAR-T cell generation could remodel the metabolic profiling is unclear. In this study, we aim to investigate the metabolic characteristics of CD26 CAR-T cells, which expressed CD26 antigens themselves. Methods The mitochondrial biogenesis of CD26 and CD19 CAR-T cells during expansion was evaluated by the mitochondrial content, mitochondrial DNA copy numbers and genes involved in mitochondrial regulation. The metabolic profiling was investigated by the ATP production, mitochondrial quality and the expression of metabolism-related genes. Furthermore, we assessed the phenotypes of CAR-T cells through memory-related markers. Results We reported that CD26 CAR-T cells had elevated mitochondrial biogenesis, ATP production and oxidative phosphorylation at early expansion stage. However, the mitochondrial biogenesis, mitochondrial quality, oxidative phosphorylation and glycolytic activity were all weakened at later expansion stage. On the contrary, CD19 CAR-T cells did not exhibit such characteristics. Conclusion CD26 CAR-T cells showed distinctive metabolic profiling during expansion that was extremely unfavorable to cell persistence and function. These findings may provide new insights for the optimization of CD26 CAR-T cells in terms of metabolism.


Introduction
T cells genetically modified to express CAR (CAR-T cells) can efficiently target tumor cells expressing specific antigens without restriction of the major histocompatibility complex [1].The remarkable efficacy of CAR-T cells in clinical trials has made two CD19 CAR-T cell products the first to obtain clinical application approval from the FDA [2].Currently, abundant CAR-T cell therapies have successively obtained FDA approval or are applied to clinical trials [3,4].Accordingly, the treatment landscape and clinical outcomes of hematological malignancies, especially B cell lymphoma and leukemia, and even solid tumors have changed considerably [5][6][7][8].
However, the lack of ideal tumor-specific antigens remains a major challenge for the application of CAR-T cells in many tumors [9].Substantial tumor antigens are also expressed in T cells, such as CD7, CD5 and CD99 [10][11][12].The shared expression of antigens in T cells and tumor cells may bring the fratricide caused by self-targeting and hinder the development of CAR-T cells, especially in T cell malignancies [13,14].Moreover, the self-antigens may stimulate CAR-T cells continually during the expansion.Importantly, activation of T cells when exposed to persistent antigens will lead to metabolic reprogramming [15,16].Cellular metabolism is considered to play an important role in T cell fate, and metabolic profiling and mitochondrial dynamics are involved in the differentiation of central memory T cells [17][18][19].For example, CAR-T cells with 4-1BB costimulatory molecules or expanded in IL-15 exhibit improved mitochondrial biogenesis and these CAR-T cells have superiority in the preservation of memory phenotypes and cell persistence [20,21].Therefore, CAR-T cells that expressed the targeted antigens may possess distinctive metabolic characteristics and cell fate due to persistent antigen exposure.
We previously designed CD26 CAR-T cells that could target chronic myeloid leukemia stem cells and T-cell lymphoma cells.CD26 was also expressed on the surface of activated T cells, and this phenomenon led to extensive fratricide of CD26 CAR-T cells [22,23].Interestingly, we found that the expression of CD26 would be downregulated to overcome fratricide when CD26 CAR-T cells were constructed with CD28 costimulatory molecules [24].Specially, we focused on whether the accompanying stimulation of self-antigens during expansion could remodel the metabolic profiling.Therefore, we explored the mitochondrial and metabolic characteristics of CD26 CAR-T cells expanded in vitro.Meanwhile, we also investigated the CD19 CAR-T cells that lack the stimulation of self-antigens.
In this study, we found that CD26 CAR-T cells exhibited enhanced mitochondrial biogenesis and ATP production at the early stage of expansion when compared with non-transfected T cells (NT cells).However, in the later stage of expansion, the mitochondrial metabolism and key glycolysis factors in CD26 CAR-T cells were dramatically downregulated.Moreover, CD19 CAR-T cells did not have the above distinct metabolic profiling during expansion.These results highlighted the importance of metabolic intervention, expansion time modulation and structural optimization in CAR-T cells that expressed self target antigens.

