Distinct and targetable role of calcium-sensing receptor in leukaemia

Abstract

Haematopoietic stem cells (HSC) reside in the bone marrow microenvironment (BMM), where they respond to extracellular calcium [eCa²⁺] via the G-protein coupled calcium-sensing receptor (CaSR). Here we show that a calcium gradient exists in this BMM, and that [eCa²⁺] and response to [eCa²⁺] differ between leukaemias. CaSR influences the location of MLL-AF9⁺ acute myeloid leukaemia (AML) cells within this niche and differentially impacts MLL-AF9⁺ AML versus BCR-ABL1⁺ leukaemias. Deficiency of CaSR reduces AML leukaemic stem cells (LSC) 6.5-fold. CaSR interacts with filamin A, a crosslinker of actin filaments, affects stemness-associated factors and modulates pERK, β-catenin and c-MYC signaling and intracellular levels of [Ca²⁺] in MLL-AF9⁺ AML cells. Combination treatment of cytarabine plus CaSR-inhibition in various models may be superior to cytarabine alone. Our studies suggest CaSR to be a differential and targetable factor in leukaemia progression influencing self-renewal of AML LSC via [eCa²⁺] cues from the BMM.

Fig. 1. Calcium ions as features of the normal and leukaemic BMM.

a Schematic diagram (Created with BioRender.com) of the calcium indicator GCaMP6s, cloned into the sleeping beauty transposon plasmid pSBbi-Pur expressing RFP (left). The mechanism of GFP fluorescence after the binding of calcium ions to calmodulin and the conformational change of calmodulin and M13 peptide are shown on the right. b Representative two-photon intravital microscopy image of the calvarium of an unirradiated Rag-2^−/− IL2R γ^−/− CD47^−/− mouse, which had been transplanted with 5 × 10⁵ BA/F3 cells expressing GCaMP6s 2 h prior to imaging. The image is representative of 3 independent experiments, and the scale bar represents 50 μm. Bone fluoresces blue due to second harmonic generation, the transplanted BA/F3-GCaMP6s cells fluoresce green. c Correlation of the integrated GFP intensity of BA/F3-GCaMP6s cells with the distance to the endosteum (μm), as measured by intravital microscopy. In stacks of 7−17 images of an individual BA/F3-GCaMP6s cell, the GFP intensity of the middle stack was used as representative value. Pearson’s correlation coefficient (r) and the P value are shown (P = 0.0405, n = 34). The n represents 34 individual cells which were analysed. The experiments were repeated three times. d Quantification of the calcium concentration (μg/μl) present in the bone marrow (BM) of irradiated BALB/c mice transplanted with empty vector control-transduced BM or moribund mice with chronic myeloid leukaemia (CML), acute myeloid leukaemia (AML) or B-cell acute lymphoblastic leukaemia (B-ALL) shortly before death. All diseases were induced with the retroviral transduction/transplantation model (n = 5 individual mice per group, one-way ANOVA, Tukey test, mean ± SD). e Concentration of calcium (μg/μl) present in the heparinised BM plasma of patients (AML (n = 7); non-AML: CML (n = 1), myelodysplastic syndrome (MDS) (n = 1), ALL (n = 2), diffuse large B-cell lymphoma (DLBCL) (n = 1), P = 0.0086, two-tailed t test, mean ± SD). n refers to individual patients. Source data are provided as a Source Data file.

Fig. 2. Calcium and CaSR impact leukaemia cell function.

a Percentage of CaSR⁺ cells of all cells in different human leukaemia cell lines (one-way ANOVA, Dunnett test, mean ± SD) (K562: BCR-ABL1⁺ (CML, n = 9), NALM-6: B-ALL (n = 3), THP1: MLL-AF9 (AML, n = 8), Kasumi: AML1-ETO (AML, n = 3)). n refers to biological replicates. b, c Representative immunoblot (b) and its quantification (c) (mean ± SD) for CaSR and GAPDH expression in lysates of THP1 cells treated with 0–7.5 mM of CaCl₂ for 4 h. The blot is representative of three independent experiments. d, e Representative immunoblot (of three independent experiments) (d) and its quantification (e) (mean ± SD) for CaSR and GAPDH expression in K562 cells treated with CaCl₂. f, g Calcium flux analysis of THP1 (f) and K562 (g). h, i Number of THP1 (h) or K562 (i) cells adhering to fibronectin (FN). Prior to adhesion, cells had been treated with calcium (n = 6, one-way ANOVA, Dunnett test, mean ± SD). j, k Number of THP1 (j) or K562 (k) cells which migrated towards C-X-C motif chemokine 12 (CXCL12). Cells had been pre-treated with calcium (n = 4, one-way ANOVA, Dunnett test, mean ± SD). From h–k, n refers to biological replicates. l–o Distance (in µm) of intravenously transplanted CMTMR-labelled THP1 (l, m) or K562 (n, o) cells to the endosteum in the calvarium of NOD SCID interleukin-2 receptor γ (IL2Rγ)^−/− (NSG) mice. NTC (non-target control) (black) versus CaSR knockout (KO) (blue) cells (l, n) and THP1 WT (black) versus CaSR-overexpressing (OE) THP1 or K562 (red) (m,o) cells had been transplanted and imaged by intravital microscopy (two-tailed t test, mean ± SD) (n = 30 in (l), n = 29 (THP1 WT), n = 30 (THP1 CaSR OE) in m, n = 27 (K562 NTC), n = 53 (K562 CaSR KO) in n, n = 40 (K562 WT), n = 50 (K562 CaSR OE) in o. p, q Distance of transplanted THP1 (p) or K562 (q) cells to the endosteum. Cells had been pre-treated with NPS-2143. Experiments were performed in three different mice per cohort (n = 35 (vehicle), n = 38 (NPS-2143) in p, n = 30 (vehicle), n = 42 (NPS-2143) in q). From l–q n refers to individual cells. Data are presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 3. CaSR alters leukaemia progression and regulates self-renewal of AML LSC.

