YIV-906 is a complex mixture prepared using a traditional hot water extract of four medicinal plants, Scutelleria baicalensis Georgi (S) and Paeonia lactiflora Pall. (P) Glycyrrhiza uralensis Fisch. (G), and Ziziphus jujuba Mill (Z) at a ratio of 3:2:2:2, respectively. It can be consistently prepared batch-to-batch and used in clinical trials. Details of the quality control of YIV-906 can be found in our previous reports (Tilton et al., 2010; Lam et al., 2018). YIV-906 water extract (100 mg/ml) was prepared as a stock for all culture assays.

RPMI1640 (Gibco, cat#11835030)+5%FBS+50ug/ml kanamycin at 37°C with 5%CO₂ incubation was used for cell culture and experiments. Authenticated Jurkat cell lines were purchased from ATCC (Jurkat, Clone E6-1), (T lymphoblast) cells were transfected with pcDNA-PD1-Dyk DNA (Genscript Biotech Corp, OHU263220D) using Lipofectamine LTX (Thermo Fisher Scientific, Cat#15338030) and then selected by G418 200 ug/ml. Authenticated Raji cell lines were purchased from ATCC (Raji-CCL-86™), (B-lymphocyte) cells were transfected with pcDNA-PD-L1-Dyk DNA (Genscript Biotech Corp, OHU22144D) and then selected by G418 600 ug/ml. Anti-PD1-PE (BioLegend, Cat#A17188B) or Anti-PD-L1(CD274)-APC (Invitrogen, Cat#17-5983-42) were used to confirm the expression of PD1 or PD-L1 in Jurkat cells or Raji cells using a laser flow cytometer (BD, Accuri 6 plus). Jurkat-PD1 cells were transfected with NFAT luciferase reporter DNA, where four repeated NFAT response elements, GGA​GGA​AAA​ACT​GTT​TCA​TAC​AGA​AGG​CGT (SEQ ID NO. 1) oligos, were inserted into the pGL4.2 vector (Promega, Cat#E6751) and then selected with 0.1ug/ml puromycin. Jurkat-PD1 NFAT-luc TCRαβ knockout cells were generated by transfecting CRISPR/Cas9-gRNA: eSpCas9-2A-GFP (Genscript Biotech Corp, SC1818) inserted into the targeting sequence using Lipofectamine LTX (Thermo Fisher Scientific, #15338030). The targeting DNA sequences for TCRa and TCRb were CTT​CAA​GAG​CAA​CAG​TGC​TG and AGG​TCG​CTG​TGT​TTG​AGC​CA, respectively. PE anti-human α/β T Cell Receptor Antibody (BioLegend, Cat#306708) was used to confirm the knockout of TCRαβ using flow cytometry. The targeting sequences for SHP1 (PTPN6) and SHP2 (PTPN11) were ACC​TGA​TCC​CCC​ACC​CTG​C and AGG​CCT​AGT​AAA​AGT​AAC​CC, respectively. Western blot analysis was used to confirm the knockout of SHP1 and SHP2 using anti-SHP1 (BioLegend, Cat#620301) and anti-SHP2 (Cell Signaling, Cat#3397S). Jurkat CART cell establishment: The pSLCAR-CD19-CD3zeta vector (Addgene_135993), pCMV-VSV-G (Addgene_8454), and psPAX2 (Addgene_12260) were transfected into HEK293T for 48h. The lentivirus was concentrated using a Lenti-X concentrator (TakaRa, Cat#631231) according to the manual. Concentrated Lentivirus- pSLCAR-CD19-CD3zeta was transduced into Jurkat-PD1 NFAT-luc cells using polybrene 10ug/ml (Millipore, Cat#TR-1003-G) and then GFP-positive cells were randomly sorted using flow cytometry. Jurka-PD1 NFAT-luc CAR-CD19-CD3zeta clones were randomly selected when NFAT luciferase activity was increased by co-culture with Raji cells overnight.

