healthmedicinet

Skeletal muscle cells express IL-15/IL-15R? complex protein in response to TNF-? and
IFN-? stimulation

Previous studies found up-regulation of Il15 RNA during myoblast differentiation 11] and pro-inflammatory cytokine stimulation 27]. Given the presence of post-transcriptional regulation of IL-15 expression and that
all circulating IL-15 are in complex with IL-15R? 9], we examined the expression of IL-15/IL-15R? complex protein by skeletal muscle cell
under the two conditions mentioned above. During a 6-day C2C12 myoblast-to-myotube
differentiation, the cells showed a greater than 10-fold and a 3-fold increase of
Il15 and Il15ra mRNA, respectively (Fig. 1a), whereas the cell lysate and culture medium contained no detectable IL-15/IL-15R?
protein by ELISA with a sensitivity of 3.9 pg/ml. This result indicates that although
the level of Il15 and Il15ra mRNA increased along myoblast-to-myotube differentiation, there was little production
of IL-15/IL-15R? protein. We next examined whether pro-inflammatory cytokines induce
IL-15 and IL-15R? expression by skeletal muscle cells. We found that TNF-? and IFN-?
each up-regulated Il15 and Il15ra mRNA in C2C12 myotubes with distinct kinetics and together exerted a synergistic
effect (Fig. 1b). A similar synergistic effect occurred in primary myotubes (Fig. 1c). TNF-? and IFN-? treatment also up-regulated the expression of the IL-15/IL-15R?
complex protein in C2C12 myoblasts and myotubes (Fig. 1d) and in primary myotubes (Fig. 1e) in vitro. Most of the IL-15/IL-15R? proteins were present in the C2C12 cell lysates
(Fig. 1d), while nearly all IL-15/IL-15R? proteins were present in the primary myotube lysate
(Fig. 1e). Injection of TNF-? and IFN-? into the quadriceps muscle in mice also greatly induced
the expression of IL-15/IL-15R? protein in the injected muscle (Fig. 1f).

IFN-?, another cytokine up-regulated in myositis muscle 26], induced Il15 and Il15ra mRNA and IL-15/IL-15R? complex protein to the level similar to those induced by IFN-?
or TNF-? (Additional file 2: Figure S1). Other pro-inflammatory factors, such as IL-1?, IL-1?, and LPS, only
transiently up-regulated Il15ra or Il15 mRNA but showed small effect on the induction of the IL-15/IL-15R? protein (Additional
file 2: Figure S1). Together, these results indicate that the expression of the IL-15/IL-15R?
complex protein by skeletal muscle cells was undetectable under steady state conditions,
while induced by TNF-?, IFN-?, or IFN-? that associates with Th1 response in myositis.

IL-15 is present on the surface of skeletal muscle cells

As the majority of the cytokine-induced IL-15/IL-15R? complex was in the lysates of
the myoblast and myotube (Fig. 1d, e), we next examined whether the complex is present on the cell surface. We found that
TNF-? and IFN-? induced expression of IL-15 and IL-15R? on the surface of C2C12 myoblasts
as detected by flow cytometry (Fig. 2a). For the cytokine-treated C2C12 myotubes, dissociation of IL-15 from IL-15R? on
the cell surface by washing cells with acidic glycine buffer resulted in an 80 % reduction
of IL-15/IL-15R? in the cell lysate (Fig. 2b), indicating that 80 % of IL-15 was presented by IL-15R? on the cell surface. This
reduction was not due to alteration of IL-15R? by the acid treatment, as the level
of IL-15/IL-15R? resumed following the addition of exogenous IL-15 (Fig. 2b). These results indicated that the majority of the cytokine-induced IL-15 was present
on the surface of skeletal muscle cells as IL-15/IL-15R? complex.

