Abstract

   Acute myeloid leukemia (AML), the most common leukemia in adults,
   exhibits immune escape characteristics like solid tumors. The
   expression of indoleamine 2,3-dioxygenase 1 (IDO1), a well-recognized
   immune checkpoint, has been detected in AML blast cells and is
   associated with poor clinical outcome. Although an imbalance of B cell
   subpopulations exists in AML patients’ bone marrow microenvironment,
   the role of B cells and their interaction with IDO1 in AML have yet to
   be elucidated. Herein, with bioinformatic analysis, we found the close
   correlations between IDO1 expression and survival and B cell
   subpopulation proportions in AML patients. Further, our investigation
   into IDO1 expression and activity, B cell subpopulation proportions and
   immunosuppressive interleukin-10 (IL-10) level in AML cells and
   clinical samples revealed significant findings. Using a co-culture
   system of healthy human PBMCs and AML cell lines, we demonstrated that
   high IDO1 expression in AML cells could alter the proportions of total
   B, regulatory B and memory B cells, and increased the level of IL-10.
   Finally, with the IDO1 inhibitor RY103 designed by our laboratory, we
   found that IDO1 inhibition had good anti-leukemic effect and restored
   the abnormal proportions of B cell subpopulations in AML mice. Our
   study is the first to reveal the modulation of IDO1 on B cell
   subpopulations in AML, making a significant breakthrough in
   understanding the immune escape mechanisms of AML. Application of IDO1
   inhibitor, such as RY103, targeting the imbalance of B cell
   subpopulations can lead to innovative treatments for AML.

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s43556-025-00262-x.

   Keywords: Acute myeloid leukemia (AML); Immunosuppression; Indoleamine
   2,3-dioxygenase 1 (IDO1); B cells; IDO1 inhibitors

Introduction

   Acute myeloid leukemia (AML), the most prevalent leukemia in adults, is
   marked by abnormal proliferation of leukemic blasts in both bone marrow
   (BM) and peripheral blood (PB). Nevertheless, the underlying
   immunological mechanism of AML is not yet fulled elaborated
   [[38]1–[39]3]. AML shares similar immune evasion mechanisms with solid
   tumors, including the upregulation of amino acid metabolic enzymes such
   as indoleamine 2,3-dioxygenase-1 (IDO1) [[40]4] and arginase-2 (ARG2)
   [[41]5], downregulation of antigen presentation molecules,
   overexpression of inhibitory T cell ligands, and the production of
   reactive oxygen species. Thus, in turn, dysfunction of natural killer
   (NK) cells and T cells [[42]6], proliferation and activation of
   immunosuppressive cell populations like regulatory T cells (Tregs)
   [[43]7], myeloid-derived suppressor cells (MDSCs) [[44]8] and
   tumor-associated macrophages (TAMs) [[45]9] can be induced.

   IDO1, an immunomodulatory enzyme, catabolizes the degradation of
   tryptophan (Trp) into N-formyl-kynurenine, which represents the intial
   and rate-limiting step of the kynurenine pathway (KP) [[46]10]. IDO1
   has emerged as a key player in suppressing anti-tumor responses and
   fostering tumor immune evasion [[47]11].The immunosuppressive effect of
   IDO1 is driven by the depletion of Trp and the accumulation of Kyn,
   that affects immune cells such as NK cells [[48]12, [49]13], effector T
   cells [[50]14], Tregs [[51]15], tolerant dendritic cells (DCs)
   [[52]16], and MDSCs [[53]17], leading to tumor immune evasion [[54]18].
   Several IDO1 selective inhibitors have been developed. IDO1 inhibitory
   activities and anti-tumor effects of epacadostat [[55]19, [56]20],
   1-methyl-L-tryptophan (1-L-MT) [[57]21, [58]22] and RY103
   [[59]23–[60]25] in CT26 colorectal cancer, KPIC pancreatic carcinoma,
   GL261 glioma and B16F10 melanoma mouse models have been demonstrated.
   It has been reported that IDO1 expresses in both BM and PB AML blasts
   [[61]26–[62]29], whose expression negatively correlates with clinical
   outcome, including the response to chemotherapy, risk of relapse and
   overall survival (OS) [[63]30–[64]33]. IDO1 constitutively expressing
   AML cells co-cultured with T cells and DCs can inhibit CD4^+and CD8^+T
   cell proliferation, promote Tregs proliferation, impair DCs maturation,
   and eventually suppress anti-tumor immune response
   [[65]34–[66]36].Despite the fact that 3 clinical studies
   ([67]NCT03491579, [68]NCT03444649, [69]NCT02835729) have been conducted
   to explore the efficacy of combination of epacadostat and chemotherapy
   in AML [[70]11], the role of IDO1 in AML and its underlying mechanism
   is still uncertain.

   There is limited research on the potential contribution of B cells to
   the anti-tumor immune response during tumor progression [[71]9].
   Despite the growing body of evidence that B cells possess the
   capability to modulate T cell activation and nonspecific immune
   responses [[72]37], its role within the bone marrow microenvironment
   (BMM) has yet to be elucidated. Studies have shown that in AML there is
   a reduction of total B cells (CD19^+ B cells) and an imbalance in the
   proportions of different B cell subpopulations: compared with healthy
   individuals, total B cells in the BM of relapsed or refractory AML
   patients are significantly reduced [[73]38]; within one year after
   intensive cytotoxic chemotherapy, AML patients exhibit abnormal
   proportions of memory and transitional B cells in peripheral blood
   mononuclear cells (PBMCs) [[74]39]; transcriptome sequencing analysis
   of BM samples from AML infants revealed that many significantly altered
   genes are associated with the development and differentiation of B
   cells [[75]40]; the proportions of regulatory B cell (Breg) cells in
   PBMCs and bone marrow mononuclear cells (BMMCs) of AML patients are
   higher than healthy subjects, and overall survival of patients with a
   higher proportion of Breg cells are shorter [[76]41]. These studies
   suggest that there is an imbalance of B cell subpopulations in PBMCs
   and BMMCs of AML patients. Alteration of proportions of different B
   cell subpopulations in BMM may exert a significant influence on the
   pathological process of AML, but the specific molecular mechanism has
   not been elucidated.