a, b Kaplan–Meier style survival curves of mice transplanted with MLL-AF9⁺ (a; n = 9 (wildtype), n = 10 (CaSR KO), Log-rank test)) or BCR-ABL1⁺ (b; (n = 5 (wildtype), n = 7 (CaSR KO), Log-rank test)) wildtype (black) or CaSR KO (blue) BM. c, d Kaplan–Meier style survival curves of BALB/c mice transplanted with wildtype MLL-AF9⁺ (c) or BCR-ABL1 + -BM (d), co-transduced with empty vector- (black) or CaSR-overexpressing retrovirus (red) (n = 6 (empty vector), n = 8 (CaSR OE), Log-rank test). e Kaplan–Meier style survival curve of wildtype mice transplanted with wildtype or CaSR KO MLL-AF9⁺ BM. The wildtype donor BM was empty vector⁺ (black), and the CaSR KO BM was empty vector⁺ (blue) or overexpressed (OE) CaSR (red) (n = 16 (wildtype+empty vector), n = 16 (CaSR KO+empty vector), n = 8 (CaSR KO+CaSR OE), Log-rank test). f Kaplan-Meier-style survival curve of wildtype secondary mice transplanted with sorted MLL-AF9⁺ (GFP⁺) Lin⁻ BM cells from primary recipients of MLL-AF9-transduced wildtype or CaSR KO BM (n = 10 (wildtype), n = 11 (CaSR KO), Log-rank test). From a–f n refers to individual mice. g Limiting dilution transplantation analysis for secondary wildtype recipient mice transplanted with sorted wildtype or CaSR KO AML cells (n = 4–5 mice per recipient group, Poisson statistics). h, i Colonies from the BM (h) or spleen (i) of wildtype mice transplanted with wildtype (black) or CaSR KO (blue) MLL-AF9-transduced BM (n = 4, two-tailed t test, mean ± SD). j Colonies from the serial replating of BM-derived cells from wildtype mice transplanted with MLL-AF9-transduced wildtype (black) or CaSR KO (blue) BM (n = 3, two-way ANOVA, Sidak’s test, mean ± SD). k Relative expression of stemness- and self-renewal-related genes in unsorted BM cells from mice transplanted with MLL-AF9-transduced wildtype, (black) CaSR KO (blue) and CaSR OE (red) leukaemia-initiating cells (n = 3 (CaSR KO, CaSR OE), n = 6 (wildtype), two-way ANOVA, Sidak’s test, mean ± SD). From h–k n refers to biological replicates. l Volcano plot summarising all differentially expressed genes in sorted MLL-AF9⁺ (GFP⁺) Lin⁻ cells in the BM of wildtype recipient mice transplanted with MLL-AF9-transduced wildtype BM co-transduced with empty vector- or CaSR OE-expressing retrovirus. Upregulated versus downregulated genes are represented in red and blue, respectively. Data are presented as mean values ± SD. Source data are provided as a Source Data file. The schematics in a–g and j were created with BioRender.com.

Fig. 4. CaSR activates MAPK and Wnt-β-catenin.

a, b Differentially expressed genes (a) and proteins (b) in sorted Lin⁻ MLL-AF9⁺ cells from the BM of WT (Mx1-Cre^−/− CaSR ^flox/flox) mice transplanted with WT or CaSR KO (Mx1-Cre^± CaSR ^flox/flox) MLL-AF9⁺ leukaemia-initiating cells. Each protein is represented by a dot, which is mapped according to its statistical significance (P value) versus magnitude of change (fold change). c, d Expression of ERK 1/2 and phospho ERK 1/2 (pERK1/2 (Thr 202/Tyr 204)) in lysates of THP1 non-target control (NTC) versus THP1 CaSR KO (c) and THP1 WT versus THP1 CaSR OE cells (d). The images are representative of three (c) or 6 (d) independent experiments. e, f Relative expression of pERK1/2 target genes in THP1 NTC versus THP1 CaSR KO cells (e) (n = 8, two-way ANOVA, Sidak’s test, mean ± SD) and THP1 WT versus THP1 CaSR OE cells (n = 9, two-way ANOVA, Sidak’s test, mean ± SD) (f). n refers to biological replicates. g Expression of β-catenin and c-MYC in lysates of THP1 WT and THP1 CaSR KO cells. The image is representative of eight independent experiments. h Expression of β-catenin (left), c-MYC and cyclin D1 (right) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) or vinculin, respectively, in lysates of THP1 WT versus THP1 CaSR OE cells. The image is representative of nine independent experiments. Source data are provided as a Source Data file.

