Anti-SIRPα antibodies as a potential new tool for cancer immunotherapy

Abstract

Tumor cells are thought to evade immune surveillance through interaction with immune cells. Much recent attention has focused on the modification of immune responses as a basis for new cancer treatments. SIRPα is an Ig superfamily protein that inhibits phagocytosis in macrophages upon interaction with its ligand CD47 expressed on the surface of target cells. Here, we show that SIRPα is highly expressed in human renal cell carcinoma and melanoma. Furthermore, an anti-SIRPα Ab that blocks the interaction with CD47 markedly suppressed tumor formation by renal cell carcinoma or melanoma cells in immunocompetent syngeneic mice. This inhibitory effect of the Ab appeared to be mediated by dual mechanisms: direct induction of Ab-dependent cellular phagocytosis of tumor cells by macrophages and blockade of CD47-SIRPα signaling that negatively regulates such phagocytosis. The antitumor effect of the Ab was greatly attenuated by selective depletion not only of macrophages but also of NK cells or CD8+ T cells. In addition, the anti-SIRPα Ab also enhances the inhibitory effects of Abs against CD20 and programmed cell death 1 (PD-1) on tumor formation in mice injected with SIRPα-nonexpressing tumor cells. Anti-SIRPα Abs thus warrant further study as a potential new therapy for a broad range of cancers.

Full text

1insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
Conflict of interest: The authors have
declared that no conflict of interest
exists.
Submitted: June 16, 2016
Accepted: November 17, 2016
Published: January 12, 2017
Reference information:
JCI Insight. 2017;2(1):e89140.
doi:10.1172/jci.insight.89140.
Anti-SIRPα antibodies as a potential new
tool for cancer immunotherapy
Tadahiko Yanagita,1,2 Yoji Murata,1 Daisuke Tanaka,1 Sei-ichiro Motegi,3 Eri Arai,4,5
Edwin Widyanto Daniwijaya,1 Daisuke Hazama,1 Ken Washio,1 Yasuyuki Saito,1 Takenori Kotani,1
Hiroshi Ohnishi,6 Per-Arne Oldenborg,7 Noel Verjan Garcia,8 Masayuki Miyasaka,8,9
Osamu Ishikawa,3 Yae Kanai,4,5 Takahide Komori,2 and Takashi Matozaki1
1Division of Molecular and Cellular Signaling, Department of Biochemistry and Molecular Biology, 2Department of Oral and
Maxillofacial Surgery, Kobe University Graduate School of Medicine, Kobe, Japan. 3Department of Dermatology, Gunma
University Graduate School of Medicine, Gunma, Japan. 4Division of Molecular Pathology, National Cancer Center Research
Institute, Tokyo, Japan. 5Department of Pathology, Keio University School of Medicine, Tokyo, Japan. 6Department of
Laboratory Sciences, Gunma University Graduate School of Health Sciences, Gunma, Japan. 7Department of Integrative
Medical Biology, Section for Histology and Cell Biology, Umeå University, Umeå, Sweden. 8Laboratory of Immunodynamics,
Department of Microbiology and Immunology, Osaka University Graduate School of Medicine, Osaka, Japan.
9MediCity Research Laboratory, University of Turku, Turku, Finland.
Introduction
The tumor microenvironment consists of a variety of stromal cell types including fibroblasts, immune cells,
and endothelial cells, as well as soluble and insoluble factors such as cytokines, chemokines, and extracel-
lular matrix (1, 2). This microenvironment plays an important role in the regulation of tumor progression
by promoting tumor cell survival, invasion, and metastasis as well as angiogenesis (1–3). Crosstalk between
tumor and immune cells in the tumor microenvironment is also thought to contribute to the evasion of
tumor cells from immune surveillance. For instance, binding of programmed cell death 1 (PD-1) on cyto-
toxic T lymphocytes to its ligand PD-L1 on tumor cells prevents killing of the latter cells by the former (4).
Indeed, Abs against PD-1 are now in clinical use for the treatment of cancers including advanced mela-
noma, renal cell carcinoma, and non–small-cell lung cancer (5). Moreover, the binding of tumor-derived
soluble MHC class I–related chain A (MICA) to its receptor NKG2D on NK cells and T cells results in
the downregulation of NKG2D and impairs the responsiveness of such cells specific for tumor antigens (6,
7). Molecules that participate in negative regulation of the antitumor response of immune cells are thus
promising targets for cancer therapy.
Signal regulatory protein α (SIRPα) is a transmembrane protein with an extracellular region compris-
ing three Ig-like domains and a cytoplasmic region containing immunoreceptor tyrosine–based inhibition
motifs that mediate binding of the protein tyrosine phosphatases SHP1 and SHP2 (8, 9). Tyrosine phos-
phorylation of SIRPα is regulated by various growth factors and cytokines as well as by integrin-mediated
Tumor cells are thought to evade immune surveillance through interaction with immune cells. Much
recent attention has focused on the modification of immune responses as a basis for new cancer
treatments. SIRPα is an Ig superfamily protein that inhibits phagocytosis in macrophages upon
interaction with its ligand CD47 expressed on the surface of target cells. Here, we show that SIRPα
is highly expressed in human renal cell carcinoma and melanoma. Furthermore, an anti-SIRPα Ab
that blocks the interaction with CD47 markedly suppressed tumor formation by renal cell carcinoma
or melanoma cells in immunocompetent syngeneic mice. This inhibitory eect of the Ab appeared
to be mediated by dual mechanisms: direct induction of Ab-dependent cellular phagocytosis of
tumor cells by macrophages and blockade of CD47-SIRPα signaling that negatively regulates such
phagocytosis. The antitumor eect of the Ab was greatly attenuated by selective depletion not only
of macrophages but also of NK cells or CD8+ T cells. In addition, the anti-SIRPα Ab also enhances
the inhibitory eects of Abs against CD20 and programmed cell death 1 (PD-1) on tumor formation
in mice injected with SIRPα-nonexpressing tumor cells. Anti-SIRPα Abs thus warrant further study
as a potential new therapy for a broad range of cancers.
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