   Although there are opinions that IDO1 is related to B cells dysfunction
   [[77]42, [78]43], the effect of IDO1 on B cells remains to be explored.
   Herein, we aimed to clarify whether IDO1 is involved in the imbalance
   of B cell subpopulations in AML microenvironment, and whether IDO1
   inhibitors are effective against AML. With clinical datasets and
   samples, the correlation between IDO1 and overall survival, immune cell
   proportions, especially B cell subpopulation proportions in AML
   patients were explored. The effect of IDO1 on B cells were studied in a
   co-culture system of healthy human PBMCs and AML cell line.
   Furthermore, the therapeutic efficacy of IDO1 inhibitors in C1498 AML
   bearing mice was examined. The present study represented a pioneering
   effort in demonstrating the impact of IDO1 on the imbalance of B cell
   subpopulations, which provided theoretical basis for immunotherapy of
   AML.

Results

IDO1 expression negatively correlated with survival and associated with
distinct B Cell subpopulation proportions of AML patients

   To identify the prognostic value of IDO1 expression and the potential
   correlation of IDO1 expression with B cell subpopulation proportions in
   AML patients, we used the Ruijin Hospital dataset to compare the
   survival between IDO1-high group and IDO1-low group. Based on the gene
   expression level of IDO1 in BM specimens (maximum rank statistic), AML
   patients from Ruijin Hospital were divided into low and high groups
   (Definition standards see Materials and Methods). Results suggested
   that AML patients with high IDO1 expression had worse overall survival
   (p = 0.014) in AML (Fig. [79]1a). The IDO1-high group did not show
   independent prognosis significance (Fig. [80]1b) in AML patients.
   Subgroup analysis indicated that the gene expression level of IDO1 was
   lower in AML-M3 subtype according to French American British (FAB)
   classification criteria (also known as acute promyelocytic leukemia)
   (Fig. [81]1c) and 20–29 age group (Fig. [82]1d), while the > = 70 years
   age group showed relatively higher gene expression of IDO1
   (Fig. [83]1d). The comparison between IDO1-low and IDO1-high groups
   indicated that high IDO1 gene expression was associated with older age,
   lower BM blasts, HGB, and higher WBC level (Table S3). In addition, the
   positive rate of gene fusions was lower in IDO1-high group, in
   particular CBFB::MYH11 and RUNX1::RUNX1T1, which are considered as
   relatively favorable genetic lesions of the ELN risk classification
   (Fig. S1). In contrast, the TET2, FLT3, RUNX1, BCOR and spliceosome
   mutations were more common in the IDO1-high group, which were commonly
   enriched in myelodysplastic syndrome (MDS)-transformed AML (Table S3).

Fig. 1.

   [84]Fig. 1
   [85]Open in a new tab

   In AML patients, IDO1 expression negatively correlated to survival and
   the proportion of naïve B cells and positively correlated to the
   proportion of memory B cells. Clinical Data from Ruijin Hospital were
   analyzed. a Kaplan–Meier survival curves of OS of the AML patients
   according to IDO1 mRNA expression level in bone marrow specimens.
   IDO1-low group (n = 871) and IDO1-high group (n = 126) were compared
   using the log-rank test. b Forest plot of multivariant analysis of
   basic clinical information and IDO1 expression groups. c and d Boxplots
   of IDO1 gene expression level in the FAB diagnosis and age groups of
   patients with AML. e The correlations between IDO1 expression and
   relative proportions of 22 cell types in the BMM of AML patients using
   CIBERSORT method coupled with LM22. Fork marks indicate statistical
   significance below 0.05. f Box plots of naïve B cells and memory B
   cells fraction in IDO1-high AML group, IDO1-low AML group and non-AML
   group. g and h Scatter plots of naïve B cells(left), memory B cells
   fractions (right) and IDO1 gene expression level (log[2] (TPM + 1)).
   i STRING gene network of up-regulated genes in the high IDO1 expression
   group. j Bar plot of enriched gene pathways using up-regulated genes in
   high IDO1 expression group. k Flow cytometry analysis of the
   proportions of naïve B cells (CD19^+IgD^+CD27^−), memory B cells
   (CD19^+CD27^+) and Breg cells (CD19^+CD24^+CD27^+) in PBMCs from
   healthy individuals and AML patients. n = 3 per group. Statistical
   significance was determined by Student’s t-test and data were expressed
   as mean ± S.D., *p < 0.05, ***p < 0.001

   CIBERSORTx-based deconvolution and the correlation analysis revealed
   the relationship between IDO1 expression level and fractions of immune
   cells, including naïve B and memory B cells (Fig. [86]1e). The
   IDO1-high group showed significantly lower proportions of naïve B cells
   (Fig. [87]1f), plasma and CD8^+ T cells (Fig. S2). On the contrary, the
   proportion of T cells suppressive monocytes was higher in the IDO1-high
   group (Fig. S2). Through Spearman correlation analysis, it was found
   that IDO1 gene expression level is negatively correlated and positively
   correlated with the proportions of naïve B and memory B cells in BMM,
   respectively (Fig. [88]1g, h). Then, we identified 71 significantly
   up-regulated genes in the IDO1-high group in AML patients
   (Fig. [89]1i). Several immune-related gene pathways were found
   (Fig. [90]1j), including immune system process, regulation of cytokine
   production, and inflammatory response, etc. Gene expression and
   clinical data from the TCGA LAML patients (n = 151) [[91]44] were used
   to successfully validate the prognosis significance of IDO1 and the
   correlations between IDO1 gene expression level and naïve and memory B
   cells proportions (Fig. S3).

   To validate the abnormal proportions of B cell subpopulations in the
   immune microenvironment of AML patients, we employed flow cytometry to
   analyze B cells subpopulations proportions in PBMCs samples from 3
   healthy individuals and 3 AML patients. The results revealed that the
   proportions of naïve B and memory B cells in PBMCs samples from AML
   patients were significantly lower than those in healthy individuals,
   while the proportion of Breg cells was significantly higher
   (Fig. [92]1k). In summary, these results suggested that IDO1 gene
   expression inversely correlated with AML patient survival and
   associated with distinct B cell subpopulation proportions.

IDO1 expression and activity correlated with B cell subpopulation proportions
in BMMCs and PBMCs of AML patients

   Although the mRNA expression of IDO1 has been recorded in the database,
   the protein expression of IDO1, particularly its pattern of expression
   in AML, remains inadequately characterized. Herein, two AML cell lines
   were chosen to investigate the expression dynamics of IDO1. As shown in
   Fig. [93]2a-b, both OA3 cells and K562 cells expressed IDO1, whereas
   IDO1 constitutive expression was higher in OA3 cells. IDO1 expression
   in both cell lines could be further induced by interferon-gamma
   (IFN-γ), and OA3 cells were more sensitive (Fig. [94]2c, d).