Fig. 2. IL-15 is presented on the surface of skeletal muscle cells. a Expression of IL-15 and IL-15R? on C2C12 myoblasts. Cells were treated with TNF-?
and IFN-? (TNF?+?IFN) or without cytokine (Control) for 24 h and then stained with anti-IL-15 and IL-15R? antibodies for flow cytometry
analysis. “Unstained” indicates cells without antibody staining. Data are representative of three independent
experiments with similar results. b Quantification of IL-15 bound by cell surface IL-15R? on C2C12 myotubes. C2C12 myotubes
were first treated with TNF-? and IFN-? for 24 h, incubated with acid glycine buffer
to dissociate IL-15 from cell surface IL-15R?, and then re-treated with exogenous
IL-15. Cell lysates from each step were collected and the amount of IL-15/ IL-15R?
complex was measured by ELISA. Data are mean?±?SEM of triplicates and representative
of two independent experiments with similar results. ***p??0.001

Skeletal muscle cells express a scanty amount of IL-15R? and only respond to a high
concentration of IL-15 hyperagonist, but not IL-15

Despite of various reported IL-15 functions in the skeletal muscle 12]–16], the IL-15-induced signals in skeletal muscle cells remains unexplored. We first
examined the expression of IL-15R? and ?
_c
on C2C12 myoblasts by flow cytometry and detected no IL-15R? expression but a low
level of ?
_c
induced by TNF-? and IFN-? treatment (Fig. 3a). The more sensitive qPCR also detected induction of ? _c
mRNA, but not Il15rb mRNA, in C2C12 and primary myotubes by the cytokines (Fig. 3b). Moreover, the level of Il15rb mRNA in the primary myotube was 16 times lower than that in the C2C12 myotube based
on the Ct value of qPCR.

Fig. 3. High concentration of IL-15 hyperagonist, but not IL-15, induces STAT5 signaling and
atrophy in skeletal muscle cells. a Analysis of IL-15R? and ?
_c
expression on C2C12 myoblasts under the same condition of Fig. 2a. Data are representative of three independent experiments with similar results. b Expression of Il15rb and ? _c
mRNA in C2C12 and primary myotube. Samples were collected 24 h after cytokine treatment
and analyzed by qPCR. Data are triplicates and representative of two independent experiments
with similar results. c Immunoblotting of STAT5 phosphorylation in C2C12 myotubes after treating with IL-15
or IL-15 hyperagonist for 30 min. Quantification data of four independent experiments
are shown below. d Anti-IL-15R? antibody diminished IL-15 hyperagonist-induced STAT5 phosphorylation
in C2C12 myotubes. Myotubes were pretreated with anti-IL-15R? or isotype control antibody
for 1 h and then treated with IL-15 hyperagonist for 30 min. Data are representative
of two independent experiments with similar results. e Immunoblotting of STAT5 phosphorylation in C2C12 myotubes that were first stimulated
with TNF-? and IFN-? for 24 h and then treated with exogenous IL-15 or IL-15 hyperagonist
for 30 min. Data are representative of two independent experiments with similar results.
f Accumulation of myosin heavy chain (MyHC) in C2C12 myotubes treated with IL-15 (50
and 100 ng/ml), IL-15 hyperagonist (50 and 100 ng/ml), or vehicle (0 ng/ml). After
48 h, cells were immunostained with FITC-labeled anti-myosin antibody to evaluate
the accumulation of MyHC. MyHC-positive area per microscopic field are shown in the
right panel and represented as the percentages of con. Quantification data were pooled from three
independent experiments. Each symbol represents the quantification data from one microscopic field. Scale bar?=?500 ?m.
Data are mean?±?SEM. *p??0.05, **p??0.01, ***p??0.001