Jurkat cells (T cells) were stably transfected with nuclear factor of activated T-cells (NFAT) luciferase reporter DNA and PD1 DNA. For stimulation, 50 ul of Jurkat cells-PD1 cells at 5 × 10⁵/ml were co-cultured with 50 ul of Raji cells at 10⁶/ml in a 1:2 ratio and staphylococcal superantigen E (SEE) (Toxin Technology, Cat#ET404) 1 ng/ml to 10 ng/ml, in the absence or presence of YIV-906 or its components were added to the mixed cells overnight at 37°C with 5%CO₂ incubation. NFAT activity was determined by measuring luciferase activity chemiluminescence. 75 ul of Jurkat-PD1 cells at 10⁶ cells/well with or without wild-type Raji cells or Raji-PD-L1 cells at 2 × 10⁶ cells/well were seeded into round-bottom 96-well plates. In some experiments, InVivoSIM anti-human PD-1 (Nivolumab Biosimilar) (BioXcell, Cat#SIM0003) 18 ug/ml was added to the cells for 2 h or 24 h before drug and see treatment. 25 ul of control medium or 25 ul of 5x concentration of YIV-906 or other drugs were added to the well. 25 ul of see (5 × 10 ng/ml) was added to the wells. After overnight incubation at 37°C with 5%CO₂, the cells were lysed using a luciferase lysis buffer which contained luciferin to generate the luminescence. The luminescence was recorded using a luminescence microplate reader. Each data point represents three experiments of triplicate samples from the NFAT luciferase reporter assay.

Jurkat-PD1 cells 2 × 10⁵/ml with or without wild-type Raji cells or Raji-PD-L1 cells (4 × 10⁵/ml) in 24-well plates were treated with YIV-906 with or without SEE (10 ng/ml and 30 ng/ml), for 48 h. Anti-CD69-FITC (BioLegend, Cat#310904) and anti-PD1-PE (BioLegend, Cat#A17188B) in PBS+1%BSA was used to stain the cells. Anti-PD1-PE was used to separate Jurkat-PD1 cells from mixed cells. The expression of CD69 in the FITC channel was presented as median fluorescence or total Jurkat cell population percentage using a laser flow cytometer (BD, Accuri 6 plus). Cell death of Raji cells (co-cultured with Jurkat-PD1-CAR-CD-19 cells/Raji cells at a 2:1 ratio) was determined by gating CD19⁺ Raji cells (Anti-CD19-perCP, BioLegend, #30228) and Annexin V-PE (BioLegend, #640947) positive and/or Helix-NR positive cells. Each data point represents the average mean of four experiments of triplicate samples of flow cytometry assays.

Following the treatment, the medium was collected for IL2, IFNg, and IL10 detection using fluorescence bead array (BioLegend: LEGENDplex assays, Human CD8/NK Panel, Cat#740267) according to the to the manufacturer’s instructions.

pNPP assays were used to determine the inhibitory effects of YIV-906 and its components on recombinant human SHP2(PTPN11) and SHP1(PTPN6) (LSBio, WA). Briefly, 20 ng of enzyme with different doses of YIV-906 or its components were reacted in the 100 ul reaction buffer (5 mM pNPP (p-Nitrophenyl Phosphate), 25 mM Hepes pH 7.3, 2.5 mM EDTA, 2.5 mM DTT, 100 ug/ml BSA) in a well of a 96-well plate at 37°C for 1 h. 100 ul of 2N H₂SO₄ were used to stop the reaction. Heated SHP1 or SHP2 (90°C for 15 min) was used as the control in parallel experiments. OD optical 450 nm was measured. The OD of wells without added drugs was normalized to 1 after subtracting the wells’ OD without enzyme. Each data point represents the average mean of three experiments of triplicate samples.

