Dendritic Cell-Specific Transmembrane Protein (DC-STAMP) Regulates Osteoclast Differentiation via the Ca2+ /NFATc1 Axis

Abstract

DC-STAMP is a multi-pass transmembrane protein essential for cell-cell fusion between osteoclast precursors during osteoclast (OC) development. DC-STAMP-/- mice have mild osteopetrosis and form mononuclear cells with limited resorption capacity. The identification of an Immunoreceptor Tyrosine-based Inhibitory Motif (ITIM) on the cytoplasmic tail of DC-STAMP suggested a potential signaling function. The absence of a known DC-STAMP ligand, however, has hindered the elucidation of downstream signaling pathways. To address this problem, we engineered a light-activatable DC-STAMP chimeric molecule in which light exposure mimics ligand engagement that can be traced by downstream Ca²⁺ signaling. Deletion of the cytoplasmic ITIM resulted in a significant elevation in the amplitude and duration of intracellular Ca²⁺ flux. Decreased NFATc1 expression in DC-STAMP-/- cells was restored by DC-STAMP over-expression. Multiple biological phenotypes including cell-cell fusion, bone erosion, cell mobility, DC-STAMP cell surface distribution, and NFATc1 nuclear translocation were altered by deletion of the ITIM and adjacent amino acids. In contrast, mutations on each of the tyrosine residues surrounding the ITIM showed no effect on DC-STAMP function. Collectively, our results suggest that the ITIM on DC-STAMP is a functional motif that regulates osteoclast differentiation through the NFATc1/Ca²⁺ axis. J. Cell. Physiol. 232: 2538-2549, 2017. © 2016 Wiley Periodicals, Inc.

Figure 1. Comparison of human and murine sequences of DC-STAMP. Susceptible (L⁴⁰⁰ to F⁴⁰⁰) point mutations in Paget's disease and ITIM are present in the DC-STAMP cytoplasmic tail

(A) The transmembrane (T) regions of DC-STAMP are labeled with T1-T7. The point mutation (L to F) identified in Paget’s disease is labeled with a red arrow; ITIM labeled with a red bracket, 3 tyrosine residues close to ITIM are labeled in red, and deleted protein sequences (TD) in GFP and PA chimeric constructs are underlined. (B) Three tyrosine residues (in red) are close to ITIM. . Red rectangle: ITIM; Blue: 6 amino-acid deletions in tail-deleted mutant (TD). (C) Two point mutations (mut#1: Y⁴⁰⁹F⁴⁰⁹; mut#2: Y⁴⁰⁹A⁴⁰⁹) were introduced in the ITIM.

Figure 1. Comparison of human and murine sequences of DC-STAMP. Susceptible (L⁴⁰⁰ to F⁴⁰⁰) point mutations in Paget's disease and ITIM are present in the DC-STAMP cytoplasmic tail

(A) The transmembrane (T) regions of DC-STAMP are labeled with T1-T7. The point mutation (L to F) identified in Paget’s disease is labeled with a red arrow; ITIM labeled with a red bracket, 3 tyrosine residues close to ITIM are labeled in red, and deleted protein sequences (TD) in GFP and PA chimeric constructs are underlined. (B) Three tyrosine residues (in red) are close to ITIM. . Red rectangle: ITIM; Blue: 6 amino-acid deletions in tail-deleted mutant (TD). (C) Two point mutations (mut#1: Y⁴⁰⁹F⁴⁰⁹; mut#2: Y⁴⁰⁹A⁴⁰⁹) were introduced in the ITIM.

Figure 2. Deletion of ITIM on DC-STAMP alters cell-cell fusion, cell volume and bone resorption

(A) DC-STAMP−/− cells were infected with GFP-tagged DC-STAMP constructs: wild-type (WT, d-f), ITIM-deleted (TD, g-i), tyrosine mutants (Y⁴⁰⁹F⁴⁰⁹ or A⁴⁰⁹, j-l), or vector (a-c). After viral infection, cells were cultured in OC-promoting media (RANKL and M-CSF). 1^st column: TRAP staining (a,d,g,j); 2^nd column: GFP fluorescent images (b,e,h,k); 3^rd column: bone wafer assay (c,f,i,l). Red asterisks identify selected erosion pits. (B) Summary of results shown in (A). (a) Number of nuclei per OC, (b) surface area per OC, (c) surface area per erosion pit. (***P=0.001; **P=0.01).