2
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
cell adhesion to extracellular matrix proteins. SIRPα is especially abundant in myeloid cells such as mac-
rophages and DCs, whereas it is expressed at only low levels in T, B, NK, and NKT cells (10–13). The
extracellular region of SIRPα interacts with its ligand CD47, which is expressed in most cell types (14) and
is also a member of the Ig superfamily (8, 9, 14).
The interaction of SIRPα on macrophages with CD47 on rbc prevents phagocytosis of Ig-opsonized
rbc by macrophages in vitro (15) and in vivo (16). Such negative regulation of macrophages is thought
to be mediated by the binding of SHP1 to the cytoplasmic region of SIRPα (15). We previously showed
that prevention of the CD47-SIRPα interaction with an Ab against SIRPα in vitro enhanced the killing
by phagocytes of human epidermal growth factor receptor 2–positive (HER2-positive) breast cancer cells
opsonized with the HER2-specific mAb trastuzumab (17), suggesting that such blockade of the CD47-
SIRPα interaction is a promising new approach to cancer treatment. An Ab against CD47 that blocks
the binding of CD47 to SIRPα was shown to promote both Ab-dependent cellular phagocytosis (ADCP)
of human non-Hodgkin lymphoma cells by macrophages in vitro and eradication of xenografts of these
cancer cells induced by the CD20-specific mAb rituximab in immunodeficient mice (18). Moreover, the
same Ab against CD47 was found to inhibit the growth of various human tumor xenografts including
solid tumors (19). However, given that CD47 is ubiquitously expressed at a high level in normal tis-
sues, efficient targeting of CD47 specifically on cancer cells is problematic. Moreover, Abs against CD47
might trigger Ab-dependent cellular cytotoxicity (ADCC) in healthy cells, such as rbc, which is not a
desirable response (20).
To further explore the potential of cancer therapy based on Abs against SIRPα, we first examined
which types of human cancers express this protein at a high level. We then tested the effect of such Abs on
the growth of renal cell carcinoma and melanoma, both of which were found to express SIRPα at a high
level. Finally, we investigated whether the combination of an Ab against SIRPα and other anticancer Abs,
such as those specific for CD20 or PD-1, might suppress tumor growth in vivo in a synergistic manner.
Results
Human renal cell carcinoma and melanoma express SIRPα at a high level. To investigate the potential antitumor
effect of Abs against SIRPα, we first examined which types of human cancer or cancer cell lines express
this protein at a high level by consulting a human protein atlas (21) and The Cancer Cell Line Encyclope-
dia (22). Database searches indicated that SIRPα mRNA or protein might be moderately or highly abun-
dant in human renal cell carcinoma and melanoma. Microarray analysis preformed previously (23) indeed
revealed that the levels of SIRPA mRNA in clear cell renal cell carcinoma (n = 95 patients) was markedly
higher than those in matched normal kidney tissue (Figure 1A). Immunohistochemical staining with poly-
clonal Abs (pAbs) against human SIRPα — the specificity of which was confirmed by immunofluorescence
and immunoblot analyses (Supplemental Figure 1, A–C; supplemental material available online with this
article; doi:10.1172/jci.insight.89140DS1) — also showed that SIRPα protein was expressed at a high level
in tumor sections from 4 patients randomly selected from the 95 patients with clear cell renal cell carci-
noma (Figure 1B, Supplemental Figure 1D, and Table 1). In addition, SIRPα immunoreactivity, which was
determined by a mAb against human SIRPα (040 mAb), was detected at a high level in sections of mela-
noma from 4 of 8 patients and at a moderate or low level in the remaining 4 patients, with such staining
corresponding well to that of the melanoma marker melanoma antigen recognized by T cells 1 (MART-1)
(24) (Figure 1C, Supplemental Figure 2A, and Table 2). Although the 040 mAb could react with SIRPα as
well as the SIRP family members SIRPβ1 or SIRPγ (25) (Supplemental Figure 3), the immunoreactivity of
SIRPβ1 or of SIRPγ in human melanoma was undetectable (Supplemental Figure 2B). It is thus likely that
the expression of SIRPα is specifically increased in these melanoma samples. Consistent with these results,
SIRPα was found to be abundant in human renal cell carcinoma (ACHN, 786-O, A498, Caki-1) and mela-
noma (WM239a, SK-MEL-28, SK-MEL-5) cell lines, although it was present at only a low level in A375
melanoma cells (Supplemental Figure 4). Taken together, these results thus suggested that SIRPα is highly
expressed in human renal cell carcinoma and melanoma.
Abs against SIRPα attenuate the growth of tumors formed by SIRPα-expressing renal cell carcinoma or melanoma
cells in syngeneic mice. We next investigated the effects of 2 different mAbs against mouse SIRPα — MY-1 (rat
IgG2a) (26) and P84 (rat IgG1) (27, 28) — on the growth of tumors formed by mouse renal cell carcinoma or
melanoma cells implanted into syngeneic mice. MY-1 binds to the NH2-terminal Ig-V–like domain of SIRPα
and thereby blocks the interaction with CD47, whereas P84 has little effect on the CD47-SIRPα interaction
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

3
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
(a detailed characterization of these two mAbs is provided in Supplemental Figures 5 and 6). In addition,
MY-1 reacted with SIRPα or the SIRP family member SIRPβ (29) overexpressed in HEK293A cells, whereas
P84 only reacted with SIRPα (Supplemental Figure 5B). Immunoblot and flow cytometric analyses showed
that mouse RENCA renal cell carcinoma and B16BL6 melanoma cells express SIRPα at a high level on the
cell surface (Figure 2A). In contrast, flow cytometric analysis also revealed that expression of SIRPβ was
minimal on the cell surface of both cell lines (Supplemental Figure 7), suggesting that immunoreactivity
for MY-1 or P84 on these tumor cells is attributable to SIRPα expression on their cell surface. Syngeneic
BALB/c mice at 8 weeks of age were injected s.c. with RENCA cells and then i.p. with either normal rat
IgG (control), MY-1, or P84 three times a week (Figure 2B). Treatment with MY-1 resulted in marked
attenuation of tumor formation by RENCA cells compared with that seen in mice treated with control
IgG, whereas P84 manifested a smaller but still significant inhibitory effect on tumor growth (Figure 2B).
Consistent with these findings, mice treated with MY-1 exhibited prolonged survival compared with those
treated with either control IgG or P84 when the treatment was discontinued after 3 weekly cycles (Figure
2B). Moreover, the inhibitory effect of MY-1 on tumor growth as well as its beneficial effect on survival rates
were also apparent when treatment was delayed until the tumor volume had achieved an average size of 100
mm3 (Figure 2C). By contrast, such delayed treatment with P84 had no significant effect on either tumor
growth or survival compared with the effect observed with control IgG (Figure 2C). We then examined
the effects of the Abs against SIRPα on metastatic tumor formation by B16BL6 cells in C57BL/6J mice.
Figure 1. High level of SIRPα expres-
sion in human renal cell carcinoma
and melanoma. (A) Microarray analysis
of signal regulatory protein α (SIRPA)
mRNA abundance in tumor and adjacent
normal tissue from patients with clear
cell renal cell carcinoma. Individual values
(n = 95) are shown. Bars indicate the
median values. ***P < 0.001, by 2-tailed
Student’s t test. (B) Paran-embedded
tumor sections prepared from a patient
with clear cell renal cell carcinoma (case
1) were subjected to H&E staining as well
as immunohistochemical staining (brown)
with polyclonal Abs (pAbs) against human
SIRPα and counterstaining with hema-
toxylin. Boxed regions in the top panels
are shown at higher magnification in
the bottom panels. Scale bars: 500 μm
(top panel) and 100 μm (bottom panel).
(C) Fresh-frozen tumor sections from
a patient with melanoma (case 1) were
subjected to immunofluorescence staining
with mAbs against melanoma antigen
recognized by T cells 1 (MART-1) (magenta)
and human SIRPα (040 mAb) (green).
Nuclei were also stained with DAPI (blue).
Scale bar: 100 μm.
Table 1. Characteristics of the 4 patients with clear cell renal cell carcinoma
Case Age (yr) Sex Source TNM Stage
1 50 Male Kidney T1bN0M0 I
2 56 Female Kidney T3aN0M0 III
3 43 Male Kidney T3aN0M0 III
470 Male Kidney T3bN0M0 III
TNM, tumor, node, metastasis system.
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