Jurkat-PD1 cells 10⁶/ml were placed in 24-well plates and treated with YIV-906 or its constituent herbs for 45 min. The cells were collected at 1000 g centrifugation for 10 min. The cells were prepared with 2X protein loading buffer (Tris pH 6.8 1M, SDS 1%; glycerol; β-mercaptoethanol; bromophenol blue; and distilled water). The samples were then heated to 95°C for 5 min to denature the proteins prior to western blotting. SDS polyacrylamide gel electrophoresis (10% Mini-Protean TGX™ Precast Protein Gels, Bio-Rad) was used to separate the proteins according to their electrophoretic mobility. 20 ug of Protein extract per 10 µL per well was used. Migration was performed in a 1X running buffer (Tris/Glycine/SDS) at 185 V for 50 min. Proteins were transferred onto a PVDF membrane in transfer buffer (Tris-CAPS AX, methanol, SDS 10%, distilled water) at 75 V for 1 h. After blotting, the membrane was cut into two parts with an approximate size of 3 cm (height) × 9 cm (width) to fit into the blocking chamber. The upper part of the membrane was used for probing target proteins with specific antibodies (as described below), and the lower part of the membrane was used for probing GAPDH as a protein loading control for normalization. Non-specific binding sites on the PVDF membrane were blocked with a blocking solution (3% milk powder and 1X TBS-T) for 30 min. The PVDF membrane was then incubated with the primary antibody against the proteins of interest overnight at 4 °C. The primary antibodies used were as follows: P-Lck-Y394 (BioLegend, Cat#933101), P-Zap70-Y319 (Cell Signaling Technology, Cat#2717), P-LAT-Y191 (Cell Signaling Technology, Cat#3584), P-SRC(Fyn)Y416 (Cell Signaling Technology, Cat#6943), P-Pyk2-Y402 (Cell Signaling Technology, Cat#3291), and GAPDH (Cell Signaling Technology Cat# 5174, RRID:AB_10622025). The membrane was washed with TBS-T 1X and incubated with a secondary antibody with horseradish peroxidase-conjugated anti-rabbit IgG 1:5000, (Thermo Fisher Scientific Cat# A27036, RRID:AB_2536099), against the immunoglobulin corresponding to the primary antibody for 1 h at room temperature. The membrane was then washed with TBS-T 1X. The protein bands were detected using chemiluminescence (Super Signal West Dura, Thermo Scientific, Cat#PI34076) and the images were acquired using an X-ray film processor (Fuji Super RX-N). Densitometric scanning was performed using an Epson V600 scanner. ImageJ software (ImageJ, RRID:SCR_003070) was used to quantify the total intensity of the immunoreactive bands. GAPDH was used as an internal control for normalization.

LC-MS analysis was performed using an Agilent 1200 series HPLC coupled with an AB SCIEX 4000 QTRAP mass spectrometer. The separation was conducted on an Alltima™ HP HPLC column (5 mm, 4.6 × 250 mm). The mobile phase was acetonitrile (A) and water with 0.1% formic acid (B) with gradient elution: 0 min, 5% A; 10 min, 20% A; 20 min, 25% A; 40 min, 30% A; 45 min, 35% A; 55 min, 45% A; 60 min, 70% A; 62 min, 90% A; 67 min, 90% A; 68 min, 5% A; and 75 min, 5% A. The flow rate was 1.0 ml/min, and the column temperature was set at 30°C. ESI negative mode mass spectrometry of scan rate 4000 amu/s was performed with the following ionization parameters: CAD: High; TEM:550.00°C; GS1:55.00; GS2:50.00; ihe: ON; IS: 4250.00; DP: 40.00; CES 0.00; CE: 5.00. The mass range for detection was 120–800 amu. Using a custom program integrated with MZmine software, the chemicals were identified based on their retention time, total mass, and ion pair fragment mass.

Data was analyzed using one- or two-way analyses of variance (ANOVA) tests (GraphPad Prism 7), correlation analyses (GraphPad Prism, RRID:SCR_002798), and Student’s t-test distribution (Microsoft Excel, RRID:SCR_016137). Differences were considered statistically significant at p < 0.05.