Fig. 2.

   [95]Fig. 2
   [96]Open in a new tab

   IDO1 constitutively expressed in AML cells and in BMMCs and PBMCs of
   AML patients, IDO1 expression correlated to B cell subpopulation
   proportions. a qPCR analysis of the mRNA expressions of IDO1 in K562
   and OA3 cells, n = 3. b Western blot analysis of the protein expression
   of IDO1 in K562 and OA3 cells. c, d Western blot analysis of the
   expression of IDO1 in OA3 and K562 cells incubated with IFN-γ. e,
   g Western blot analysis of the protein expression of IDO1 in BMMCs and
   PBMCs from 5 patients with AML. f, h HPLC analysis of Trp and Kyn
   levels in medium of BMMCs and PBMCs from 5 patients with AML. i, j Flow
   cytometry analysis of the proportions of total B cells (CD19^+), naïve
   B cells (CD19^+IgD^+CD27^−), memory B cells (CD19^+CD27^+) and Breg
   cells (CD19^+CD24^+CD27^+) in BMMCs and PBMC from 5 patients with AML.
   k, l ELISA analysis of IL-10 and TNF-α concentration in the medium of
   BMMCs and PBMCs from 5 patients with AML. Statistical significance was
   determined by Student’s t-test. Data were expressed as mean ± S.D.,
   *p < 0.05, **p < 0.01

   To investigate the expression pattern of IDO1 in AML and its impact on
   B cell subpopulation proportions, BMMCs and PBMCs samples from five
   patients with newly diagnosed AML were used. Based on the expression
   level of IDO1 in BMMCs and PBMCs, the patients were designated as
   Patient 1–5. Patient 1 had the lowest IDO1 expression, while Patient 5
   had the highest IDO1 expression (Fig. [97]2e, g). The variation of KP
   activity was consistent with that of the IDO1 expression. Patient 5 had
   the lowest Trp level, the highest Kyn level and Kyn/Trp ratio in both
   BMMCs and PBMCs (Fig. [98]2f, h). We defined Patients 1 and 2 as the
   IDO1-low group, and Patients 3, 4, and 5 as the IDO1-high group. The
   IDO1-low group had the higher proportions of total B and naïve B cells,
   and the lower proportions of memory B and Breg cells in both BMMCs and
   PBMCs, while the data of the IDO1-high group were the opposite
   (Fig. [99]2i, j). ELISA detection of immunosuppressive cytokines also
   revealed that IDO1-high group had the higher levels of interleukin 10
   (IL-10) and tumor necrosis factor alpha (TNF-α) in the BMMCs and PBMCs
   (Fig. [100]2k, l).

   Collectively, these results depicted that higher IDO1 expression in
   BMMCs and PBMCs of AML patients was possibly associated with lower
   proportions of total B and naïve B cells, and higher proportions of
   memory B and Breg cells.

IDO1 in AML cells affected the proportions and functions of B cell
subpopulations in the Co-culture of PBMCs and AML cells

   A co-culture of AML cells with healthy human PBMCs was established to
   simulate the tumour immune microenvironment and was used to test the
   effects of IDO1 in AML cells on B cell subpopulations including their
   proportions and functions. First, the B cell subpopulation proportions
   were compared among PBMCs alone, PBMCs co-cultured with OA3 cells,
   PBMCs co-cultured with IFN-γ treated OA3 cells, and PBMCs co-cultured
   with hIDO1-OE OA3 cells. As shown in Fig. [101]3a-d, both IFN-γ-induced
   and constitutive IDO1 expression in OA3 cells had significant effects
   on B cells and their subpopulations. The increase in IDO1 expression
   led to a decrease in the proportions of total B and naïve B cells, and
   an increase in the proportions of memory B and Breg cells. HPLC
   analysis also found that in the co-culture system containing IFN-γ
   treated OA3 cells and hIDO1-OE OA3 cells, the concentration of Trp was
   significantly reduced and KP activity was significantly increased
   (Fig. [102]3e). Recent studies have disclosed that the secretion of
   immunosuppressive cytokine IL-10 and IL-35 is contributed to the
   suppressive capability of Breg cells, and the ratio of IL-10 to TNF- α
   (IL-10/TNF-α) has been suggested serving as an anti-inflammatory index
   [[103]45–[104]47]. The levels of IL-35, IL-10, and TNF-α in PBMC were
   compared with different co-culture systems. As shown in Fig. [105]3f,
   the levels of TNF-α and IL-35 were higher in the co-culture of PBMCs
   and OA3 cells than in that of PBMCs alone, with the highest level in
   the co-culture of PBMCs and hIDO1-OE OA3 cells. Similarly, the level of
   IL-10 and IL-10/TNF-α were increased significantly in this co-cultured
   system.

Fig. 3.

   [106]Fig. 3
   [107]Open in a new tab

   In co-culture of PBMCs and AML cells, high IDO1 expression reduced the
   proportions of total B and naïve B cells, increased the proportions of
   memory B and Breg cells. PBMCs were co-cultured with OA3 cells, IFN-γ
   treated OA3 cells, and hIDO1-OE OA3 cells, stimulated by IL-4
   (10 ng/mL) and LPS (20 μg/mL) for 72 h. Ctrl represented PBMCs.
   a-d Flow cytometry analysis of the proportions of total B cells
   (CD19^+), naïve B cells (CD19^+IgD^+CD27^−), memory B cells
   (CD19^+CD27^+) and Breg cells (CD19^+CD24^+CD27.^+). Representative
   histograms and quantification were shown. n = 3 per group. e HPLC
   analysis of Trp and Kyn levels in the supernatants. n = 3 per group.
   f qPCR analysis of relative expression of IL-10, TNF-α, IL-10/ TNF-α
   ratio and IL-35. n = 3 per group. Statistical significance was
   determined by one-way ANOVA followed by Dunnett’s post hoc test Data
   were presented as the mean ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001