4
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
B16BL6 cells were injected i.v. into 8-week-old
mice, which were then treated with either con-
trol IgG, MY-1, or P84 three times a week. MY-1
significantly reduced the number of metastatic
nodules formed in the lungs compared with
either control IgG or P84 (Figure 2D). Togeth-
er, these results thus suggested that Abs against
SIRPα that block the CD47-SIRPα interaction
markedly attenuate tumor formation by SIRPα-
expressing cancer cells in vivo.
We also examined possible adverse effects
of treatment with MY-1 in mice. Hematologic
and blood biochemical analyses showed that
treatment of C57BL/6J mice with MY-1 result-
ed in a small increase in the number of reticulocytes in the blood, but no other adverse effects (including
anemia), compared with treatment with vehicle or control IgG (Supplemental Table 1).
Importance of macrophages for the antitumor effect of MY-1 in vivo. We next investigated whether macro-
phages contribute to the inhibitory effect of MY-1 on tumor formation by SIRPα-expressing cancer cells. We
depleted mice of macrophages by administering clodronate liposomes beginning before injection of cancer
cells. Effective depletion of F4/80+CD11b+ macrophages in the spleen was apparent as early as 3 days after
the first i.v. injection of clodronate liposomes (Figure 3A). Such macrophage depletion was associated with
a marked attenuation of the inhibitory effect of MY-1 on the growth of tumors formed by RENCA cells
(Figure 3B), suggesting that macrophages indeed contribute, at least in part, to the antitumor effect of MY-1.
Given that SIRPα is expressed at a high level in RENCA and B16BL6 cells, it appeared likely that
MY-1 inhibited tumor formation by these cells in vivo at least in part by directly promoting ADCP of tumor
cells by macrophages. Indeed, compared with control IgG, MY-1 markedly promoted the phagocytosis of
CFSE-labeled RENCA cells by BM-derived macrophages (BMDMs) (Figure 3, C and D). In addition, the
Figure 2. Abs against SIRPα attenuate the growth
of tumors formed by SIRPα-expressing renal cell
carcinoma or melanoma cells in syngeneic mice. (A)
Whole lysates of RENCA or B16BL6 cells were subject-
ed to immunoblot analysis with mAbs against mouse
signal regulatory protein α (SIRPα) (P84) and against
β-tubulin as a loading control (left panel). RENCA and
B16BL6 cells were also subjected to flow cytometric
analysis of SIRPα expression on the surface of viable
cells after staining with propidium iodide (PI), a
biotin-conjugated mAb against mouse SIRPα (P84)
(or isotype control), and allophycocyanin-conjugated
(APC-conjugated) streptavidin (right panel). Data are
representative of 3 separate experiments. (B and C)
Tumor volume (left panels) and survival curves (right
panels) for BALB/c mice injected s.c. with RENCA
cells and treated i.p. with control IgG or mAbs against
mouse SIRPα (MY-1 or P84) according to the indicated
schedule, beginning either immediately after injec-
tion of tumor cells (B) or after tumor volume had
achieved an average size of 100 mm3 (C). (D) Number
of tumor nodules in the lungs of C57BL/6J mice 14
days after i.v. injection of B16BL6 cells and the onset
of i.p. administration of control IgG or mAbs against
mouse SIRPα (MY-1 or P84) according to the indicated
schedule. Data represent the mean ± SEM for n = 8
(B), n = 11 (C), or n = 10 (D) mice per group examined
in 2 separate experiments. *P < 0.05, **P < 0.01, and
***P < 0.001, by 2-way ANOVA with Tukey’s test (left
panels in B and C) or log-rank test (right panels in B
and C), or by 1-way ANOVA with Tukey’s test (D).
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

5
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
stimulatory effect of F(ab′)2 fragments of MY-1 on such phagocytosis was significantly smaller than that
of the intact Ab (Figure 3E). Knockdown of SIRPα in CFSE-labeled RENCA cells markedly attenuated
the MY-1–promoted phagocytosis by BMDMs (Figure 3F). Moreover, coculture of CFSE-labeled RENCA
cells with BMDMs preincubated with either intact MY-1 or F(ab′)2 fragments of MY-1 failed to stimulate
the phagocytosis of the tumor cells by BMDMs (Supplemental Figure 8A), suggesting that opsonization
of RENCA cells by intact MY-1 contributes to the promotion of macrophage-mediated ADCP. On the
other hand, the effect of MY-1 on phagocytosis was much greater than that of P84 (Figure 3, C and D).
Moreover, F(ab′)2 fragments of MY-1 retained the inhibitory effect of the intact Ab on the CD47-SIRPα
interaction (Supplemental Figure 6C) and had a significant stimulatory effect on phagocytosis compared
with control IgG (Figure 3E). By contrast, incubation of BMDMs with CFSE-labeled RENCA cells prein-
cubated with either intact MY-1 or F(ab′)2 fragments of MY-1 did not promote phagocytosis of the tumor
cells by BMDMs (Supplemental Figure 8B). These results thus suggested that blockade of the inhibitory
signal provided by the interaction of CD47 (on cancer cells) with SIRPα (on macrophages) also contributes
to the promotion of phagocytosis in macrophages by MY-1. In contrast, treatment of RENCA cells with
MY-1 did not influence cell viability (Supplemental Figure 9).
Macrophages are broadly classified into M1 and M2 types, which are thought to have antitumori-
genic and protumorigenic functions, respectively (30, 31). We found that the frequency of macrophages
in tumors formed by RENCA cells 14 days after cell injection did not differ between mice treated with
MY-1 or control IgG (Figure 4A). However, the ratio of M1 to M2 macrophages in the tumors of
MY-1–treated mice was significantly higher than that in the tumors of control IgG–treated mice (Figure
4A), suggesting that MY-1 treatment does not affect the number of macrophages in tumors but rather
increases the M1/M2 ratio.
Contribution of NK cells and CD8+ T cells to MY-1–induced inhibition of tumor growth. To investigate wheth-
er other types of immune cells participate in the inhibition of tumor growth by MY-1 in vivo, we first
examined the effect of MY-1 on the population of immune cells in tumors formed by implanted RENCA
cells. Fourteen days after tumor cell injection, the numbers of both NK cells and T cells in the tumors
were markedly increased in mice treated with MY-1 compared with those treated with control IgG (Figure
4B). In particular, the number of CD8+ T cells, but not that of CD4+ T cells, was significantly increased
in the tumors of MY-1–treated mice (Figure 4B). By contrast, the number of CD11b+Gr-1+ cells including
tumor-associated neutrophils, which possess protumourigenic or antitumour activity (32), in the tumors
of MY-1–treated mice was similar to that in the tumors of mice treated with control IgG (Figure 4C).
Moreover, CD3+CD4+Foxp3+ Tregs, which are thought to suppress tumor immunosurveillance (33), in
the tumors of MY-1–treated mice were increased compared with those in the tumors of mice treated with
control IgG (Figure 4C). We therefore next examined whether NK cells or CD8+ T cells contribute to the
inhibition of tumor growth by MY-1. Treatment of BALB/c mice with Abs against asialoganglioside GM1
(asialo-GM1) (34) effectively depleted NK cells from the spleen as well as markedly attenuated the antitu-
mor effect of MY-1 for RENCA cells (Figure 5A). In addition, treatment with a mAb against CD8α (2.43)
(35) eliminated CD8+ T cells from the spleen as well as inhibited the antitumor effect of MY-1 (Figure 5B).
By contrast, depletion of CD4+ Τ cells with a mAb against CD4 failed to influence the inhibitory effect of
Table 2. Expression of SIRPα in tumor samples from patients with malignant melanoma
Case Age (yr) Sex Source TNM Stage Type SIRPα expression
1 55 F Right upper arm T4aN1M0 IIIA NM ++
276 MRight sole T2aN0M0 IB ALM ±
3 80 M Toe of left foot T3bN0M0 IIB NM ++
4 55 M Left upper arm T4bN0M0 IIC NM ++
5 39 F Head T4bN3M0 IIIC SSM ±
649 FRight lower back T4bN3M0 IIIC NM +
7 67 M Abdomen T4bN0M0 IIC NM +
8 66 M Left heel T2aN0M0 IB NM ++
M, male; F, female; ALM, acral lentiginous melanoma; NM, nodular melanoma; SSM, superficial spreading melanoma; SIRPα, signal regulatory protein α;
TNM, tumor, node, metastasis system; ++, high; +, moderate; ±, low.
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