Due to the possibility of a very low level of IL-15R? expression, we examined whether
IL-15 or an IL-15 hyperagonist induces signal transduction in C2C12 myotubes. The
latter is a fusion protein of IL-15 and IL-15R?-sushi domain, which possesses higher
binding affinity for IL-15R?/?
_c
(K _d
?=?780 pM) and promotes a stronger proliferation of IL-15R?/?
_c
-bearing cells (EC
₅₀
?=?25 pM) than IL-15 (K _d
?=?13.5 nM, EC
₅₀
?=?3 nM) 40]. We found that IL-15 did not induce STAT5 phosphorylation at a concentration up to
400 ng/ml (26.7 nM) (Fig. 3c). Whereas the IL-15 hyperagonist induced moderate but significant STAT5 phosphorylation
at concentrations of 100 ng/ml (3.4 nM) and higher (Fig. 3c), which are much higher than that required for the proliferation of IL-15R?/?
_c
-bearing cells (EC
₅₀
?=?25 pM) and for STAT5 phosphorylation in pre-activated murine CD8
⁺
cells (EC
₅₀
?=?10 pM) 40], 41]. In addition to STAT5 phosphorylation, IL-15 and IL-15 hyperagonist did not induce
phosphorylation of STAT3, AKT, and ERK in skeletal muscle cells (data not shown).
The IL-15 hyperagonist-induced STAT5 phosphorylation was completely blocked by the
IL-15R?-blocking antibody TM-b1 (Fig. 3d). However, neither IL-15 nor IL-15 hyperagonist induced STAT5 phosphorylation in
C2C12 myotubes pre-treated with TNF-? and IFN-? (Fig. 3e), which is in line with the decrease of Il15rb mRNA after cytokine treatment (Fig. 3b). Consistent with the STAT5 phosphorylation results (Fig. 3c), 100 ng/ml of IL-15 hyperagonist, but not IL-15, induced atrophy of C2C12 myotubes
(Fig. 3f). In summary, C2C12 and primary skeletal muscle cells expressed a scanty IL-15R?
under steady state condition and TNF-?/IFN-? stimulation and only responded to a high
concentration of IL-15 hyperagonist, but not IL-15. Moreover, we found that Il15 ^?/?
mice showed normal skeletal muscle mass (Additional file 3: Figure S2), myoblast differentiation (Additional file 4: Figure S3), cardiotoxin-induced muscle regeneration (Additional file 5: Figure S4), and compensatory hypertrophy of plantaris muscle (Additional file 6: Figure S5). The in vitro and in vivo results collectively suggest that the skeletal
muscle cells do not use IL-15.

IL-15 deficiency does not affect the response of primary myotube to TNF-? and IFN-?
stimulation

As TNF-? and IFN-? induced abundant expression of IL-15/IL-15R? on skeletal muscle
cells, the local concentration of IL-15/IL-15R? trans-presentation may be high enough
to trigger signaling through the limited number of IL-15R?/?
_c
on adjacent muscle cells. We thus examined whether IL-15 affects the response of the
skeletal muscle to TNF-? and IFN-? by comparing wild type (wt) and Il15 ^?/?
primary myotubes. As TNF-? and IFN-? induce muscle wasting under cancer cachexia and
IL-15 was shown to prevent it 16], 42], 43], we first examined genes that regulate muscle mass. Upon TNF-? and IFN-? stimulation,
wt and Il15 ^?/?
primary myotubes showed similar reduction in the hypertrophy-related genes Igf1 and Myh4 and increase in the atrophy-related gene iNos (Fig. 4a). We next examined the immune regulatory genes affected by IL-15 in immune cells
44]–48]. We found that TNF-? and IFN-? induced the expression of immune regulatory genes
in wt and Il15 ^?/?
primary myotubes to similar extents, including molecules in the MHC class I antigen
presentation pathway and for T-cell co-stimulation and inhibition, chemokines, and
cell adhesion molecules (Fig. 4b, c). These results indicate that skeletal muscle IL-15 does not affect the expression
of protein homeostasis and immune regulation genes by skeletal muscle cells in response
to TNF-? and IFN-?, which is consistent with the idea that the muscle cells do not
use their own IL-15.

Fig. 4. IL-15 deficiency does not affect the responses of primary myotube to TNF-? and IFN-?
stimulation. a–c Expression profiling of genes involved in the regulation of skeletal muscle mass
and immune system in wt and Il15 ^?/?
primary myotubes. Samples were collected 24 h after TNF-? and IFN-? treatment and
analyzed by qPCR. Data are mean?±?SEM of triplicates. Data are representative of two
independent experiments with similar results. *p??0.05, **p??0.01, ***p??0.001

Skeletal muscle cells stimulated with TNF-? and IFN-? present antigen and provide
IL-15 to memory-like CD8
⁺
T cells