A cell culture model of Jurkat cells co-cultured with staphylococcal superantigen E (SEE)-Raji cells was established to examine the effects of YIV-906 on T-cell activation. In this cell culture model, YIV-906 was able to promote NFAT activity by approximately one-fold in either the absence (Figure 1A) or presence (Figure 1B) of SEE, which could stimulate NFAT activity about 40-fold in the range of 80 ug/ml to 320 ug/ml, which are non-toxic doses for Jurkat cells and Raji cells.

Effects of different component herbs (at equivalent YIV-906 concentrations): Glycyrrhiza uralensis Fisch (G), Paeonia lactiflora Pall (P), Scutellaria baicalensis Georgi (S), and Ziziphus jujuba Mill (Z) on NFAT-mediated transcriptional activity of Jurkat cells was compared. S showed a very similar dose response to YIV-906 in both the absence and presence of SEE (Figures 1A, B); G and P could modulate no more than 20% NFAT activity in the absence or presence of SEE (Figures 1A, B); and Z had no discernable impact on NFAT activity (Figures 1A, B). When comparing two-herb combinations or three-herb combinations (Supplementary Figures S1A–D), only the combinations that included S showed an approximately 1-fold enhancement of NFAT activity in the absence or presence of SEE. We demonstrated that YIV-906 modulates nuclear factor of activated T-cells (NFAT) activity in Jurkat cells. S played the most important role in the modulation of NFAT, conversely G and P might play minor roles in the modulation of NFAT.

To study the impact of PD1-PD-L1 on NFAT activity during T cell activation, Jurkat cells-PD1 cells were co-cultured with PD-L1 over-expressed Raji cells, with or without SEE. YIV-906, S, and combinations with S had similar effects on the basal NFAT activity of Jurkat-PD1 cells, either incubated with Raji (Figure 1A, Supplementary Figures S1A–C) or PD-L1 over-expressed Raji cells (Raji-PD-L1) (Figure 1C, Supplementary Figures S1E–G). PD1 Interactions with Jurkat cells and PD-L1 with Raji cells inhibited NFAT activity of Jurkat cells by ∼50% (Figure 1B vs. 1D). YIV-906, S, and combinations with S rescued the depressed NFAT activity of Jurkat-PD1 cells (Figure 1D, Supplementary Figures S1F, H). G had a slight impact on NFAT activity under the above conditions (Figures 1C,D and Supplementary Figures S2E–H).

Since NFAT is a very important T cell activation transcriptional factor, the impact of YIV-906 on T cell activation was further studied. CD69 was used as a T cell activation marker. As shown in Figure 1E, YIV-906 enhanced CD69⁺ population (Figure 1E) and the median CD69-fluorescence (Supplementary Figure S2A) of Jurkat-PD1 cells when incubated with Raji cells without SEE, or with 3 ng or 10 ng/ml SEE. The interaction between PD1 and PD-L1 reduced CD69 expression induced by SEE, but YIV-906 could help to restore CD69 expression in Jurkat-PD1 cells (Figure 1F and Supplementary Figure S2B). This result suggested that YIV-906 could modulate NFAT activity of T cells, thereby leading to stronger T cell activation even under PD1-PD-L1 interaction conditions. Since Jurkat cells are a leukemia T cell, their responses may not perfectly reflect the normal conditions. T cells isolated from mice spleen were used to confirm if YIV-906 and S could still promote T cell activation. Results demonstrated that YIV-906 and S treatment could increase CD69⁺ population of CD4⁺ cells (Supplementary Figure S2C) and CD8⁺ cells (Supplementary Figure S2D) isolated from mouse spleen. This result further supported that YIV-906 and S have potential to promote T cell activation. In further, we will seek collaboration to study the impact of YIV-906 on primary human T lymphocytes.