6
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
MY-1 on tumor formation by RENCA cells (Supplemental Figure 10). We also found that treatment of
BALB/c mice bearing established RENCA tumors with Abs against asialo-GM1 or a mAb against CD8α
effectively depleted NK cells or CD8+ T cells from the spleen and tumors as well as markedly reduced the
antitumor effect of MY-1 (Supplemental Figure 11). Together, these results thus suggested that both NK
cells and CD8+ T cells, but not CD4+ T cells, participate in the suppressive effect of MY-1 on tumor forma-
tion and growth by SIRPα-expressing cancer cells in vivo.
Enhancement by MY-1 of rituximab-induced inhibition of tumor formation by Burkitt’s lymphoma (Raji) cells.
Given that Abs against SIRPα were previously shown to enhance the killing of trastuzumab-opsonized
HER2-positive breast cancer cells by phagocytes in vitro (17), we next examined the effects of mAbs
against SIRPα on ADCP-dependent inhibition of tumor growth in vivo. Human CD47 was shown to bind
to SIRPα from NOD mice but not to that from other mouse strains such as C57BL/6 (36). CD47 expressed
on Raji cells, a human Burkitt’s lymphoma cell line that does not express SIRPα, would thus be expected
to bind to SIRPα on NOD mouse macrophages. We therefore transplanted Raji cells s.c. into 6-week-
old NOD/SCID mice. Beginning 1 week after cell injection, the animals were treated i.p. twice a week
with the anti-CD20 mAb rituximab (37) either alone or together with MY-1 or P84. Treatment with MY-1
alone had little effect on the growth of tumors formed by Raji cells (Supplemental Figure 12), whereas the
combination of MY-1 with a suboptimal dose of rituximab greatly attenuated tumor growth compared
with the effect of rituximab alone (Figure 6A). In addition, this potentiating effect of MY-1 was mark-
edly greater than that of P84 (Figure 6A), suggesting that disruption of the CD47-SIRPα interaction is
Figure 3. Importance of macrophages for the antitumor eect of
the MY-1 mAb against mouse SIRPα in vivo. (A) BALB/c mice were
injected i.v. with either PBS liposomes as a control (Ctrl) or clodro-
nate liposomes (Clodronate). Splenocytes isolated from the mice 3
days later were stained with propidium iodide (PI), a brilliant violet
(BV) 510–conjugated mAb against CD45, a BV 421–conjugated mAb
against CD11b, and a phycoerythrin-conjugated (PE-conjugated)
mAb against F4/80 for analysis by flow cytometry. The relative
number of F4/80+CD11b+ cells (macrophages) is expressed as a per-
centage of all viable CD45+ cells on each plot. (B) BALB/c mice were
injected with PBS liposomes or clodronate liposomes, RENCA cells,
and either MY-1 or control IgG according to the indicated schedule.
Tumor volume was measured at the indicated time points. (C and
D) CFSE-labeled RENCA cells were incubated for 4 hours with
BALB/c mouse BM-derived macrophages (BMDMs) in the presence
of the indicated Abs, after which cells were harvested, stained
with a biotin-conjugated mAb against F4/80 and allophycocyanin-
conjugated (APC-conjugated) streptavidin as well as with PI, and
analyzed by flow cytometry. The relative number of CFSE+F4/80+
BMDMs (BMDMs that had phagocytosed CFSE-labeled RENCA
cells) is expressed as a percentage of all viable F4/80+ cells in the
representative plots (C) as well as for a representative experiment
(D). (E) CFSE-labeled RENCA cells were incubated for 4 hours with
BMDMs in the presence of control IgG or of intact or F(ab′)2 frag-
ments of MY-1. The percentage of CFSE+F4/80+ BMDMs among
viable F4/80+ cells was then determined as in C. (F) RENCA cells
were transfected with signal regulatory protein α (Sirpa) or control
siRNAs, after which cells were labeled with CFSE and stained with
PI, a biotin-conjugated mAb against mouse SIRPα (P84) (or iso-
type control), and APC-conjugated streptavidin for analysis by flow
cytometry (left panel). The CFSE-labeled RENCA cells were also
incubated for 4 hours with BMDMs in the presence of the indicated
Abs. The percentage of CFSE+F4/80+ BMDMs among viable F4/80+
cells was then determined as in C (right panel). Data are represen-
tative of 3 separate experiments (A and C–F); the mean ± SEM of
triplicate determinations (n = 3) (D, E, and right panel in F), or the
mean ± SEM for n = 8 mice per group in 2 separate experiments
(B). ***P < 0.001, by 2-way (B) or 1-way (D, E, and right panel in F)
ANOVA with Tukey’s test.
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