T cell isolation and CAR-T cell construction
After informed consent and approval by the Institutional Review Board of Tongji Medical College and the Hubei committee, peripheral blood samples were obtained from healthy adult donors.Peripheral blood mononuclear cells were separated by density gradient centrifugation using lymphocyte separation solution (TBD Science, China).Then T cells were isolated using CD3 microbeads (Miltenyi Biotech) according to the manufacturer's protocol.T cells were stimulated with T Cell Activation/Expansion kit (Miltenyi Biotech) at a ratio of 2:1 (T cell:bead) in X-VIVO-15 medium (Lonza) supplemented with 5% fetal bovine serum (Gibco) and 200 U/ml recombinant human IL-2 (PeproTech).After 48 h of stimulation, activated T cells were transduced with lentivirus encoding the CD19-CAR or CD26-CAR as previously described [24].The anti-CD19-ScFv-CD8-CD137-CD3Ζ expression cassette was contained in CD19-CAR lentiviral construct.

Flow cytometry
For the measurement of mitochondrial content, CAR-T cells were incubated with 200 nM Mito-tracker Green (Beyotime, China) working solution for 15 min at 37 °C in the dark and washed with X-VIVO-15 medium (pre-warmed to 37 °C), and they were suspended in PBS and analyzed by LSR Fortessa X-20 flow cytometry (BD Biosciences).The mitochondrial content was evaluated by the mean fluorescence intensity of Mito-tracker Green (FITC) and normalized using the control group (NT cells).The intracellular reactive oxygen species (ROS) and mitochondrial ROS were detected using dichloro-dihydro-fluorescein diacetate (DCFH-DA, Beyotime, China) and mitochondrial superoxide indicator (MitoSOX Red, Yeasen, China), respectively.The detection procedure was the same as that used for measuring mitochondrial content.The levels of ROS and mitochondrial ROS were evaluated by the mean fluorescence intensity of DCFH-DA (FITC) and MitoSOX Red (PE), respectively.The data were also normalized using the control group (NT cells).The mitochondrial membrane potential (MMP) was detected using 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethyl benzimidazolylcarbocyanine iodide probe (JC-1, Beyotime, China) according to the manufacturer's protocol.Then the level of MMP was evaluated by the ratio of the median fluorescence intensity of JC-1 aggregates (Red) to JC-1 monomers (Green) and normalized to NT cells.For the expression of GLUT1, CD45RA and CCR7, CAR-T cells were stained with GLUT1-Alexa 647, CD45RA-APC and CCR7-PE antibodies (BD Biosciences) and detected using LSR Fortessa X-20 flow cytometry.All data were analyzed by FlowJo 10.7.1 software.

Measurement of ATP content
CAR-T cells (1 × 10 5 ) were collected and washed with PBS, and the cell pellet was examined by ATP Assay Kit (Beyotime, China) according to the manufacturer's protocol.The ATP content of 1 × 10 5 CAR-T cells was calculated and normalized using the control group (NT cells).

Quantitative real-time PCR (RT-qPCR)
The DNA and RNA of CAR-T cells were extracted using the Cell Genomic DNA Extraction Kit (Tiangen, China) and Trizol reagent (Takara), respectively.Then, the RNA product was reverse transcribed to cDNA using the PrimeScript RT Master Mix (Takara).RT-qPCR reactions were performed according to the manufacturer's instructions.Primers (Tsingke, China) are described in Supplementary Table S1.For the assessment of mitochondrial DNA copy numbers, mitochondrial gene tRNALeu (UUR) was normalized to a nuclear gene, β2-microglobulin.For the mRNA expression level of candidate genes, each gene was normalized to a housekeeping gene, β-actin.

Statistical analysis
Unless otherwise noted, the results are presented as the Mean ± SD of at least 3 independent experiments and analyzed using GraphPad Prism 8 software.One-way ANOVA test was performed to calculate the significance among groups.P values below 0.05 were considered significant.