   With the co-culture of PBMCs and hIDO1-OE OA3 cells, the influence of
   IDO1 inhibitor on B cell subpopulations was investigated. First, the
   efficacy of epacadostat and RY103 on the blockade of KP were examined.
   The results indicated that either epacadostat or RY103 significantly
   suppressed the upregulated KP (decreasing Kyn levels and Kyn/Trp ratio)
   in hIDO1-OE OA3 cells (Fig. S5a), but they did not affect the KP
   activity in healthy PBMCs (Fig. S6a) and the co-culture system of PBMCs
   and OA3 cells (Fig. S7e). In addition, neither epacadostat nor RY103
   appeared to affect the viabilities of hIDO1-OE OA3 cells and healthy
   PBMCs (Fig. S5b, S6b). Then, the effects of epacadostat and RY103 on
   the B cells and its subpopulations in the co-culture were investigated.
   As shown in Fig. [108]4a-d, the supplement of epacadostat and RY103
   increased the proportions of total B and naïve B cells while decreased
   the proportions of memory B and Breg cells. On the contrary, IDO1
   inhibitors did not alter the proportions of B cell subpopulations in
   PBMCs alone (Fig. S6c-f) and the co-culture of PBMCs and OA3 cells
   (Fig. S7a-d), which confirmed the important immunoregulatory role of
   IDO1 on B cells. The levels of IL-10 and TNF-α as well as IL-10/TNF- α
   in the co-culture of PBMCs with hIDO1-OE OA3 cells were reduced by the
   treatment of epacadostat or RY103 (Fig. [109]4e, f).

Fig. 4.

   [110]Fig. 4
   [111]Open in a new tab

   IDO1 inhibitors increased the proportions of total B and naïve B cells
   and decreased the proportions of memory B and Breg cells in the
   co-culture of PBMCs and AML cells. PBMCs were co-cultured with hIDO1-OE
   OA3 cells stimulated by IL-4 (10 ng/mL) and LPS (20 μg/mL), in the
   absence or presence of IDO1 inhibitor (2 μM) for 72 h. Ctrl represented
   the co-culture of PBMCs and hIDO1-OE OA3 cells, Epa. represented
   epacadostat. a-d Flow cytometry analysis of the proportions of total B
   cells (CD19^+), naïve B cells (CD19^+IgD^+CD27^−), memory B cells
   (CD19^+CD27^+), Breg cells (CD19^+CD24^+CD27^+). Representative
   histograms and quantification of B cells were shown. n = 3 per group.
   e ELISA analysis of IL-10 concentration in the supernatants. n = 3 per
   group. f qPCR analysis of the levels of IL-10, TNF-α, and IL-10/TNF-α.
   n = 3 per group. Statistical significance was determined by one-way
   ANOVA followed by Dunnett’s post hoc test. Data were presented as the
   mean ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001

   These results indicated that high expression of IDO1 in AML cells
   reduced the proportions of total B and naïve B cells, increased the
   proportions of memory B and Breg cells, and suppressed the expression
   of cytokines in the co-culture of PBMCs and AML cells.

IDO1 inhibitors exhibited anti-leukemic effect in AML mice

   To validate the regulatory effect of IDO1 on the proportion and
   function of B cells in AML in vivo, we conducted experiments using
   C1498 cells in AML mouse model. C1498 cells are murine AML cells that
   can express IDO1 (Fig. [112]5a). We carried out the in vivo
   anti-leukemic assay of IDO1 inhibitors (1-L-MT and RY103) in C1498
   AML-bearing mice. (Fig. [113]5b). At the end of two-week treatment,
   IDO1 inhibitors restored the body weight loss caused by AML to a
   similar level of non-AML mice (Fig. [114]5d). Similarly, IDO1
   inhibitors reduced spleen weight (Fig. [115]5c). As a typical and
   diagnostic manifestation, the accumulation of leukemic blast cells in
   peripheral blood along with the decreased red blood cell count and
   mature neutrophil count occur in AML. As shown in Fig. [116]5e, the
   treatment with IDO1 inhibitors restored the elevated levels of leukemia
   blast cells and lower levels of red blood cells and mature neutrophils
   in AML model mice to the levels similar to non-AML. Next, the effects
   of RY103 and 1-L-MT on KP blockade in serum were assessed. As shown in
   Fig. [117]5f, compared to non-AML mice, serum Trp concentration was
   significantly reduced and the Kyn/Trp ratio was significantly elevated
   in AML mice. Meanwhile, serum Trp concentration in 1-L-MT group and
   RY103 group were higher and the Kyn/Trp ratio was lower compared to
   those in AML group. Taken together, our results indicated that IDO1
   inhibitors suppressed the tumor growth in C1498 AML-bearing mice.

Fig. 5.

   [118]Fig. 5
   [119]Open in a new tab

   IDO1 inhibitors inhibited tumor growth and ameliorated the abnormal
   proportions of B cell subpopulations in C1498 AML-bearing mice.
   a Western blot analysis of the protein expressions of IDO1 in C1498
   cells with or without IFN-γ (100 ng/mL) for 48 h. b Schematic
   representation illustrating the tumor inoculation in C57BL/6 female
   mice with 1 × 10^6 C1498 cells per mouse. The construction of mouse
   models and the designation of the different treatments were described
   in the Materials and Methods section. c Spleens isolated from mice were
   weighted, n = 3–5 per group. d Body weight was recorded every 36 h and
   statistical analysis was shown, n = 3–5 per group. e Representative
   images of morphology of peripheral blood cells after Wright-Giemsa
   staining from different groups, n = 3–5 per group. The red arrow
   represents red blood cells, the yellow arrow represents AML blast
   cells, and the blue arrow represents neutrophils. Three photos were
   taken from the peripheral blood smear of each mouse, The magnification
   of the microscope is 10 and 20 times, and the scales are 50 and 20 µm,
   respectively. f HPLC analysis of serum Trp and Kyn levels, and Kyn/Trp
   ratio was calculated, n = 3–5 per group. g ELISA analysis of serum
   IL-10 levels from different groups, n = 3–5 per group. h-i Flow
   cytometry analysis of the proportions of total B cells (CD19^+), naïve
   B cells (CD19^+IgD^+CD27^−), memory B cells (CD19^+CD27^+) and Breg
   cells (CD19^+CD24^+CD27.^+) in spleen (h) and bone marrow (i) from
   different groups, n = 3–5 per group. Statistical significance was
   determined by one-way ANOVA followed by Dunnett’s post hoc test. Data
   were presented as the mean ± SD. * p < 0.05, ** p < 0.01, *** p < 0.001