To investigate whether YIV-906 and its components activates NFAT activity through direct activation of TCR, the TCRαβ gene was knocked out from Jurkat-PD1 cells using CRISPR/Cas9 technology (Figure 2A). Jurkat-PD1 TCRαβ knockout (KO) cells did not respond to SEE when co-cultured with Raji cells (Figure 2B). As shown in Figures 2C, D, YIV-906 and S had very similar impacts on NFAT activity in Jurkat-PD1 and Jurkat-PD1 TCRαβ KO cells. These results suggest that TCRαβ are not essential for the action of YIV-906 and its components on NFAT. YIV-906 and its components may have direct downstream targets of TCR αβ.

10R-VIVIT peptide, which is a cell-permeable peptide, could inhibit NFAT activity by blocking calcineurin-mediated dephosphorylation of NFAT. 10R-VIVIT at 25 uM, a non-toxic dose, could inhibit about 70% of NFAT activity of Jurkat-PD1 cells when co-cultured with Raji or Raji-PD-L1 cells (Supplementary Figure S3A). Under 10R-VIVIT (25uM) treatment conditions, YIV-906 and S could still promote NFAT activity of Jurkat-PD1 cells with similar magnitude (about 2 folds) without 10R-VIVIT treatment conditions (Supplementary Figure S3B, 3C). Therefore, calcineurin might not be the major target for YIV-906 and S.

To test whether YIV-906 and its components could directly impact downstream signaling of TCR, their effects on the phosphorylation of T cell receptor downstream cascades, including Lck, Zap70, LAT, Fyn, and Pyk2 in Jurkat cells was examined using western blot analysis. As shown in Figure 3A, treatment with YIV-906 at a dose of 320 ug/ml for 45 min induced the phosphorylation of Lck, Zap70, LAT, Fyn, and Pyk2. By comparing individual herb effects at equivalent concentrations (320 ug/ml), component S was found to play a major role in protein phosphorylation induction by YIV-906 (Figure 3A). Using a C18 column to fractionate YIV-906, higher relative induction activity of protein phosphorylation was observed in 45%–75% acetonitrile/methanol (A/M) fractions (Figure 3B). S was found to have an activity profile similar to YIV-906 (Figure 3C). The 45% A/M fraction of S had a strong effect on P-Zap70-Y319, whereas the 60% A/M fraction of S had a strong effect on P-Lck-Y39, P-LAT-Y191, P-SRC(Fyn)-Y416, and P-Pyk2-Y402 (Figure 3C). The chemicals present in the S fractions are listed in Supplementary Figure S4A and they had been reported in our previous reports (Ye et al., 2007; Zhang et al., 2010). In Supplementary Figure S4A, protein phosphorylation induced by these fractions was indicated using red upward arrows. It should be noted that these chemicals might have different modes of action on the targets.

Some chemicals from active S fractions were studied for their ability to induce protein phosphorylation of T cell receptor downstream cascades, including Lck, Zap70, LAT, Fyn, and Pyk2 in Jurkat cells using western blot analysis. Baicalein strongly induced phosphorylation of all the examined proteins (Figure 3D). Both baicalin and oroxylin A 7-O-beta-D-glucuronide (OA-7-S), at 40 uM, showed some induction effects on P-Zap70-Y394, P-SRC(Fyn)-Y416, and P-Pyk2-Y402 (Figure 3D). These compounds potentially could play a critical role in T cell activation by YIV-906. Some S compounds promoted NFAT activity: Oroxylin A 7-O-beta-D-glucuronide (OA-7-side), baicalein, oroxylin A, and chrysin were able to modulate the basal NFAT activity of Jurkat-PD1 cells incubated with Raji, or PD-L1 over-expressed Raji cells (Figures 3E, F). The measured modulatory effects were independent of TCRαβ expression (Supplementary Figure S4B, C), but the optimum doses were different. Oroxylin A 7-O-beta-D-glucuronide at 40 uM promote NFAT activity by 2 folds, whereas baicalein, oroxylin A, and chrysin required only 10 to 2.5 uM to have a similar impact on NFAT. In the presence of SEE, oroxylin A 7-O-beta-D-glucuronide, baicalein, oroxylin A, and chrysin showed similar promotional effects as they did on basal NFAT activity. It was also observed that wogonoside, baicalein, oroxylin A, and chrysin inhibited SEE-triggered NFAT activity at concentrations from 20 to 40 uM (Figures 3G, H). Some compounds mentioned above exhibited a biphasic effect on NFAT activity, possibly because these compounds might have multiple modes of action on their target(s) that are associated with the NFAT pathway. In the tumor microenvironment, overall activity might depend on the composition and internal interactions of compounds within YIV-906, hence no single compound or target can simply explain all activities.