CD26 CAR-T cells demonstrated decreased mitochondrial biogenesis at the later stage of expansion
To evaluate whether the mitochondrial activity of CD19 or CD26 CAR-T cells changed with the culture time, we detected the mitochondrial content of CAR-T cells cultured for up to 28 days weekly.The results demonstrated that the mitochondrial mass of CD19 CAR-T cells was comparable to that of NT cells on days 7, 14 and 21 (Figure 1(a) and Supplemental Figure 1).However, the mitochondrial mass of CD26 CAR-T cells was significantly higher than that of NT cells on day 14.Both CAR-T cells showed lower mitochondrial mass than NT cells on day 28.Consistently, the analysis of the ratio of mitochondrial DNA (mtDNA) to nuclear DNA (nDNA) in CAR-T cells showed that mtDNA copy numbers in CD26 CAR-T cells were notably greater than that in NT cells on days 7 and 14 (Figure 1(b)), while there was no difference between CD19 CAR-T cells and NT cells.CD19 and CD26 CAR-T cells both exhibited decreased mtDNA levels on day 28.In order to further clarify the elevated mitochondria in CD26 CAR-T cells at early stage, we examined the expression of candidate genes involved in mitochondrial biogenesis.As expected, PGC-1α, the master regulator of mitochondrial biogenesis [25], increased significantly in CD26 CAR-T cells on day 14 (Figure 1(c)).The other two related genes, TFAM and MTCO1 [26], were also upregulated on day 14 (Figure 1(d,e)), but all these three genes downregulated on day 28.In contrast, there was no significant difference in the expression of these genes between CD19 CAR-T cells and NT cells at any time point.It was reported that enhanced mitochondrial biogenesis was a strong indicator of immune memory [27,28].Accordingly, we tested the expression profiles of CD45RA and CCR7 on days 7 and 14.The results revealed that the formation of central memory T cells (TCM) and stem cell memory T cells (TSCM) was strengthened in CD26 CAR-T cells on day 14 (Figure 2(a-c)), while effector memory T cells (TEM) were downregulated (Figure 2(d)).This was in line with the enhanced mitochondrial biogenesis of CD26 CAR-T cells on days 7 and 14.Effector memory T cells were also downregulated slightly in CD19 CAR-T cells, but central memory T cells and stem cell memory T cells showed little change.These findings suggested that the mitochondrial biogenesis in CD26 CAR-T cells was significantly promoted at early stage while impaired at later stage of culture.However, CD19 CAR-T cells did not have such dynamic variation.

CD26 CAR-T cells demonstrated progressively downregulated mitochondrial metabolism
Based on the characteristics of mitochondrial biogenesis in CD26 CAR-T cells, we next sought to examine their metabolic dynamics over time.We observed that the ATP content of CD26 CAR-T cells was significantly higher than that of NT cells on days 7 and 14, while the ATP content was decreased on day 28 (Figure 3(a)).This was consistent with the changes in mitochondrial biogenesis (Figure 1(a,b)).However, the ATP content of CD19 CAR-T cells was higher than that of NT cells only on day 7 but it was comparable to NT cells at other time points.Furthermore, we measured the expression level of genes involved in oxidative phosphorylation, which was the main source of ATP.The results showed that the expression level of FABP5, a molecule that mediated lipid uptake and intracellular transport [29], was higher in CD26 CAR-T cells than that in NT cells on days 7, 14 and 21 (Figure 3

(b)).
There was no significant difference on day 28.However, the expression level of FABP5 in CD19 CAR-T cells was comparable to that in NT cells at any time point.Another gene CPT1A, encoding the rate-limiting enzyme for long-chain fatty acid oxidation [30], was downregulated markedly in CD26 CAR-T cells on day 28 (Figure 3(c)).CD19 CAR-T cells also showed reduced expression of CPT1A on day 28 but without statistical significance.Thus, these findings indicated that the energy metabolism in CD26 CAR-T cells may be enhanced at early stage of culture due to the increased mitochondrial biogenesis, but gradually weakened at later stage.