   Furthermore, to determine whether the suppression of tumor growth in
   C1498 AML-bearing mice caused by IDO1 inhibition was related to changes
   in the proportions and functions of B cell subpopulations, we examined
   the proportions of B cell subpopulations in spleen and BM and the
   concentration of IL-10 in serum. Our results showed that in the spleen,
   the proportions of total B and naïve B cells in AML mice decreased and
   the proportions of memory B and Breg cells increased significantly
   compared to non-AML mice. After treatment with IDO1 inhibitors, the
   abnormal proportions of total B, memory B and Breg cells were
   recovered, and the effect of RY103 was better between the two
   (Fig. [120]5h). Similar results were observed in the BM as in the
   spleen, and treatment with IDO1 inhibitors significantly recovered the
   abnormal proportions of naïve B, memory B, and Breg cells
   (Fig. [121]5i). At the same time, we found that the proportion of CD8^+
   T cells in AML-bearing mice was significantly reduced. However, IDO1
   inhibitors were unable to recover it, further confirming that IDO1
   inhibitors reversed the immunosuppressive state of mice by ameliorating
   the abnormal proportions of B cell subpopulations (Fig. S8). Besides,
   the concentration of IL-10 in the serum of AML mice was significantly
   higher than that of non-AML mice (Fig. [122]5g). It has also been shown
   that this increase was reversed by IDO1 inhibitions.

   In summary, our results indicated that IDO1 played a crucial role in
   AML progression, affecting B cell subpopulations proportions and
   cytokine levels, ultimately altering the immune status. IDO1 inhibitors
   showed promising anti-leukemic effects.

IDO1 potentially affected the proportions and functions of B cell
subpopulations in AML through PI3K-AKT signaling pathway

   Given that we had found in the above-mentioned studies that IDO1 in AML
   cells can affect the proportions and functions of B cell subpopulations
   in AML, we attempted to elucidate the specific mechanisms involved.
   RNA-sequencing analysis was performed and compared on healthy PBMCs and
   the co-culture of PBMCs with hIDO1-OE OA3 cells. 1843 differentially
   expressed genes were detected, of which 1043 were upregulated and 800
   were downregulated (Fig. [123]6a-b). Downregulated genes were
   predominantly clustered in genes related to lymphocytes, including
   SASH1, STAB1, CD69, and FOS. The Kyoto Encyclopedia of Genes and
   Genomes (KEGG) pathway enrichment analysis demonstrated that the genes
   with the most significant differential expression were annotated to the
   PI3K-AKT pathway. (Fig. [124]6c). Class IA PI3K activates the
   expression of B cell transcription factor Pax5, making it necessary for
   B cell differentiation [[125]48], and in a recently published work, the
   PI3K-AKT signaling pathway has been also found in age-related genes of
   AML patients, which involves inflammation and inhibitory immune
   microenvironment [[126]49]. In addition to the PI3K pathway, our
   results also identified several pathways related to immune responses,
   including the B cell receptor (BCR) signaling pathway and cytokine
   signaling pathway, as well as other pathways related to cell
   proliferation and differentiation. We validated through qPCR
   experiments that high IDO1 expression increased mRNA expression of
   PIK3CA, PIK3R1 subunits of PI3K kinase and Akt, while the addition of
   IDO1 inhibitors eliminated this phenomenon in the co-culture of PBMCs
   and hIDO1-OE OA3 cells (Fig. [127]6e).

Fig. 6.

   [128]Fig. 6
   [129]Open in a new tab

   IDO1 activated PI3K-AKT signaling pathway to affect the proportions and
   functions of B cell subpopulations in the co-culture of PBMCs and
   hIDO1-OE OA3 cells. RNA sequencing analysis was performed on PBMCs and
   PBMCs co-cultured with hIDO1-OE OA3 cells. a Differential expression
   gene clustering analysis heatmap. b Differentially expressed genes were
   presented as a volcano plot. The software for differential expression
   analysis is DESeq2, with a screening threshold of padj < 0.05 and
   |log2FoldChange|> 1. c KEGG Pathway enrichment analysis of
   differentially expressed genes. d GSEA Pathway enrichment analysis
   results. Perform GSEA enrichment analysis on GO and PATHWAY content
   using cluster Profile, respectively. e qPCR analysis of the levels of
   PIK3CA, PIK3R1, PDK1 and AKT in the PBMCs, PBMCs co-cultured with OA3
   and hIDO1-OE OA3 cells, PBMCs co-cultured with RY103 treated hIDO1-OE
   OA3 cells. Ctrl represented PBMCs. n = 3 per group. Statistical
   significance was determined by one-way ANOVA followed by Dunnett’s post
   hoc test. Data were presented as the mean ± SD. * p < 0.05, **
   p < 0.01, *** p < 0.001

   GSEA enrichment analysis was performed on KEGG Pathway, and it was
   found that high expression of IDO1 was negatively correlated with
   positive regulation of B cell activation pathway, BCR signaling
   pathway, and immune response pathway (Fig. [130]6d). GSEA enrichment
   was performed on Gene Ontology (GO) terms, and the results also showed
   significant enrichment of pathways related to tryptophan metabolism and
   B cell receptor signaling pathways.

Discussion

   B cell subpopulations imbalance has been observed in PBMCs and BMMCs of
   AML patients [[131]9, [132]50], which may contribute to the
   pathological progression and immune evasion of AML, but the specific
   molecular mechanism behind has not been elucidated. We speculate that
   the imbalance of B cell subpopulations is potentially related to IDO1
   that has been suggested to be involved in AML immune evasion.

   In the current research, we analyzed the correlation between IDO1
   expression and the proportions of B cell subpopulations using AML
   patient samples at both bioinformatics and clinical data levels. We
   found that IDO1 expression and activity were negatively associated with
   the proportions of total B and naïve cells, and positively correlated
   with the proportions of memory B and Breg cells (Figs. [133]1, [134]2).
   Existing studies only focus on the relationship between IDO1 and AML
   prognosis, as well as that between B cells subpopulations imbalance and
   AML, but not the relationship between IDO1 and the imbalance of B cell
   subpopulations. We demonstrated for the first time that IDO1 expression
   corelated with the proportions of B cell subpopulations.