IL-15 is a well-known survival and activation factor for memory CD8
⁺
T cells. As inflammatory cytokines induce the expression of IL-15/IL-15R? and antigen
presentation molecules by myoblasts and myotubes 49], we designed a muscle-cell-T-cell co-culture system to assess whether the muscle
cells directly activate CD8
⁺
T cells and the role of IL-15 in this process. We generated a C2C12 myoblast subline
that stably expresses H-2K
^b
-EGFP (C2C12-K
^b
). As overexpression of H-2K
^b
impairs myoblast differentiation in this study and 50], this co-culture system is for myoblast and T cells. We then transduced full-length
OVA gene into C2C12-K
^b
myoblast as an endogenous antigen and generated C2C12-K
^b
/OVA myoblast. TNF-? and IFN-? greatly induced the expression of IL-15 and H-2K
^b
on C2C12-K
^b
/OVA cells (Fig. 5a). This high induction of H-2K
^b
might partly result from the up-regulation of ?2-microglobulin (B2m) by the cytokines
(Fig. 4b), because B2m is essential for the stabilization of MHC class I molecule in correct
conformation to receive the peptide in the ER and to move from the ER to the cell
surface 51]. The cytokine-stimulated C2C12-K
^b
/OVA, but not C2C12-K
^b
, cells induced production of granzyme B (grB) and IFN-? by memory-like OT-1 cells
(Fig. 5b). We then found that an IL-15R?-blocking antibody, but not IL-2-neutralizing antibody,
suppressed grB and IFN-? production by the memory-like OT-1 cells (Fig. 5c). As the cytokine-stimulated C2C12-K
^b
/OVA cells were washed before co-culturing with OT-1 cells in the presence of the
exocytosis inhibitor brefeldin A, IL-15 was presumably only present on the muscle
cell surface. These results indicate that myoblasts stimulated with TNF-? and IFN-?
present antigen and IL-15 to memory-like CD8
⁺
T cells to promote their effector function. Given that both myoblasts and myotubes
function as antigen-presenting cells under inflammation, what we observed in the myoblast-CD8
⁺
T-cell co-culture is likely applicable to myotubes.

Fig. 5. Skeletal muscle cells presented antigen and provided IL-15 to promote the effector
function of CD8
⁺
T cells under TNF-? and IFN-? treatment. a Expression of IL-15 and H-2K
^b
on parental C2C12 and C2C12-K
^b
/OVA cells. Cells were treated with TNF-? and IFN-? (TNF?+?IFN) or without (Control) for 24 h, stained with anti-IL-15 or anti-H-2Kb antibody, then analyzed by flow
cytometry. “Unstained” indicates C2C12 or C2C12-K
^b
/OVA myoblast without antibody staining. b Expression of grB and IFN-? by CD8
⁺
T cells co-cultured with cytokine-treated C2C12-K
^b
or C2C12-K
^b
/OVA myoblasts for 8 h. After co-culture, CD8
⁺
T cells were stained with anti-grB and IFN-? antibody by intracellular staining then
analyzed by flow cytometry. Data are representative of three independent experiments
with similar results. c Effect of anti-IL-2 (?IL-2, S4B6) or anti-IL-2/15R? (?R?, TM-?1) antibody on grB
and IFN-? production by CD8
⁺
T cells co-cultured with TNF?+?IFN-pretreated C2C12-K
^b
/OVA myoblasts. The percentages of grB
⁺
and IFN-?
⁺
cells in CD8
⁺
T cells are indicated in the dot plots. Data from independent experiments were compiled as % normalized activation using
the percentage of grB
⁺
or IFN-?
⁺
cells in CD8
⁺
T cells of the TNF?+?IFN-pretreated C2C12-K
^b
/OVA cell co-culture group of each experiment as 100 % (bottom panels). Isotype controls, Rat IgG2a (Iso) and Rat IgG2b (Iso). Each symbol is representative of one independent experiment. **p??0.01, ***p??0.001

Skeletal muscle IL-15 promotes the progression of autoimmune myositis

TNF-? and IFN-? are commonly expressed in the skeletal muscle of patients suffering
from inflammatory myopathies, in which CD8
⁺
T cells infiltrate and play a critical role in disease progression 23], 25], 52]. The enhancement of memory-like CD8
⁺
T-cell effector function by myoblast IL-15 in vitro prompted us to examine the role
of skeletal muscle IL-15 in autoimmune myositis in vivo. We first generated skeletal-muscle-specific
Il15 ^?/?
mice by crossing Il15 ^f/f
mice with ACTA–cre mice. The ACTA-Il15 ^?/?
mice showed an 80 % reduction of Il15 mRNA specifically in the skeletal muscle (Fig. 6a) with normal levels of IL-15/IL-15R? complex and NK and memory CD8
⁺
T cells in the peripheral blood (Fig. 6b, c).