CD26 CAR-T cells demonstrated decreased mitochondrial ROS and MMP at the later stage of expansion
In addition to biogenesis activity, mitochondrial quality is equally important in energy metabolism and cell function.We next measured the levels of ROS in CD19 and CD26 CAR-T cells at different time points.The results showed that the total intracellular ROS levels in CD26 CAR-T cells were all markedly higher than that in NT cells on days 7, 14, 21 and 28 (Supplemental Figure 2(a) and Figure 4(a)).However, there was no obvious difference in ROS production between CD19 CAR-T cells and NT cells at any time point.CD26 CAR-T cells also displayed greater mitochondrial ROS than NT cells on days 7 and 14 (Supplemental Figure 2(b) and Figure 4(b)).This was consistent with the elevated mitochondrial biogenesis in CD26 CAR-T cells.The amounts of mitochondrial ROS in these two CAR-T cells were both lower than that in NT cells on day 28.As excessive ROS may cause damage to mitochondria and cell functions [31], and therefore we examined the MMP to evaluate the mitochondrial state.We observed that the MMP of CD26 CAR-T cells was reduced when compared with NT cells on days 7, 14 and 28 (Supplemental Figure 3 and Figure 4(c)).However, the MMP of CD19 CAR-T cells was similar to that of NT cells.Taken together, CD26 CAR-T cells had increased mitochondrial ROS at the early stage of expansion and decreased mitochondrial ROS and MMP at the later stage of expansion.This indicated an abnormality in mitochondrial function may contribute to attenuated energy metabolism in the later period of culture.

CD26 CAR-T cells demonstrated downregulated glycolytic activity
As mentioned above, the quantity and quality of mitochondria in CD26 CAR-T cells changed dynamically along the expansion.We further sought to identify alterations in glycolytic activity.Glucose uptake was mediated by glucose transporter (GLUT), which played a leading role in glycolytic metabolism [32].Our results revealed that the expression of GLUT1 in CD26 CAR-T cells was dramatically decreased on days 14 and 21 (Figure 5(a-c)).Although the proportion of GLUT1-expressing cells in CD19 CAR-T cells was not significantly different from that in NT cells, the expression intensity of GLUT1 decreased.Consistent with this, the expression level of GLUT1 gene in CD19 and CD26 CAR-T cells was also downregulated on days 14, 21 and 28 (Figure 5(d)).This suggested that the glycolytic activity in CAR-T cells may be attenuated.We sought to confirm this by detecting the expression of (c) the mitochondrial membrane potential (mmp) in cd19 and cd26 car-t cells on days 7, 14, 21 and 28. the result was calculated by the ratio of the median fluorescence intensity of Jc-1 aggregates (red) and Jc-1 monomers (green) and normalized to nt cells.representative of 3 donors.data represented as mean ± Sd. differences between car-t cells and nt cells were shown and calculated by one-way anoVa test followed by dunnett's multiple comparisons test.
G6PD and PDK1 genes, which were involved in in glucose metabolism [33,34].The results showed that G6PD and PDK1 were significantly downregulated in CD19 and CD26 CAR-T cells on days 21 and 28 (Figure 6).In summary, these findings confirmed that the glycolytic activity of CD19 and CD26 CAR-T cells decreased gradually during the expansion.