   Subsequently, we also validated the relationship between IDO1 and B
   cells in vitro. We used co-culture of healthy human PBMCs and AML cells
   to better simulate the tumor microenvironment. The results showed that
   high IDO1 expression could reduce the proportions of total B and naïve
   B cells, increase the proportions of memory B and Breg cells, and the
   levels of immunosuppressive cytokines (Fig. [135]3). The supplement of
   IDO1 inhibitors epacadostat and RY103 could restore these abnormal
   proportions of B cells subpopulations (Fig. [136]4). Co-culture of AML
   cells and immune cells such as macrophages and T cells have been
   usually utilized [[137]51]. In this study, we first used co-culture
   models to study the proportions of B cell subpopulations. However, this
   co-culture model still had some limitations. Since the two types of
   cells were physically in contact with each other, it was difficult to
   separate each types of cells for further downstream assays [[138]51].

   We further investigated the signaling pathways that may be involved in
   the regulation of IDO1 on proportions of B cell subpopulations through
   RNA-seq, and found that the PI3K-AKT pathway had the most significant
   correlation with IDO1 (Fig. [139]6). It has been reported that the
   PI3K-AKT pathway has a crucial role in the progression of AML,
   promoting the proliferation and spread of AML blast cells through
   different mechanisms such as inhibiting cell apoptosis, enhancing the
   migration and invasion ability of AML blast cells [[140]52, [141]53].
   It has also been shown that IDO1 binds to the PI3K p85 subunit via a
   YxxM motif [[142]54]. Our results connected IDO1 with the PI3K pathway,
   expanding the knowledge on regulatory role of IDO1 in AML.

   IDO inhibitor has been a research and development pipeline of
   well-known pharmaceutical companies in the past few years.
   Approximately 100 clinical trials of at least 8 small molecule IDO1
   inhibitors aiming to characterize the safety and efficacy profile have
   been initiated and the majority of them focused on solid tumors
   [[143]11]. It should be emphasized that very few research has focused
   on AML. In our study, the anti-tumor activity of IDO1 inhibitors RY103
   and 1-L-MT were tested in C1498 AML bearing mice. The results showed
   that RY103 had better ability to block KP and better anti-tumor
   efficacy than 1-L-MT in AML mice. RY103 reversed the immunosuppressive
   state of AML mice by ameliorating the abnormal proportions of B cell
   subpopulations (Fig. [144]5). Previous reports have shown that various
   lymphocytes including NK cells and T cells are inhibited in AML’s BMM
   [[145]55–[146]57]. Our study revealed for the first time the effect of
   IDO1 on the proportions and functions of B cell subpopulations,
   providing a theoretical basis for IDO1 inhibitor therapy for AML.

   Our inability to utilize more AML patient samples and AML cell lines
   indicates that our research results are not perfect. Besides, our
   future study may explore the mechanism by which IDO1 regulates B cell
   differentiation via PI3K-AKT pathway.

   In summary, our study revealed the impact of IDO1 on both the
   proportions of B cell subpopulations and immunosuppressive cytokines
   within BMM and therefore provided new evidence for the contribution of
   IDO1 in immune evasion, extending beyond its well-recognized impact on
   the T cells in AML (Fig. [147]7). Our study elucidated the potential
   therapeutic potential of IDO1 inhibitor as a valuable treatment of AML,
   which warranted the proof from clinical study.

Fig. 7.

   [148]Fig. 7
   [149]Open in a new tab

   The schematic demonstration of IDO1 expression contributes to the
   malignancy of acute myeloid leukemia by affecting the proportions of B
   cell subpopulations. IDO1 in AML blast cells depletes Trp and
   accumulates Kyn in the TME, thereby altering the proportions and
   functions of B cell subpopulations, specifically manifested as a
   decrease in the proportion of naïve B cells, an increase in the
   proportions of memory B and Breg cells and the expression of
   immunosuppressive cytokines

Materials and methods

Bioinformatic analysis

   Previously published 1,007 adults AML from Ruijin Hospital Affiliated
   to Shanghai Jiao Tong University School of Medicine [[150]49] were
   included in this study. Available prognosis data was obtained in more
   than 99% (n = 997) patients. To validate the conclusions, we also
   reproduced the core results based on the public data from TCGA LAML
   patients (n = 151). The gene expression level was normalized using TPM
   (transcripts per million) measures. Using the Surv-Cutpoint function in
   the survminer package of R, which repeatedly tested all potential cut
   points to establish the maximum rank statistic, the patients were
   divided into two groups: the high IDO1 expression group (IDO1-high) and
   the low IDO1 expression group (IDO1-low). Immune cell deconvolution was
   performed using the CIBERSORTx software [[151]58]. Spearman correlation
   analysis was used to infer potential interactions between
   gene-immunocyte pairs. Differentially expressed genes were calculated
   using DESeq2 variance stabilizing transformation (VST) normalization
   and the Limma package. The gene sets enrichment analysis was performed
   using the clusterProfiler [[152]59] and Broad GSEA programs (adjusted
   p < 0.05 and log2 (fold change) > 1).

Human samples

   BM and PB samples were obtained from 5 patients with non-promyelocytic
   AML at Ruijin Hospital Affiliated to Shanghai Jiao Tong University
   School of Medicine and used to isolate BMMCs and PBMCs. PBMCs of
   healthy individuals were purchased from Milestone Biotechnologies
   company. The baseline and disease characteristics information was
   provided in Table. S1 (for Supplementary information).

Cell culture

   OCI-AML3 (human AML cell line, OA3) and K562 (human chronic myeloid
   leukemia cell line) were procured from Shanghai Institute of
   Hematology. C1498 (mouse AML cell line) was purchased from Haixing
   Biosciences Co., Ltd. OA3, K562 and C1498 cells were cultured in RPMI
   1640 (Gibco, Cat# 31800022) supplemented with 10% FBS, 100 U/mL of
   penicillin (Aladdin, Cat# S432673) and 100 μg/mL of streptomycin
   (Aladdin, Cat# S432673). Cells were grown at 5% CO2 in a humidified
   incubator at 37℃.

In vitro PBMCs co-culture with AML cells

   Healthy human PBMCs were co-cultured with OA3 cells stably
   over-expressing IDO1 (hIDO1-OE OA3) or WT OA3 cells to simulate the
   tumor immune microenvironment, and the ratio of PBMCs to AML cells in
   the co-culture system was 5:1, and the system was stimulated by
   recombinant human interleukin (IL)−4 Protein (Yeasen, Cat# 90105ES08)
   at 10 ng/ml and lipopolysaccharide (LPS) (Sigma, Cat# L2880) at
   20 μg/ml for 72 h.