Figure 2. Deletion of ITIM on DC-STAMP alters cell-cell fusion, cell volume and bone resorption

(A) DC-STAMP−/− cells were infected with GFP-tagged DC-STAMP constructs: wild-type (WT, d-f), ITIM-deleted (TD, g-i), tyrosine mutants (Y⁴⁰⁹F⁴⁰⁹ or A⁴⁰⁹, j-l), or vector (a-c). After viral infection, cells were cultured in OC-promoting media (RANKL and M-CSF). 1^st column: TRAP staining (a,d,g,j); 2^nd column: GFP fluorescent images (b,e,h,k); 3^rd column: bone wafer assay (c,f,i,l). Red asterisks identify selected erosion pits. (B) Summary of results shown in (A). (a) Number of nuclei per OC, (b) surface area per OC, (c) surface area per erosion pit. (***P=0.001; **P=0.01).

Figure 3. DC-STAMP cell surface distribution and OC migration altered by ITIM deletion

(A) Murine DC-STAMP−/− BMM were infected with GFP-tagged WT- (top row, a & c) or TD-(bottom row, b & d) DC-STAMP. Cellular localization of DC-STAMP was traced by GFP fluorescence. a & b , c & d were photographed under different magnifications as shown by the scale bar; c & d were photographed under the same magnification, TD-DC-STAMP expressing cells demonstrated a more uneven distribution pattern (red asterisks) of DC-STAMP compared to WT-DC-STAMP (d vs. c). (B) DC-STAMP WT proteins were overexpressed in DC-STAMP−/− cells by retroviral infection, cultured in OC-promoting media (RANKL+M-CSF) on bone wafers, and visualized by scanning electron microscope (SEM). (C) DC-STAMP TD proteins were overexpressed in DC-STAMP−/− cells following the same experimental conditions as (B). WT DC-STAMP: a, b, d, e, g, h, j; TD DC-STAMP: c, f, i, k. (a)-(f): BMM-derived OC on the surface of bone wafers; (g)-(i): pits eroded by BMM-derived OC; (j)-(k): villous extensions on the cell surface of osteoclasts. Three different scale bars: 1, 2, & 10 µm are shown. The infection efficiency (>70%) by retrovirus on the bone wafer was assessed by GFP expression in a parallel experiment.

Figure 3. DC-STAMP cell surface distribution and OC migration altered by ITIM deletion

(A) Murine DC-STAMP−/− BMM were infected with GFP-tagged WT- (top row, a & c) or TD-(bottom row, b & d) DC-STAMP. Cellular localization of DC-STAMP was traced by GFP fluorescence. a & b , c & d were photographed under different magnifications as shown by the scale bar; c & d were photographed under the same magnification, TD-DC-STAMP expressing cells demonstrated a more uneven distribution pattern (red asterisks) of DC-STAMP compared to WT-DC-STAMP (d vs. c). (B) DC-STAMP WT proteins were overexpressed in DC-STAMP−/− cells by retroviral infection, cultured in OC-promoting media (RANKL+M-CSF) on bone wafers, and visualized by scanning electron microscope (SEM). (C) DC-STAMP TD proteins were overexpressed in DC-STAMP−/− cells following the same experimental conditions as (B). WT DC-STAMP: a, b, d, e, g, h, j; TD DC-STAMP: c, f, i, k. (a)-(f): BMM-derived OC on the surface of bone wafers; (g)-(i): pits eroded by BMM-derived OC; (j)-(k): villous extensions on the cell surface of osteoclasts. Three different scale bars: 1, 2, & 10 µm are shown. The infection efficiency (>70%) by retrovirus on the bone wafer was assessed by GFP expression in a parallel experiment.

Figure 4. Nuclear translocation of NFATc1 is decreased in cells infected with DC-STAMP-TD constructs

(A) DC-STAMP−/− BMM were transiently infected with GFP-tagged WT (top row) or TD (bottom row) DC-STAMP. a & d: NFATc1/nucleus overlay; b & e: DC-STAMP/nucleus overlay; c & f: overlay of NFATc1/DC-STAMP/nucleus. (B) Percentage of cells with NFATc1 nuclear localization. The percentage of cells with NFATc1 nuclear translocation was calculated by examining 100 GFP+ cells at two time points post-infection (48 & 120 hours). Figure S4 depicts WT- and TD-DC-STAMP-expressing cells in a lower magnification.