SHP2 (SH2-containing phosphatase 2, a protein tyrosine phosphatase) is an immediate downstream mediator of PD1 (Marasco et al., 2020; Patsoukis et al., 2020). Once PD1 binds to PD-L1, PD-1 will phosphorylate its immune receptor tyrosine–based inhibitory motif (ITIM) and immune receptor tyrosine–based switch motif (ITSM) to recruit and activate SHP2 (Marasco et al., 2020; Patsoukis et al., 2020). Activated SHP2 initiates T-cell inactivation by dephosphorylating TCR downstream cascade proteins (Marasco et al., 2020; Patsoukis et al., 2020). In the absence of SHP2, SHP1 can replace SHP2 (Celis-Gutierrez et al., 2019). SHP2 inhibitors could block PD1 action and promote T-cell activation for facilitating cancer treatment (Zhao et al., 2019; Yuan et al., 2020). As shown in Figure 4A, YIV-906 showed inhibitory effects on both SHP1 and SHP2 enzyme activities, with stronger potency against SHP1. G and S component herbs of YIV-906 were found to be the key herbs responsible for SHP1 and SHP2 inhibition (Supplementary Figure S5A, B); and some individual chemical compounds in G and S were found to have inhibitory effects on SHP1 and SHP2 (Supplementary Figure S5C, D). These compounds demonstrate selectivity for SHP1 and SHP2, for example, isoliquiritigenin was more potent against SHP1 than SHP2, whereas glycyrrhizic acid showed stronger inhibition of SHP2 than SHP1 (Supplementary Figure S5C, D). Baicalin, oroxylin A 7-O-beta-D-glucuronide, wogonoside, baicalein, chrysin, and oroxylin A from S showed stronger inhibitory effects on SHP2 than on SHP1 (Supplementary Figure S5C, D).

To determine whether TCR downstream protein phosphorylation induced by YIV-906 is dependent on the presence of SHP1 and/or SHP2, SHP1 and/or SHP2 genes were knocked out in Jurkat-PD1 cells using CRISPR-gRNA (Supplementary Figure S6). It appeared that SHP2 and SHP1/2 knockout cells had higher basal levels of protein phosphorylation (Figure 4B and Supplementary Figures S7A–G). SHP2 might be a critical enzyme for maintaining lower levels of TCR downstream protein phosphorylation as well. Comparing to Jurkat cells, SHP1 knockout reduced the phosphorylation of all proteins triggered by YIV-906 (Figure 4B and Supplementary Figures S7A–G). SHP2 knockout did not affect the protein phosphorylation of LCK and LAT triggered by YIV-906 (Figure 4B and Supplementary Figures S7A–G). However, the level of phosphorylation of Zap70, Fyn, and PYK2 in SHP2 knockout cells only slightly increased under YIV-906 treatment (Figure 4B and Supplementary Figures S7A–G). It is interesting to note that LCK, Zap70, and PYK2 in SHP1/2 knockout cells had little or no response to YIV-906 and that a high dose of YIV-906 decreased LAT and Fyn phosphorylation (Figure 4B and Supplementary Figures S7C–G). Overall SHP1 and SHP2 could have different impacts on regulating the basal levels of protein phosphorylation of TCR downstream. Both SHP1 and SHP2 were important to YIV-906 triggering Zap70 and Fyn phosphorylation (Supplementary Figure S7G). Additionally, SHP1 and SHP2 double knockout minimized TCR downstream protein phosphorylation triggered by YIV-906 (Supplementary Figure S7G).