Reagents

   HPβCD ((2-hydroxypropyl)-β-cyclodextrin, Aladdin, Cat# H108813), 1-L-MT
   (Sigma-Aldrich, Cat# 447,439), epacadostat (INCB024360, Selleck, Cat#
   S7910), murine Interferon-γ (IFN-γ, Beyotime, Cat# P6137), human IFN-γ
   (Yeasen, Cat# 91211ES10) were used in this study. RY103 was designed
   and synthesized by our lab.

Cell treatment conditions

   OA3 and K562 cells: incubated with 100 ng/mL of human IFN-γ for
   24/48/72/120 h.

   C1498 cell: incubated with 100 ng/mL of murine IFN-γ for 48 h.

   hIDO1-OE OA3, PBMCs co-culture with hIDO1-OE OA3 cells: RY103
   (incubated with 2 μM of RY103 for 72 h); epacadostat (incubated with
   2 μM of epacadostat for 72 h).

Animal model and treatment

   Female, 4 to 5-week-old, C57BL/6 mice were purchased from Shanghai
   ShengChang.

   The mice were divided into non-AML and AML groups. To construct AML
   mice, C1498 cells (1 × 10^6 cells /mouse) were intravenously injected.
   Beginning on day 6 after the implantation, mice were randomized and
   given the treatments intraperitoneally (i.p.) on the same day.
   Treatments were listed as follows: non-AML and AML mice received 10%
   HPβCD. Mice RY103 and 1-L-MT groups received 6 mg/kg of RY103 in 10%
   HPβCD and 100 mg/kg of 1-L-MT in 10% HPβCD, respectively every 36 h.
   Once the two-week treatment was completed, the mice were sacrificed.
   The spleen was isolated and weighted, and the BM was collected. B cells
   in spleen and BM were detected by flow cytometry.

High-performance liquid chromatography (HPLC) analysis of Trp and Kyn

   The cell culture supernatant and mice serum were treated with 5%
   perchloric acid precipitation protein method. After centrifugation, the
   supernatant was taken and the content of Trp and Kyn was detected by
   HPLC based on the retention time and the UV absorption (280 nm for Trp,
   360 nm for Kyn) [[153]25]. (Kyn/Trp) × 100 reflects the activity of
   IDO1.

RT-PCR and quantitative real-time PCR (qPCR)

   qPCR was performed by using the SYBR Green PCR Master Mix kit
   (ABclonal, Cat# RK21203) for the detection of IDO1, IL-10, TNF-α,
   IL-35, PIK3CA, PIK3R1, PDK1, AKT, β-actin mRNA expression. The primers
   used for the qPCR are shown in Table S2.

Western blot (WB)

   Protein samples obtained from AML cells were transferred onto PVDF
   membranes after separated by SDS-PAGE. The following primary antibodies
   against anti-IDO1 (1:3000, Proteintech, Cat# 66,528–1-Ig), anti-GAPDH
   (1:10,000, Proteintech, Cat# 60,004–1-Ig) and anti-β-actin (1:5000,
   ABclonal, Cat# AC004) antibodies were used. The proteins were treated
   with an HRP-conjugated secondary antibody and were visualized using ECL
   reagents (Tanon, Cat# 1,805,001).

Enzyme-linked immunosorbent assay (ELISA)

   The concentrations of IL-10 in the medium and serum were measured by
   ELISA kit (HUABio, Cat# EM0005 and EH0009). All procedures were
   performed following the product manual.

Cell counting kit-8 (CCK-8) assay

   After treating the cells according to the above “Materials and Methods
   4.6”, the cell culture medium was replaced with fresh culture medium
   containing 10% CCK-8 solution. After culturing the cells for 4 h, the
   absorbance at a wavelength of 450 nm was measured using a microplate
   reader.

Flow cytometry analysis

   Cells from co-culture system, BMs and spleens of mice were labeled with
   the following fluorescence-conjugated antibodies purchased from
   Biolegend: PerCP/Cyanine5.5 anti-human CD19 (Cat# 302,230), FITC
   anti-human IgD (Cat# 348,206), PE anti-human CD24 (Cat#311,106),
   PE/Cyanine7 anti-human CD27 (Cat#302,838), PerCP/Cyanine5.5 anti-mouse
   CD19 (Cat#152,405), FITC anti-mouse IgD (Cat# 405,703), PE anti-mouse
   CD24 (Cat#101,807), PE/Cyanine7 anti-mouse CD27 (Cat#124,215), FITC
   anti-mouse CD45 (Cat# 103,108), APC anti-mouse CD3 (Cat#100,235), PE
   anti-mouse CD8α (Cat#100,707). Zombie NIR Fixable Viability Kit
   (Biolegend, Cat# 423,105) was used to label viable cells. Cells with
   different labels were detected using the Beckman Coulter’s Gallios flow
   cytometer, and the data were analyzed using FlowJo software. Flow
   cytometry gating strategies were shown in Fig. S4 (for Supplementary
   information).

Wright–Giemsa staining

   The slides were soaked in 10% hydrochloric acid for 24 h in advance and
   washed thoroughly to dry. 10 μL of fresh mouse blood was evenly spread
   on a glass slide and placed in a 37 °C incubator to keep warm and dry.
   The rapid Wright-Giemsa staining solution (Yisheng, Cat# 60529ES01) was
   used for post-staining observation. All procedures were performed
   following the product manual.

RNA sequencing

   RNA from the co-culture of PBMCs and IDO1-OE OA3 cells (experimental
   group) and PBMCs alone (control group) were extracted (n = 3 in each
   group). Library construction and sequencing were performed on an
   Illumina platform by Wuhan Benagen Technology. The raw data were first
   processed with FastQC (v0.11.9, default) to filter out adapters and
   low-quality sequences. Then, clean reads were mapped to the human
   genome using Star (v2.7.9a, default). Gene expression level for each
   sample were calculated using RSEM (v1.3.3, default), and expressed as
   fragments per kilobase of transcript per million fragments mapped. The
   gene expression levels in stressed samples were analyzied againt those
   in control samples to identify the differentially expressed genes
   (DEGs) using DESeq2 (v1.34.0, default). The DEGs were identified as
   previously described, using the parameters: Foldchange ≥ 2.00 and
   Probability ≥ 0.8 with a significant false discovery rate-adjusted P
   value < 0.05 based on the three biological replicates. GO and KEGG
   enrichment analyses for the DEGs were conducted using cluster Profiler
   version 3.8.