Discussion
With the extensive development of CAR-T cell applications, the choice of targets had become increasingly diversified.Inevitably, some antigens including CD26 that were specifically expressed on tumor cells were also presented in CAR-T cells.Such targets would provide antigen stimulation signals in the production of CAR-T cells, which may cause remodeling of mitochondrial biogenesis and metabolism profiling.Here, we observed that CD26 CAR-T cells demonstrated elevated mitochondrial biogenesis at the early expansion stage.However, the mitochondria were depressed at the later stage as evidenced by decreased mitochondrial mass and decreased mtDNA copy numbers on day 28.In addition, we identified that the expression of crucial factors involved in oxidative phosphorylation and glycolysis were downregulated in CD26 CAR-T cells during expansion.
The role of mitochondrial biogenesis, oxidative phosphorylation and glycolysis in T cell function had been extensively studied.Naïve T cells usually adopted mitochondrial oxidative phosphorylation or fatty acid oxidation as the main metabolic mode [35].Upon encountering antigen, activated T cells were considered to rapidly switch to glycolysis with marked increases in biosynthetic intermediates.This conversion ability supported the rapid proliferation of T cells and the exertion of effector functions [36].Besides, it was shown that enhanced mitochondrial biogenesis was responsible for the central memory T cell formation and superior cytotoxicity of CAR-T cells [20,28].Due to the transient fratricide in the early stage of culture [24], the initial expansion of CD26 CAR-T cells was delayed and the production process would representative of 4 donors.data represented as mean ± Sd. differences between car-t cells and nt cells were shown and calculated by one-way anoVa test followed by dunnett's multiple comparisons test.be prolonged.Therefore, targeting the defects of mitochondrial and glycolytic metabolic profiling in the later stage may be particularly important.
PGC-1α, as a nuclear co-activator in mitochondrial biogenesis, controlled the mtDNA maintenance and transcriptional expression through activation of TFAM [37].Although tumor-infiltrating T cells demonstrated defects in mitochondrial mass and function, activating mitochondrial biogenesis through PGC-1α resulted in superior effector function [38,39].In our results, the expression of PGC-1α and TFAM was consistent with the mitochondrial biogenesis, which upregulated in the early stage and then downregulated in the later stage.This suggested that manipulating PGC-1α or TFAM may be an effective strategy for regulating mitochondrial biogenesis.Besides, to achieve the quick proliferation and powerful function, it was particularly important to upregulate GLUT1 expression and glucose uptake to fuel glycolysis in T cells [32,40].Unexpectedly, the expression of GLUT1 in CD26 CAR-T cells was dramatically downregulated instead of increased.The downregulation of key enzymes involved in glycolysis, G6PD and PDK1, further confirmed the weakening of glycolysis.This indicated that it may be also necessary and favorable to manipulate these glycolysis-related factors in the later stage.
CD26 CAR-T cells exhibited metabolic characteristics that were detrimental to T cells during expansion, and we also observed the fratricide effects and slow expansion.Whether these distinctive metabolic characteristics would impact the anti-tumor effect and functional status of CD26 CAR-T cells was unclear.Besides, it was reported that persistent stimulation of antigens caused T cell exhaustion, which was strongly associated with metabolic remodeling [41][42][43].Therefore, the effect function, activation and exhaustion transition of CD26 CAR-T cells during the expansion process still needed to be further clarified.In addition, previous studies had reported CD26 as a T cell activation antigen and played an important role in T cell function [44,45].However, the role of CD26 in mitochondrial biogenesis, metabolism and cell differentiation was still not defined in T cells.During the expansion, the downregulation of CD26 molecules and the targeting-CD26 CAR-T cell effects occurred simultaneously.Whether this metabolic remodeling revealed above was related to continuous antigen stimulation or CD26 downregulation remained unclear.Therefore, further exploration on the biological role of CD26 and the antigen stimulation in CAR-T cells was needed.Regulating the activity of CD26 using CD26 inhibitors such as sitagliptin or knocking out the CD26 gene may contribute to clarifying the mechanisms of metabolic remodeling in CD26 CAR-T cells [46].Besides, stimulation of CD19 CAR-T cells with CD19 antigens or exploring the metabolism of other CAR-T cells that self-expressed antigens would be insightful.
In conclusion, we revealed the variations in the metabolic profiling of CD26 CAR-T cells which self-expressed CD26 antigens.We discovered that while CD26 CAR-T cells survived the fratricide effect, they were in an unfavorably metabolic state at the later stage of expansion.There were serious defects in mitochondrial biogenesis, oxidative phosphorylation and glycolysis.Therefore, the mechanisms of metabolism reprogramming in CD26 CAR-T cells and optimization strategies were worth further investigation.