YIV-906’s impacts on promoting NFAT response by SHP1 and SHP2 were studied. As shown in Figure 4C, knockout of SHP1 and/or SHP2 in Jurkat cells reduced but did not completely block NFAT activity triggered by YIV-906 in the absence of SEE. Under PD1-PD-L1 interaction conditions, single knockout of SHP2 but not SHP1 significantly increased NFAT activity (Figure 4D). SHP1/2 knockouts completely rescued the depressed NFAT activity due to PD1-PD-L1 interaction (Figure 4D). These results support previous findings that PD-1 preferentially binds with SHP-2 over SHP-1 to inhibit TCR response (Celis-Gutierrez et al., 2019). Under Jurkat-PD1-Raji plus SEE conditions, knocking-out SHP1 and/or SHP2 did not have a significant impact on the cells’ NFAT response to YIV-906 (Figure 4E). When Jurkat-PD1 cells were co-cultured with Raji-PD-L1 plus SEE, SHP1 or SHP2 knockout only slightly decreased the NFAT response to YIV-906 (Figure 4F), but SHP1/2 double knockout significantly reduced the NFAT response to YIV-906. Overall, our results indicated that YIV-906 could overcome PD1-PD-L1 suppression and promote NFAT activity by affecting SHP1 and SHP2 activity.

As YIV-906 can modulate the antitumor activity of anti-PD1 against tumors in animal models, we investigated whether YIV-906 could cooperate with anti-PD1 to promote NFAT activity for T cell activation. When co-cultured with Jurkat-PD1 and Raji plus SEE, the addition of anti-PD1 (non-therapeutic biosimilar antibody to nivolumab , with the same variable regions ) did not affect the NFAT response of Jurkat-PD1 cells to YIV-906, S, or S compounds (Supplementary Figures S8A–D). As shown earlier, co-cultured Jurkat-PD1 cells and Raji-PD-L1 cells inhibited the NFAT activity of Jurkat cells triggered by SEE (Figures 5A, B). YIV-906, S, and S constituent compounds restored any depressed NFAT activity in Jurkat-PD1 cells (Figures 5A, B). As expected, anti-PD1 rescued the depressed NFAT activity of Jurkat-PD1 cells due to PD1 and PD-L1 interactions (Figures 5C, D). Most importantly, YIV-906, S, and S compounds (with different optimum dose) could further enhance anti-PD1 action to promote NFAT activity in Jurkat PD1 cells (Figures 5C, D). The above results demonstrated that YIV-906 and its components could cooperate with anti-PD1 to promote T-cell activation for immune therapy.

It is well known that MHC-antigen-TCR interaction recruits CD3 receptors to transduce signals for NFAT activation. Most chimeric antigen receptors (CAR) are composed of an antigen recognition domain, an extracellular hinge region, a transmembrane domain, and co-stimulatory domains such as immunoreceptor tyrosine–based activation motifs (ITAM) of CD3ζ(zeta) (Lindner et al., 2020). Once a chimeric antigen receptor (CAR) binds to its target, a signal is transduced to its ITAM of CD3ζ which further stimulates the downstream cascade leading to NFAT activation (Lindner et al., 2020). Because MHC-antigen-TCR interactions or CAR-target interactions both have the same downstream cascade stimulating NFAT activity, and YIV-906 demonstrated the potential to enhance MHC-antigen-TCR triggered NFAT response, we asked whether YIV-906 could also modulate CAR-triggered NFAT activity. To study this, Jurkat-PD1 NFAT luciferase receptor cells were transduced with CAR-CD19⁻CD3ζ(zeta) lentivirus to target CD19 on Raji cells.