Statistical analysis

   All experiments were repeated at least three times. Prism 10 software
   (GraphPad Software) was used to create graphs and perform statistical
   analyses. Data were expressed as the means ± SD. One-way analysis of
   variance (ANOVA) was used to compare several treatment groups with
   control groups. Student’s t-test was used to determine the difference
   between two groups. Significance values were set at *p < 0.05,
   **p < 0.01, ***p < 0.001 and ****p < 0.0001.

Supplementary Information

   [154]43556_2025_262_MOESM1_ESM.docx^ (3.2MB, docx)

   Supplementary Material 1: Fig. S1. Differential gene comparison between
   IDO1-low and high groups of AML patients from Ruijin hospital. Each row
   represents a clinical or genomic characteristic. Each column represents
   one patient. The patients are divided into left (low) and right (high)
   panels according to the gene expression level of IDO1. The cutoff of
   low and high is determined by the prognosis significance. The right
   percentages and bar plots show the overall mutation rate and the number
   of mutations (genes or pathways). Fig. S2. Other significant difference
   in immune fractions between IDO1-low and high groups of AML patients
   from Ruijin hospital. The statistical significance is calculated based
   on the Wilcoxon Rank Sum Test. Fig. S3. Correlations between IDO1 gene
   expression and clinical prognosis, and between IDO1 gene expression and
   proportions of naïve/memory B cells in the TCGA LAML cohort. (a)
   Kaplan-Meier survival curves of OS of the AML patients according to
   IDO1 mRNA expression level in bone marrow specimens. IDO1-low group (n
   = 71) and IDO1-high group (n = 80) were compared using the log-rank
   test; (b) Forest plot of multivariant analysis of basic clinical
   information and IDO1 expression groups; (c) The correlations between
   IDO1 gene expression and relative proportions of 22 cell types in the
   BMM of AML patients from the TCGA LAML database using CIBERSORT method
   coupled with LM22. Fork marks indicate statistical significance below
   0.05; (d) Scatter plots of the proportions of naïve (left), memory B
   cells (right) and IDO1 gene expression level (log2 (TPM+1)). Fig. S4.
   Flow cytometry gating strategy. (a) FCM gating strategy for detection
   of total B cells (CD19^+) cells, naïve B cells (CD19^+IgD^+CD27^-),
   memory B cells (CD19^+CD27^+), Breg cells (CD19^+CD27^+CD24^+). (b) FCM
   gating strategy for detection of CD8^+ T cells (CD45^+CD3^+CD8^+). All
   gates were set using fluorescence-minus-one (FMO) controls. Fig. S5.
   IDO1 inhibitors significantly inhibited the upregulated KP without
   cytotoxicity in hIDO1-OE OA3 cells. (a) HPLC analysis of Trp and Kyn
   levels in medium from OA3 cells (Ctrl), hIDO1-OE OA3 cells and IDO1
   inhibitor (2 μM) treated hIDO1-OE OA3 cells. (b) CCK-8 analysis of
   cytotoxicity of IDO1 inhibitor on hIDO1-OE OA3 cells (Ctrl). Epa.
   represented epacadostat. Statistical significance was determined by
   one-way ANOVA followed by Dunnett’s post hoc test. Data were presented
   as the mean ± SD. * p< 0.05, ** p < 0.01, *** p < 0.001. Fig. S6. IDO1
   inhibitor did not affect the basal KP and the proportions of B cell
   subpopulations without cytotoxicity in healthy human PBMCs. PBMCs were
   treated with IDO1 inhibitor (2 μM unless stated otherwise) stimulated
   by IL-4 (10 ng/mL) and LPS (20 μg/mL) for 72 h, Epa. represented
   epacadostat. (a) HPLC analysis of Trp and Kyn levels in medium. (b)
   CCK-8 analysis of the viability of PBMCs. (c-f) Flow cytometry analysis
   of the proportions of total B cells (CD19^+), naïve B cells
   (CD19^+IgD^+CD27^-), memory B cells (CD19^+CD27^+) and Breg cells
   (CD19^+CD24^+CD27^+). Fig. S7. RY103 only changed the proportions of B
   cell subpopulations and KP activity in the co-culture of PBMCs and
   hIDO1-OE OA3 cells. PBMCs were co-cultured with OA3 cells or hIDO1-OE
   OA3 cells stimulated by IL-4 (10 ng/mL) and LPS (20 μg/mL), in the
   absence or presence of RY103 (2 μM) for 72 h. (a-d) Flow cytometry
   analysis of the proportions of total B cells (CD19^+), naïve B cells
   (CD19^+IgD^+CD27^-), memory B cells (CD19^+CD27^+) and Breg cells
   (CD19^+CD24^+CD27^+). Representative histograms and quantification were
   shown. n=3 per group. (e) HPLC analysis of Trp and Kyn levels in
   medium. Statistical significance was determined by one-way ANOVA
   followed by Dunnett’s post hoc test for comparison. Data were presented
   as the mean± SD. * p < 0.05, ** p < 0.01, *** p < 0.001. Fig. S8. IDO1
   inhibitor did not alter the proportion of CD8^+ T cells in the spleen
   and bone marrow of AML mice. The construction of mouse models and the
   assignment of the different treatments were described in the Materials
   and Methods section. (a-b) Flow cytometry analysis of the proportion of
   CD8^+T cells (CD45^+CD8^+) in spleen and BMMCs, n=5 per group.
   Representative histograms and quantification were shown. Statistical
   significance was determined by one-way ANOVA followed by Dunnett’s post
   hoc test. Data were presented as the mean ± SD. * p < 0.05, ** p< 0.01,
   *** p < 0.001. Table S1. Clinical characteristics of 5 AML patients
   from Ruijin Hospital included in this study. Table S2. PCR Primers for
   quantification. Table S3. Clinical and molecular features in IDO1-low
   and high groups of AML patients from Ruijin hospital.
   [155]Supplementary Material 2.^ (316.8KB, docx)

Acknowledgements