7
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
likely important for this effect of MY-1. MY-1 also
enhanced the inhibitory effect of rituximab on
tumor growth when treatment was delayed, until
tumors had achieved an average size of 150 to 200
mm3 (Figure 6B). Furthermore, rituximab promot-
ed the phagocytosis of CFSE-labeled Raji cells by
BMDMs from NOD mice, and MY-1 markedly
enhanced this effect of rituximab (Supplemental
Figure 13). P84 also enhanced such rituximab-
induced phagocytosis, albeit to a lesser extent than
did MY-1 (Supplemental Figure 13).
Impact of combination therapy with MY-1 and a
mAb against PD-1 on tumor growth in vivo. Immuno-
therapy with Abs against PD-1 that block the inhib-
itory PD-1/PD-L1 axis has recently been found to
provide substantial clinical benefit in patients with
a wide range of cancer types (5). Such Abs also
exert antitumor activity in immunocompetent mice
implanted with various types of mouse cancer cells
including colon cancer cells (38, 39). We therefore
examined the impact of combination therapy with
MY-1 and an Ab against PD-1 (4H2) that blocks
the PD-1–PD-L1 interaction (39) on the growth of
tumors formed by mouse CT26 colon cancer cells in
BALB/c mice. Flow cytometry with a mAb against
PD-L1 revealed that CT26 cells express PD-L1
on the cell surface, whereas flow cytometry with
either P84 or MY-1 revealed that these cells do not
express SIRPα (Figure 6C). The expression level of
SIRPα on tumors in mice transplanted with CT26
cells was also minimal (data not shown). Whereas
MY-1 alone had no effect on the growth of tumors
formed by CT26 cells, it markedly enhanced the
suppressive effect of the Ab against PD-1 on tumor
growth (Figure 6D).
Discussion
Both renal cell carcinoma and advanced or meta-
static melanoma have a poor prognosis (40–42).
Immunotherapy with either IFN-α or IL-2 or che-
motherapy with dacarbazine is thus largely inef-
fective for patients with these cancers, and these
therapies are also associated with serious adverse
effects (42–44). In contrast, immunotherapy with
ipilimumab (a mAb against CTLA-4) or nivolum-
ab (a mAb against PD-1), both of which are thought to abrogate the inhibitory checkpoint for cytotoxic T
cell activity, has recently been approved as a promising treatment for both advanced melanoma and renal
cell carcinoma (42–44). Effective immunotherapy with Abs that target tumor-specific antigens on the cell
surface of these carcinomas and thereby kill the cancer cells (by ADCC or ADCP) has yet to be developed.
We have now shown that the amount of SIRPA mRNA in tumor tissue from patients with renal cell car-
cinoma was significantly increased compared with that in matched normal tissue. The expression of SIRPα
protein was also prominent in the tumor tissue from patients with renal cell carcinoma or melanoma as well
as in corresponding cancer cell lines. Treatment with the MY-1 mAb against SIRPα, which blocks the binding
of CD47 to SIRPα, resulted in a marked reduction in the tumor burden of immunocompetent mice injected
Figure 4. Impact of MY-1 on the proportion of immune cells in RENCA tumors. (A–C) BALB/c
mice were injected s.c. with RENCA cells and treated with control IgG or MY-1 as in Figure 2B.
Tumors were harvested 14 days after cell injection for isolation of infiltrating cells. The isolated
cells were stained with propidium iodide (PI) and a brilliant violet (BV) 510–conjugated mAb
against CD45 as well as with either a phycoerythrin-conjugated (PE-conjugated) mAb against
F4/80 and an allophycocyanin-conjugated (APC-conjugated) mAb against Ly6C (A, left panel),
a PE-conjugated mAb against F4/80, an APC-conjugated mAb against Ly6C, an FITC-conju-
gated mAb against mouse CD206, a biotin-conjugated mAb against MHC class II, and PE- and
Cy7-conjugated streptavidin (A, right panel), a PE-conjugated mAb against CD3ε and an FITC-
conjugated mAb against CD49b (B, left top panel), an FITC-conjugated mAb against CD3ε (B,
right top panel), an FITC-conjugated mAb against CD3ε, an APC-conjugated mAb against CD4,
and a PE-conjugated mAb against CD8α (B, bottom panels), a BV 421–conjugated mAb against
CD11b and a PE-conjugated mAb against Ly6G/Ly6C (C, top panel), or an FITC-conjugated
mAb against CD3ε, an APC-conjugated mAb against CD4, and a PE-conjugated mAb against
Foxp3 (C, bottom panel). The cells were then subjected to flow cytometry for determination
of the frequencies of F4/80+Ly6Clo cells (Macrophages) (A, left panel), CD3ε–CD49b+ (NK) cells,
CD3ε+ (T) cells, CD3ε+CD4+CD8α– (CD4+ T) cells, and CD3ε+CD4–CD8α+ (CD8+ T) cells (B), or of
CD11b+Gr-1+ cells and CD3ε+CD4+Foxp3+ cells (Tregs) (C), as well as the ratio of F4/80+Ly6CloMHC
IIhiCD206lo cells (M1 macrophages) to F4/80+Ly6CloMHC IIloCD206hi cells (M2 macrophages) (A,
right panel), among all viable CD45+ cells. Data represent the mean ± SEM for n = 6 (IgG) or n
= 7 (MY-1) mice in 2 separate experiments (A and B); or for n = 8 (IgG) or n = 9 (MY-1) mice in 2
separate experiments (C). *P < 0.05 and **P < 0.01, by 2-tailed Student’s t test.
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

8
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
with syngeneic renal cell carcinoma (RENCA)
or melanoma (B16BL6) cells. In addition,
depletion of macrophages by injection of clo-
dronate liposomes attenuated the antitumor
effect of MY-1. We also found that intact MY-1
greatly enhanced the phagocytosis of SIRPα-
expressing tumor cells by macrophages in vitro,
whereas the enhancing effect of F(ab′)2 frag-
ments of MY-1 was less pronounced. In addi-
tion, knockdown of SIRPα in RENCA cells
resulted in the attenuation of MY-1–induced
phagocytosis by macrophages, suggesting that
the anti-SIRPα Ab prevents tumor formation
in part by promoting ADCP of cancer cells by
macrophages. In contrast, treatment of mice
injected with RENCA or B16BL6 cells with a
mAb against SIRPα (P84) that does not block
the interaction of CD47 with SIRPα failed to
prevent tumor growth, suggesting that blockage
of the CD47-SIRPα interaction is also impor-
tant for the inhibitory effect of MY-1 on tumor
growth. Targeting of SIRPα by Abs such as
MY-1 may therefore constitute a potential new
immunotherapeutic approach to the treatment
of cancers such as renal cell carcinoma and
melanoma that express SIRPα at a high level,
with the antitumor effect of such Abs being
dependent on a dual mechanism of action.
We also investigated the effect of MY-1
on the immune microenvironment of tumors
formed by RENCA cells in syngeneic mice.
Whereas treatment with MY-1 did not affect
the number of macrophages in the tumors, it
resulted in a significant increase in the pro-
portion of M1 macrophages. Macrophages in
human malignant tumors are thought to dif-
ferentiate predominantly into those of the M2
type, which possess protumorigenic activity
and are implicated both in the abrogation of
antitumor immunity and in tumor progression
(30, 45). The suppression of tumor growth
by MY-1 is thus also likely achieved in part
through regulation of the M1-versus-M2 differentiation of tumor-associated macrophages. Indeed, knock-
down of SIRPα in macrophages cocultured with hepatoma cells was found to result in increased proinflam-
matory cytokine production through activation of the NF-κB signaling pathway (46), a phenotype consis-
tent with that of M1 macrophages (45), suggesting that SIRPα regulates a switch in macrophage phenotype
in the tumor microenvironment. In contrast, it remains unknown whether such an effect of MY-1 on the
M1/M2 ratio is relevant to the antitumor effect of MY-1. Indeed, Leidi et al. showed that M2 macrophages
exhibited higher phagocytic activity toward B-chronic lymphocytic leukemia or lymphoma opsonized with
an Ab against CD20 than did M1 macrophages in vitro (47).
We found that treatment of mice bearing RENCA cell tumors with MY-1 also increased the fre-
quency of NK cells and CD8+ T cells in the tumors. Moreover, depletion of either of these cell types
resulted in marked attenuation of the antitumor effect of MY-1, implicating these immune cells in this
effect. It has recently been demonstrated that the efficacy of CD47 blockage against tumors required
Figure 5. Contribution of NK cells and CD8+ T cells to inhibition of tumor growth by MY-1. (A)
BALB/c mice were injected with either vehicle (Ctrl) or polyclonal Abs (pAbs) against asialoganglio-
side GM1 (asialo-GM1) (α-GM1), 4 days after which splenocytes were isolated from the mice, sub-
jected to staining with a brilliant violet (BV) 510–conjugated mAb against CD45, a phycoerythrin-
conjugated (PE–conjugated) mAb against CD3ε, and an FITC-conjugated mAb against CD49b, as
well as staining with propidium iodide (PI), and analyzed by flow cytometry. The relative number
of NK cells is expressed as a percentage of all viable CD45+ splenocytes on each plot (top panel).
BALB/c mice were also treated with either vehicle or pAbs against asialo-GM1, injected with REN-
CA cells, and treated with MY-1 or control IgG according to the indicated schedule for measurement
of tumor volume at the indicated time points (bottom panel). (B) BALB/c mice were treated with
either vehicle (Ctrl) or a mAb against CD8α (α-CD8α), 4 days after which splenocytes were isolated
from the mice, subjected to staining with a BV 510–conjugated mAb against CD45, an FITC-conju-
gated mAb against CD3ε, an allophycocyanin-conjugated (APC-conjugated) mAb against CD4, and
a PE-conjugated mAb against CD8α as well as staining with PI, and analyzed by flow cytometry.
The relative number of CD8+ T cells is expressed as a percentage of all viable CD45+CD3ε+ spleno-
cytes on each plot (top panel). BALB/c mice were also treated with either vehicle or a mAb against
CD8α, injected with RENCA cells, and treated with MY-1 or control IgG according to the indicated
schedule for measurement of tumor volume at the indicated time points (bottom panel). Data are
representative of 3 separate experiments (A and B, top panels) or represent the mean ± SEM for n
= 10 mice per group in 2 separate experiments (A, bottom panel); or for n = 10 (IgG, MY-1, or MY-1 +
α-CD8α) or n = 9 (IgG + α-CD8α) mice in 2 separate experiments (B, bottom panel). ***P < 0.001, by
2-way ANOVA with Tukey’s test.
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