Figure 1 .
Figure 1. the mitochondrial biogenesis in cd26 and cd19 car-t cells.(a, b) the mitochondrial content (a) and mitochondrial dna: nuclear dna (mtdna/ndna) ratio (b) in cd19 and cd26 car-t cells on days 7, 14, 21 and 28. the mitochondrial content was indicated by the mean fluorescence intensity of mito-tracker green and normalized to nt cells.representative of 4 donors.(c-e)the mrna expression of pgc-1α (c), tFam (d) and mtco1 (e) in cd19 and cd26 car-t cells normalized to nt cells on days 7, 14, 21 and 28.representative of 3 donors.data represented as mean ± Sd. differences between car-t cells and nt cells were shown and calculated by one-way anoVa test followed by dunnett's multiple comparisons test.

Figure 2 .
Figure2. the memory characteristics of cd26 and cd19 car-t cells.(a) the flow cytometry analysis of memory profiling of cd26 and cd19 car-t cells on days 7 and 14. tcm (central memory t cells, ccr7 + cd45ra -), tScm (stem cell memory t cells, ccr7 + cd45ra + ), tEm (effector memory t cells, ccr7 -cd45ra -).(b-d) the ratio of tcm (b), tScm (c) and tEm (d) in cd26 and cd19 car-t cells.representative of 4 donors.data represented as mean ± Sd. differences between car-t cells and nt cells were shown and calculated by one-way anoVa test followed by dunnett's multiple comparisons test.

Figure 3 .
Figure 3. the mitochondrial metabolism in cd26 and cd19 car-t cells.(a) the atp production in cd19 and cd26 car-t cells that normalized to nt cells on days 7, 14, 21 and 28.representative of 4 donors.(b, c) the mrna expression of FaBp5 (b) and cpt1a (c) in cd19 and cd26 car-t cells that normalized to nt cells on days 7, 14, 21 and 28.representative of 4 donors.data represented as mean ± Sd. differences between car-t cells and nt cells were shown and calculated by one-way anoVa test followed by dunnett's multiple comparisons test.

Figure 4 .
Figure 4. Increased mitochondrial roS and decreased mmp in cd26 car-t cells.(a, b) the roS (reactive oxygen species) level (a) and mitoSox level (b) in cd19 and cd26 car-t cells on days 7, 14, 21 and 28. the data was represented by the mean fluorescence intensity and normalized to nt cells.representative of 4 donors.(c)the mitochondrial membrane potential (mmp) in cd19 and cd26 car-t cells on days 7, 14, 21 and 28. the result was calculated by the ratio of the median fluorescence intensity of Jc-1 aggregates (red) and Jc-1 monomers (green) and normalized to nt cells.representative of 3 donors.data represented as mean ± Sd. differences between car-t cells and nt cells were shown and calculated by one-way anoVa test followed by dunnett's multiple comparisons test.

Figure 5 .
Figure 5. downregulated glycolytic activity in cd26 car-t cells.(a-c) the expression of glut1 in nt cells, cd19 and cd26 car-t cells on days 14 and 21.representative of 5 donors.representative flow cytometry was showed in a, the ratio of glut1 positive cells and the mean fluorescence intensity of glut1 were shown in b and c, respectively.(d) the mrna expression of glut1 gene in cd19 and cd26 car-t cells that normalized to nt cells on days 7, 14, 21 and 28.representative of 4 donors.data represented as mean ± Sd. differences between car-t cells and nt cells were shown and calculated by one-way anoVa test followed by dunnett's multiple comparisons test.

Figure 6 .
Figure 6.downregulated glycolysis-related genes in cd26 car-t cells.(a, b) the mrna expression of g6pd (a) and pdK1 (b) gene in cd19 and cd26 car-t cells that normalized to nt cells on days 7, 14, 21 and 28.representative of 4 donors.data represented as mean ± Sd. differences between car-t cells and nt cells were shown and calculated by one-way anoVa test followed by dunnett's multiple comparisons test.