Fig. 6. Skeletal muscle IL-15 contributes to the progression of autoimmune myositis. a The expression of Il15 mRNA in various tissues of wt (?=?3), Il15 ^f/f
(?=?5), and ACTA-Il15 ^?/?
(?=?5) mice was detected by qPCR. WAT white adipose tissue. Data are mean?±?SEM. ***p??0.001. b The amount of IL-15/IL-15R? complex protein in the serum of Il15 ^f/f
and ACTA-Il15 ^?/?
mice (?=?9–10) was measured by ELISA. Data are mean?±?SEM. c Comparison of memory CD8
⁺
T cell (mCD8) and NK cell (NK) level in the peripheral blood among wt, Il15 ^?/?
, Il15 ^f/f
, and ACTA-Il15 ^?/?
mice (?=?5 in each group). Fold change was calculated by normalizing the percentage of indicated
cell type in mutant mice to that in wt mice. mCD8 were H57
⁺
CD19
^?
CD8
⁺
CD44
^hi
CD122
^hi
, and NK were H57
^?
CD19
^?
NK1.1
⁺
. Data are mean?±?SEM. ***p??0.001. d Mononuclear cell infiltration in the quadriceps muscles of Syt7 ^?/? Il15 ^f/f
and Syt7 ^?/? ACTA-Il15 ^?/?
mice 14 days after C protein immunization (Left). Mononuclear cell infiltration was found in the endomysium (black square), perimysium (black arrowhead), and perivascular region (white arrowhead). Focal lymphatic invasion of myofibers were observed in C-protein-immunized Syt7 ^?/? ACTA-Il15 ^?/?
mice as indicated by the white arrowhead (Middle). The histopathology scores were compiled in the right panel with each symbol representing one mouse (Right). Scale bar in left?=?100 ?m; middle?=?25 ?m. **p??0.01. e Expression of mononuclear cell markers, Cd4, Cd8, and F4/80, mRNA in the quadriceps muscles of C-protein-immunized mice measured by qPCR. Each
symbol representing one mouse. *p??0.05, **p??0.01. f Expression of immune-relevant genes mRNA in the quadriceps muscles of C-protein-immunized
mice measured by qPCR. Each symbol representing one mouse. *p??0.05, **p??0.01

As two previously reported autoimmune myositis models, wt mice immunized with C protein
38], 53] and Syt7 ^?/?
mice 54], 55], did not develop myositis in our hand, we immunized mice of Syt7 ^?/?
background with C protein because impaired muscle membrane sealing due to Syt7 deficiency
facilitates myositis induction 54], 56]. Similar to previously reported pathology in autoimmune myositis, we found that C-protein-immunized
Syt7^?/?Il15 ^f/f
mice developed mononuclear cell infiltration (Fig. 6d). The mononuclear cells predominantly infiltrated into the endomysium as well as
the perimysium and perivascular region (Fig. 6d (left)). The immunized Syt7^?/?Il15 ^f/f
mice also developed focal lymphatic invasion of muscle fibers that features CD8
⁺
T-cell-mediated myositis (Fig. 6d (middle)). Whereas C-protein-immunized Syt7^?/?ACTA–Il15 ^?/?
mice showed a significantly reduced mononuclear cell infiltration and histopathology
score (Fig. 6d). We also found elevation of Cd4, Cd8?, and F4/80 mRNA in the skeletal muscle of the immunized Syt7^?/?Il15 ^f/f
mice (Fig. 6e), suggesting infiltration of CD4
⁺
and CD8
⁺
T cells and macrophages. The expression of MHC class I subunit B2m; pro-inflammatory cytokines Tnfa, Ifng, and Il1?: and effector molecule Prf1 were also induced in the skeletal muscle of the immunized Syt7^?/?Il15 ^f/f
mice (Fig. 6f). Whereas all these molecules examined were not induced in the C-protein-immunized
Syt7^?/?ACTA–Il15 ^?/?
mice (Fig. 6e, f). These results together demonstrate that genetic ablation of skeletal muscle IL-15
greatly reduced the pathogenesis of autoimmune myositis.