9
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
adaptive immune responses in immu-
nocompetent mouse tumor models
(48–50). The inhibition of the CD47-
SIRPα interaction by SIRPα or CD47
blockage is thus probably impor-
tant for enhancing T cell–mediated
destruction of tumors. The mecha-
nism by which treatment with MY-1
promotes the antitumor immune
response mediated by NK cells and
CD8+ T cells remains unclear, how-
ever. We found that MY-1 did not
promote NK cell–mediated cytotoxic
activity toward RENCA cells in vitro
(Supplemental Figure 14). Given
that macrophages are likely the pri-
mary effector cells for the antitumor
activity of MY-1, it is possible that
they participate in the promotion of
NK cell–dependent killing of tumor
cells by the Ab. Indeed, NK cells
are thought to be primed by interac-
tion with activated M1 macrophages
or DCs mediated either directly
through cell-cell contact or indirectly
by cytokines, and the primed cells
are then thought to contribute to the
eradication of tumor cells (51, 52).
Macrophages or DCs also recognize
N-glycan structures on tumor cells
through the innate immune receptor
Dectin 1, resulting in enhancement of
NK cell–meditated killing of tumor
cells (53). In addition, macrophages
engulf tumor cells and then cross-
present tumor antigens to CD8+ T
cells via the MHC class I route, there-
by enhancing the activity of tumor
antigen–specific T cells toward tumor
cells. CD169+ macrophages were thus
shown to engulf dead tumor cells and
to cross-present tumor antigens to CD8+ Tcells, thereby inducing antitumor immunity (54). Moreover,
CD47 blockage enhanced the cross-priming of CD8+ T cell responses by DCs, but not by macrophages,
and thereby contributed to tumor control in mice injected with syngeneic tumor cells (48). DCs might
participate in CD8+ T cell–mediated antitumor effects of Abs against SIRPα.
Our study has shown that Abs against SIRPα in combination with other anticancer Abs may be effec-
tive even for the treatment of tumors that do not express SIRPα. We thus found that MY-1, but not
P84, markedly enhanced the suppressive effect of rituximab on the growth of tumors formed by human
Burkitt’s lymphoma Raji cells in immunodeficient mice. Consistent with this finding, MY-1 also enhanced
the phagocytosis of rituximab-opsonized Raji cells by macrophages, suggesting that disruption of the
CD47-SIRPα interaction by MY-1 promotes rituximab-induced ADCP, resulting in enhanced inhibition
of tumor growth. We also found that the combination of MY-1 with an Ab against PD-1 yielded a syn-
ergistic antitumor effect against SIRPα-negative mouse colon cancer cells in vivo. In addition, such anti-
tumor activity induced by the combination of MY-1 with an Ab against PD-1 was not affected by the
Figure 6. Impact of combination therapy with MY-1 and either rituximab or a mAb against PD-1 on
tumor growth in vivo. (A and B) NOD/SCID mice were injected s.c. with Raji cells and then treated with
the indicated combinations of Abs according to the indicated schedule, beginning either when the
tumors became palpable (on day 7) (A) or when they had achieved an average size of 150 to 200 mm3
(B). (C) CT26 cells were incubated with a biotin-conjugated mAb against PD-L1 (anti–PD-L1) (or isotype
control) or with mAbs against mouse signal regulatory protein α (SIRPα) (P84 or MY-1) (or an isotype con-
trol). The cells were then stained with propidium iodide (PI) and with either allophycocyanin-conjugated
(APC-conjugated) streptavidin or Alexa Fluor 488–conjugated polyclonal Abs (pAbs) against rat IgG for
determination of cell-surface expression of PD-L1 and SIRPα by flow cytometry. (D) Tumor volume for
BALB/c mice injected s.c. with CT26 cells and treated with control IgG, MY-1, or a mAb against PD-1 (α–
PD-1), beginning after tumors had achieved an average size of 100 mm3. Data represent the mean±SEM
for n = 5 (A and B) or n = 6 (D) mice per group or are representative of 3 separate experiments (C). **P <
0.01 and ***P < 0.001, by 2-way ANOVA with Tukey’s test.
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