Fig. 5. FLNA, downstream of CaSR, may play a role in AML progression.

a Immunoblot for CaSR and FLNA in lysates of K562 and THP1 cells overexpressing CaSR, in which an anti-FLNA antibody was used to co-immunoprecipitate CaSR. The image is representative of 2 independent experiments. b Immunofluorescence staining for CaSR (purple) and FLNA (green) in sorted MLL-AF9⁺ (GFP⁺) Lin⁻ cells from WT mice with AML. The scale bar represents 5 μm. c Expression of FLNA in lysates of unsorted, total BM cells from individual WT mice transplanted with WT BM transduced with MLL-AF9 and co-transduced with empty vector- or CaSR-overexpressing retrovirus. Each column represents a single mouse. The image is representative of three biological replicates. d Expression of FLNA in lysates of non-target control (NTC), CaSR KO, WT and CaSR OE THP1 cells. The image is representative of 4 (NTC vs CaSR KO) and 7 (WT vs CaSR OE) independent experiments. e Kaplan–Meier style survival of NOD SCID interleukin (IL)-2 receptor γ KO (NSG) recipient mice transplanted with THP1 NTC (black) or THP1 FLNA KO (purple) cells (n = 8 (individual mice), Log-rank test). f Log2 expression of FLNA in sorted HSC from healthy individuals (n = 6) or unsorted BM cells from patients with AML t(15;17) (n = 54), AML inv(16)/t(16;16) (n = 47), AML t(8;21) (n = 60), MLL-rearranged AML (t(11q23)/MLL) (n = 43) or AML with complex karyotype (n = 46) (one-way ANOVA, Tukey test). g Kaplan–Meier style survival curve of patients with AML with above average (n = 96) or below average (n = 66) expression of FLNA (Log-rank test). From f, g n refers to individual patients. h, i Calcium flux represented by quantification of the area under the curve (AUC) at baseline in THP1 non-target control (NTC) (black) versus CaSR knockout (KO) (blue) cells (h) (n = 6, two-tailed t test) and THP1 WT (black) versus THP1 CaSR-overexpressing (OE) cells (red) i (n = 8, two-tailed t test). n represents biological replicates. Data are presented as mean values ± SD. Source data are provided as a Source Data file.

Fig. 6. Possible targeting of CaSR in AML.

a Percentage of WT THP1 cells of single cells, which are live, apoptotic or necrotic after treatment with the CaSR inhibitor NPS-2143 (n = 3 (biological replicates), two-way ANOVA, Dunnett’s test, mean ± SD). b–d Expression of β-catenin (b), cyclin D1 (c), and c-MYC (d) in WT or CaSR OE THP1 cells after treatment with NPS-2143. The image is representative of 7 independent experiments. e, f Schematic representation (e, (Created with BioRender.com)) and Kaplan–Meier style survival curve (f) of recipient WT BALB/c mice with MLL-AF9-induced AML treated with vehicle (black), NPS-2143 (red; 2 mg/kg, administered from day 12 to day 50 after transplantation as a daily regimen), ara-C (green; 50 mg/kg, administered on five consecutive days per cycle for three cycles every two weeks, starting on day 18 after transplantation) and the combination of both NPS-2143 and ara-C (blue) (as above) (n = 8, Log-rank test). n refers to individual mice. g Treatment scheme (Created with BioRender.com) for NOD SCID interleukin-2 receptor γ knockout (NSG) mice transplanted with BM cells from five individual patients with AML. Each sample was transplanted into 4–7 mice, whereby some mice were treated with vehicle, ara-C (50 mg/kg, administered for three consecutive days per cycle, for three cycles every two weeks) or the combination of ara-C (as above) and NPS-2143 (2 mg/kg, administered as a daily regimen). h, i White blood cell count (WBC) (h) and percentage of human CD45⁺ leukocytes (i) in the peripheral blood of NSG mice transplanted with human AML cells as in g and treated with vehicle (n = 9), ara-C (n = 10) or the combination of ara-C and NPS-2143 (n = 11), 30 days post transplantation. Recipients of the same human AML sample are indicated by the same colour (Kruskal–Willis, Dunn’s test, mean ± SD). j Kaplan–Meier style survival curve of NSG mice transplanted with human AML cells as in g–i and treated with vehicle (black), ara-C (green) or the combination of ara-C and NPS-2143 (blue) (n = 9 (vehicle), n = 10 (ara-C), n = 11 (NPS-2143), Log-rank test). From h–j, n refers to individual mice. Data are presented as mean values ± SD. Source data are provided as a Source Data file.