Figure 5. Relationship between DC-STAMP and NFATc1 protein expression

(A) Ex vivo expression of DC-STAMP. DC-STAMP−/− cells demonstrate reduced expression of NFATc1 protein in bone marrow, thymus, and spleen. Western blots depict proteins isolated from DC-STAMP +/+ (WT) & DC-STAMP −/− (KO) littermates. (B) Expression of DC-STAMP in WT and DC-STAMP−/− cells cultured with RANKL and M-CSF. NFATc1 protein expression was induced when DC-STAMP+/+ and DC-STAMP−/− BMMs were cultured in the presence of RANKL+M-CSF. The maximal NFATc1 expression in +/+ and −/− cells was noted at 96 and 48 hours post-infection, respectively. The expression of NFATc1 on DC-STAMP+/+ (left) and DC-STAMP−/− (right) cells after 8, 48, 96 and 120- hours in the presence of RANKL+M-CSF is shown. (C) Reduced NFATc1 protein level in DC-STAMP−/− cells is restored by DC-STAMP overexpression. Reduced NFATc1 protein expression in DC-STAMP−/− cells (Fig 2B) was restored following retroviral infection with WT DC-STAMP construct. Total proteins were isolated at 4 time points post-infection (48, 72, 96, 120 hours) and subjected to western blot analysis.

Figure 6. Intracellular Ca²⁺ flux was altered by ITIM deletion on the DC-STAMP cytoplasmic tail

(A) The design of the photoactivatable DC-STAMP (PA-DC-STAMP) chimeric construct. Rhod, rhodopsin. (B) Ca²⁺ flux assays on WT- or TD-DC-STAMP-overexpressing cells. 293T cells were transiently transfected with WT- or TD-DC-STAMP constructs, labeled with the Fluo-4 Ca²⁺ dye, and activated with 488-nm light. Cells expressing the PA constructs were selected and identified by the expression of red m-Cherry proteins. Top row: Three videos were recorded for 6 minutes after light activation. Middle row: 120 m-cherry+ cells in each video were selected and the Ca²⁺ signals, as represented by the green fluorescence, were depicted and converted into curves in a 0.01 second interval using the Amira software. Bottom row: Curves derived from cumulative Ca²⁺ signals from 120 m-cherry+ cells in each video. The starting and end points of experiments are defined by the light activation (red arrows) and ionophore addition (blue arrows). Click the button to start each video. Ca²⁺ pulses are labeled with black asterisks, and the arrows close to the spikes indicate the trend of Ca²⁺ signal strength. (C) Fold change of intracellular Ca²⁺ flux before (a) and after (b) light activation on 293T cells transfected with vector, WT-, or –TD-PA-DCSTAMP. A significant difference in Ca²⁺ flux was detected between the ITIM-deleted PA mutant and vector control (mean +/− SEM; * P ≤ 0.05). (D) Percentage of WT- and TD- DCSTAMP-expressing 293T cells that showed Ca²⁺ flux after light activation. * P ≤ 0.05. (E) Duration time of intracellular Ca²⁺ signals after light activation in WT-, TD-DCSTAMP, & vector- transfected 293T cells. *P ≤ 0.05.

Figure 6. Intracellular Ca²⁺ flux was altered by ITIM deletion on the DC-STAMP cytoplasmic tail

(A) The design of the photoactivatable DC-STAMP (PA-DC-STAMP) chimeric construct. Rhod, rhodopsin. (B) Ca²⁺ flux assays on WT- or TD-DC-STAMP-overexpressing cells. 293T cells were transiently transfected with WT- or TD-DC-STAMP constructs, labeled with the Fluo-4 Ca²⁺ dye, and activated with 488-nm light. Cells expressing the PA constructs were selected and identified by the expression of red m-Cherry proteins. Top row: Three videos were recorded for 6 minutes after light activation. Middle row: 120 m-cherry+ cells in each video were selected and the Ca²⁺ signals, as represented by the green fluorescence, were depicted and converted into curves in a 0.01 second interval using the Amira software. Bottom row: Curves derived from cumulative Ca²⁺ signals from 120 m-cherry+ cells in each video. The starting and end points of experiments are defined by the light activation (red arrows) and ionophore addition (blue arrows). Click the button to start each video. Ca²⁺ pulses are labeled with black asterisks, and the arrows close to the spikes indicate the trend of Ca²⁺ signal strength. (C) Fold change of intracellular Ca²⁺ flux before (a) and after (b) light activation on 293T cells transfected with vector, WT-, or –TD-PA-DCSTAMP. A significant difference in Ca²⁺ flux was detected between the ITIM-deleted PA mutant and vector control (mean +/− SEM; * P ≤ 0.05). (D) Percentage of WT- and TD- DCSTAMP-expressing 293T cells that showed Ca²⁺ flux after light activation. * P ≤ 0.05. (E) Duration time of intracellular Ca²⁺ signals after light activation in WT-, TD-DCSTAMP, & vector- transfected 293T cells. *P ≤ 0.05.