10
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
depletion of macrophages in mice bearing CT26 cell tumors (data not shown), suggesting that MY-1
synergizes the efficacy of the PD-1 Ab without the promotion of macrophage-mediated ADCP toward
SIRPα-nonexpressing tumor cells. Interestingly, it was recently demonstrated that CD47 blockage syner-
gized with PD-L1 antagonism to potentiate the attenuation of tumor growth in immunocompetent mice
injected s.c. with syngeneic B16F10 melanoma cells, whereas CD47 blockage alone or in combination
with a tumor-specific Ab failed to prevent the tumor formation (49). Together, these findings suggest that
targeting both the CD47/SIRPα and PD-1/PD-L1 axes provides a new approach to immunotherapy for
a broad range of cancers. Further investigation is warranted to elucidate the mechanism underlying these
antitumor effects of the anti-SIRPα Ab, however.
Finally, we confirmed that MY-1 had no marked adverse effects on hematologic or blood biochemical
parameters in mice. Although Abs against CD47 are thought to hold promise for the treatment of various
types of cancer (55), they have also been shown to have undesired effects such as a marked reduction in the
number of rbc (anemia), probably as a result of their triggering of ADCC or ADCP directed toward rbc,
which express CD47 at a high level (56). With regard to adverse effects, therefore, Abs against SIRPα may
be a better choice for the development of anticancer drugs that target the CD47/SIRPα axis.
Methods
Additional details on methods can be found in the supplemental methods and figures.
Abs and reagents. Rat mAbs against mouse SIRPα (MY-1 [rat IgG2a] and P84 [rat IgG1]; provided by Carl
F. Lagenaur, University of Pittsburgh, Pittsburgh, Pennsylvania, USA) and against mouse SIRPβ were gener-
ated and purified as described previously (26–29). Α mouse mAb against the Myc epitope tag (clone 9E10)
was from Santa Cruz Biotechnology Inc. Rituximab (mAb against human CD20) was obtained from Chugai
Pharmaceutical. A mAb against mouse PD-1 (4H2, a chimeric rat Ab with a murine IgG1 constant region)
was from Ono Pharmaceutical. A rabbit mAb against MART-1 (clone EP1422Y) and rabbit pAbs against
SIRPα (ab53721 and ab139698), which were generated against the cytoplasmic region of human SIRPα, were
from Abcam. An FITC-conjugated mAb against mouse CD8α (clone 53-6.7), a phycoerythrin-conjugated (PE-
conjugated) mAb against mouse CD3ε (clone 145-2C11), an allophycocyanin-conjugated (APC-conjugated)
mAb against mouse CD4 (clone RM4-5), and a biotin-conjugated rat IgG2bκ isotype control (clone A95-1)
were from BD Biosciences. A PE-conjugated mAb against F4/80 (clone BM8), a rat mAb against mouse
CD16/CD32 (clone 93), a biotin-conjugated mAb against MHC class II (clone M5/114.15.2), a biotin-con-
jugated mAb against SIRPα (clone P84), an APC-conjugated mAb against human CD47 (clone B6H12), and
an APC-conjugated mouse IgG1κ isotype control (clone P3.6.2.8.1) were from eBioscience. APC-conjugated
streptavidin, FITC-conjugated streptavidin, an FITC-conjugated mAb against mouse CD49b (clone DX5),
an FITC-conjugated mAb against SIRPα (clone P84), PE- and Cy7-conjugated streptavidin, a brilliant violet
421–conjugated mAb against mouse CD11b (clone M1/70), a brilliant violet 510–conjugated mAb against
mouse CD45 (clone 30-F11), an Alexa Fluor 488–conjugated mAb against mouse CD3ε (clone 145-2C11),
a PE-conjugated mAb against mouse CD8α (clone 53-6.7), an FITC-conjugated mAb against mouse CD206
(clone C068C2), a biotin-conjugated mAb against F4/80 (clone BM8), a biotin-conjugated mAb against
PD-L1 (clone 10F.9G2), an APC-conjugated mAb against mouse Ly6C (clone HK1.4), a PE-conjugated mAb
against Ly6G/Ly6C (clone RB6-8C5), a PE-conjugated mAb against mouse Foxp3 (clone 150D), a mAb
against human CD172b (SIRPβ, clone B4B6), a mAb against CD172g (SIRPγ, clone LSB2.20), a PE- and
Cy7-conjugated mouse IgG1κ isotype control (clone MOPC-21), and a biotin-conjugated rat IgG1κ isotype
control (clone RTK-2071) were from BioLegend. HRP-conjugated goat pAbs against rabbit, mouse, or rat IgG;
Cy3-conjugated goat pAbs against rabbit, rat, or mouse IgG, as well as normal rat or mouse IgG were from
Jackson ImmunoResearch Laboratories. A mouse IgG1 isotype control (clone G3A1) was from Cell Signal-
ing Technology. Rat IgG1 (clone 43414) and IgG2 (clone 54447) isotype controls were from R&D Systems. A
mouse mAb against β-tubulin and propidium iodide (PI) were from Sigma-Aldrich. Rat mAbs against mouse
CD8α (clone 2.43) and against CD4 (clone GK1.5) were from Bio X cell. Rabbit pAbs against mouse asialo-
GM1 were from Wako. Alexa Fluor 488–conjugated goat pAbs against rat, mouse, or rabbit IgG and Alexa
Fluor 647–conjugated goat pAbs against rat IgG as well as CFSE were from Thermo Fisher Scientific.
Animals. NOD, NOD/ShiLtJ-Prkdcscid (NOD/SCID), C57BL/6J, and BALB/c mice, which were
obtained from Charles River Laboratories Japan, Japan SLC (Shizuoka, Japan), or CLEA Japan (Tokyo,
Japan), were maintained in the Institute for Experimental Animals at the Kobe University Graduate School
of Medicine under specific pathogen–free conditions.
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