YIV-906, S, and S compounds had similar effects on promoting the basal activity of NFAT in both Jurkat-PD1-CAR-CD19 cells (Supplementary Figures S9A, B) and Jurkat-PD1 cells alone (Figure 2C and Supplementary Figures S4B). When Jurkat-PD1-CAR-CD19 cells were co-cultured with Raji cells, a 40-fold increase in NFAT-driven luciferase activity was induced (Figure 6A). Under these conditions, YIV-906, S, and S compounds further promoted NFAT activity triggered (Figure 6A and Supplementary Figures S9C). PD1 and PD-L1 interactions depressed NFAT activity triggered by the interaction CAR-CD19 and CD19 (Figure 6B). YIV-906, S, and S constituent compounds (oroxylin A 7-O-beta-D-glucuronide, wogonoside, baicalein, oroxylin A, and chrysin) could restore this depressed NFAT activity (Supplementary Figures S9D). Depending on the compound, the optimum dose for modulation may be different.

The production of IL2, IFNg and IL10 were used to monitor the response of Jurkat-PD1-CAR-CD19 cells with or without co-culturing Raji cells or Raji PD-L1 cells. As predicted, co-culturing Jurkat-PD1-CAR-CD19 cells with Raji cells could trigger higher production of IL2 and IFNg as comparing to Jurkat-PD1 plus Raji cells or Raji PD-L1 cells as well as their single culture (Figures 6C, D, Supplementary Figures S10A–F). YIV-906 and S could further enhance production of IL2 and IFNg production when

Jurkat-PD1-CAR-CD19 cells were co-cultured with Raji cells (Figures 6C, D). When Jurkat-PD1-CAR-CD19 cells were co-cultured with Raji PD-L1 cells, PD1 and PD-L1 interaction depressed IL2 and IFNg production (Figures 6C, D). YIV-906 and S could also promote IL2 and IFNg production under PD1 and PD-L1 interaction conditions (Figures 6C, D). YIV-906 and S had no big impacts on IL10 production in all culturing conditions (Figure 6E and Supplementary Figures S10G–I). Our results indicated that YIV-906 and S could further promote CAR T cell activation by enhancing IL2 and IFNg production when CAR T cells interact to its target cells.

We further demonstrated that YIV-906 could promote Jurkat-PD1-CAR-CD19 cells to kill Raji cells. As shown in Supplementary Figures S11A, B, Raji or Raji-PD-L1 cell death could be increased by increasing Jurkat-PD1-CAR-CD19 cells in the co-culture. It should be noted that PD1-PD-L1 interaction did reduce Raji-PD-L1 cell death triggered by Jurkat-PD1-CAR-CD19 cells (Fig S11A and S11B). YIV-906 could enhance the Jurkat-PD1-CAR-CD19 cells’ killing capability on either Raji or Raji-PD-L1 cells (Figures 6F, G and Supplementary Figures S11A, B). Among the component herbs, S was found to play a key role of YIV-906 to promote Raji or Raji-PD-L1 cell death (Figures 6F, G and Supplementary Figures S11C, D) caused by Jurkat-PD1-CAR-CD19 cells.”

K562 cells, which are myelogenous leukemia cell, did not express CD19 as comparing to Raji cells (Supplementary Figures S11E). When K562 cells were co-cultured with Jurkat-PD1-CAR-CD19 cells, no induction of NFAT activity or K562 cell death was observed (Supplementary Figures S11F, G). In addition, YIV-906 and S did not promote cell death of K562 cells when co-cultured with Jurkat-PD1-CAR-CD19 cells. This result demonstrated that Jurkat-PD1-CAR-CD19 cells had selective cytotoxicity on cells with CD19 expression.

Baicalein, chrysin, and oroxylin A of S (up to 10uM, did not cause significant cell death in Raji (Supplementary Figures S12A) or Raji-PD-L1 (Supplementary Figures S12B) when co-cultured, however these compounds promoted the cell killing capability of Jurkat-PD1-CAR-CD19 cells on Raji (Supplementary Figures S12C) or Raji-PD-L1 (Supplementary Figures S12D). These results suggest that YIV-906 and its components might have potential to facilitate CAR T cell therapy for cancer treatment.