11
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
Patients and tissue samples. Paired tumor and noncancerous renal cortex tissue specimens were
obtained from material surgically resected from 95 patients with primary clear cell renal cell carcinoma
at the National Cancer Center Hospital (Tokyo, Japan). Noncancerous renal cortex tissue consisted
mostly of proximal tubules, which are the origin of clear cell renal cell carcinoma. Tissue specimens
were provided by the National Cancer Center Biobank (Tokyo, Japan). Human melanoma tissue was
obtained during surgical resection from 8 patients treated at the Department of Dermatology, Gunma
University Hospital (Gunma, Japan).
Analysis of SIRPA mRNA abundance in human clear cell renal cell carcinoma. Total RNA was isolated
from paired cancerous tissue and noncancerous renal cortex specimens from 95 patients with clear cell
renal cell carcinoma using TRIzol reagent (Thermo Fisher Scientific) and was subjected to expression
microarray analysis as described previously (23). In brief, fluorescent complementary RNA (cRNA)
was produced from the total RNA (200 ng) and subjected to hybridization with a SurePrint G3 Human
Gene Expression 8 × 60 K microarray (Agilent Technologies). The signal for the probe correspond-
ing to SIRPA (ID: A_23_P210708) was extracted with the use of Feature Extraction software (Agilent
Technologies). Microarray analysis data were deposited in the publicly available Integrative Disease
Omics Database (http://gemdbj.ncc.go.jp/omics/biomart/martform/#!/Analysis/summary_mRNA_
array?datasets=cancer_kidney), as previously described (23).
Generation of mouse mAbs against human SIRPα. BALB/c mice were injected with a fusion protein
consisting of the extracellular portion of human SIRPα fused to the Fc portion of human IgG1, and
hybridomas were selected on the basis of positive staining by the released Abs of human neutrophils,
monocytes, or THP-1 or U937 cells as detected by flow cytometry (57). The mAbs were purified from
serum-free culture supernatants of the selected hybridoma cells by column chromatography with protein
G Sepharose 4 Fast Flow (GE Healthcare). The 040 mAb thus generated was used in the present study.
H&E staining and immunostaining. Tumor tissue from patients with clear cell renal cell carcinoma was
fixed with formalin, embedded in paraffin, sectioned at a thickness of 3 μm, and stained with H&E. Par-
affin-embedded sections were also subjected to immunohistochemical staining with the use of EnVision+
System-HRP (Agilent Technologies). In brief, sections were depleted of paraffin, rehydrated, immersed in
10 mM citrate buffer (pH 6.0), and heated in a pressure cooker for 3 minutes to facilitate antigen retrieval.
They were then washed with TBS-T (20 mM Tris, pH 7.5, 140 mM NaCl, 0.01% Tweeen-20), treated for
5 minutes with Peroxidase Block (EnVision+ System-HRP) to quench endogenous peroxidase activity,
exposed for 30 minutes to TBS-T containing 1% BSA, and incubated for 30 minutes with primary Abs
diluted in TBS-T. The sections were again washed with TBS-T, incubated for 30 minutes with peroxidase-
labeled polymer conjugated to goat Abs against rabbit IgG (EnVision+ System-HRP), and washed further
with TBS-T, after which immune complexes were detected by exposure to 3,3′-diaminobenzidine chromo-
gen solution (EnVision+ System-HRP). The sections were finally counterstained with Mayer’s hematoxylin
before observation with a BX51 microscope (Olympus).
Fresh-frozen sections (4 μm thickness) prepared from human melanoma tissue were fixed with 4%
paraformaldehyde in PBS, exposed for 1 hour to PBS containing 3% nonfat dried milk and 5% normal
goat serum, and then incubated overnight with primary Abs. The sections were then washed with PBS,
incubated for 1 hour with corresponding Cy3- or Alexa Fluor 488–conjugated secondary Abs, and stained
with DAPI. Fluorescence images were acquired with a BX51 fluorescence microscope (Olympus). For
immunofluorescence staining of cultured cells, the cells were fixed for 10 minutes with 4% paraformalde-
hyde, incubated for 30 minutes with buffer G (PBS containing 5% goat serum and 0.1% Triton X-100), and
stained with primary Abs in the same buffer. Immune complexes were detected with dye-labeled secondary
Abs in buffer G, and the cells were then examined with an Olympus BX51 fluorescence microscope.
Cell culture. A human Burkitt’s lymphoma cell line (Raji), human melanoma cell lines (A375, SK-
MEL-5, and SK-MEL-28), human renal cell carcinoma cell lines (ACHN, 786-O, A498, and Caki-1), a
mouse renal carcinoma cell line (RENCA), and a mouse colon cancer cell line (CT26) were obtained
from American Type Culture Collection (ATCC). HEK293A and FreeStyle 293-F cells were obtained from
Thermo Fisher Scientific. The human melanoma cell line WM239a and the mouse melanoma cell line
B16BL6 were provided by Meenhard Herlyn (The Wistar Institute, Philadelphia, Pennsylvania, USA) and
Kazuyoshi Takeda and Ko Okumura (Juntendo University, Tokyo, Japan), respectively. Raji, 786-O, REN-
CA, CT26, and B16BL6 cells were maintained in RPMI 1640 medium (Wako) supplemented with 10%
FBS. A375 cells were cultured in DMEM (Wako) supplemented with 10% FBS. SK-MEL-5, SK-MEL-28,
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

12
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
ACHN, and A498 cells were maintained in Eagle’s minimum essential medium (Wako) supplemented with
10% FBS. Caki-1 cells were cultured in McCoy’s 5A medium (Thermo Fisher Scientific) supplemented with
10% FBS. WM239a cells were maintained in medium W489, a 4:1 (v/v) mixture of MCDB153 (Sigma-
Aldrich), and L15 (Thermo Fisher Scientific), supplemented with 2 mM CaCl2, 2% FBS, and bovine insulin
(5 μg/ml). CHO cells stably expressing an active form of H-Ras (CHO-Ras cells) were provided by Sane-
take Shirahata (Kyushu University, Fukuoka, Japan), and CHO-Ras cells stably expressing mouse SIRPα,
mouse CD47, or human SIRPα were provided by Nakayuki Honma (Kyowa Hakko Kirin, Tokyo, Japan)
(58, 59). CHO-Ras cells and their derivatives were cultured in α-modified minimum essential medium
(Sigma-Aldrich) supplemented with 2mM L-glutamine, 10mM HEPES-NaOH (pH 7.4), and 10% FBS.
Immunoblot analysis. Cells were washed with ice-cold PBS and then homogenized in a solution con-
taining 50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1 mM EDTA, and 1% SDS. The lysates were heated at
95°C for 5 minutes and then centrifuged at 17,500 g for 30 minutes at room temperature, and the resulting
supernatants were subjected to immunoblot analysis as previously described (59).
Cell preparation and flow cytometry. For isolation of tumor-infiltrating cells, tumors were harvested,
minced with forceps, and then digested with HBSS containing collagenase (400 U/ml; Wako), trypsin
inhibitor (50 μg/ml; Wako), and DNase I (40 μg/ml; Roche). The undigested material was removed by
filtration through a 70-μm cell strainer (BD Biosciences), and the remaining cells were washed twice with
PBS, suspended in 6.4 ml of DMEM containing 46% OptiPrep (Alere Technologies) and 3% FBS, and then
overlaid consecutively with 3 ml DMEM containing 26% OptiPrep and 7.4% FBS and with 500 μl DMEM
containing 10% FBS. The resulting gradient was centrifuged at 800 g for 25 minutes at 20°C, after which
cells at the interface of the top 2 layers were collected, washed twice with PBS, and subjected to flow cyto-
metric analysis. For isolation of splenocytes, mouse spleen was ground gently with autoclaved frosted-glass
slides in PBS, fibrous material was removed by filtration through a 70-μm cell strainer, and rbc in the filtrate
were lysed with BD Pharm Lyse (BD Biosciences). The remaining cells were washed twice with PBS and
then subjected to flow cytometric analysis.
For flow cytometric analysis, cells were first incubated with a mAb specific for mouse CD16/CD32 to
prevent nonspecific binding of labeled mAbs against FcγR and were then labeled with specific mAbs. For
staining of Foxp3, cells were labeled with a PE-conjugated mAb against mouse Foxp3 with the use of a Tran-
scription Factor Staining Kit (TONBO Biosciences). Labeled cells were analyzed by flow cytometry using a
FACSVerse instrument (BD Biosciences), and all data were analyzed with FlowJo 9.9.3 software (Tree Star).
For determination of the expression of SIRPα on mouse cancer cell lines, cells were incubated with
a mAb against mouse CD16/CD32, washed with PBS, and then incubated first with a biotin-conjugated
mAb against SIRPα (P84) or isotype control and then with APC-conjugated streptavidin and PI. Alterna-
tively, mouse cancer cells were treated with a mAb against mouse CD16/CD32 and then stained with PI
as well as with the MY-1 mAb (or an isotype control) followed by dye-labeled secondary Abs. Stained cells
were subjected to flow cytometry, and data were analyzed with FlowJo software.
Tumor cell engraftment and treatment. RENCA cells (5 × 105 in 50 μl of PBS) were mixed with an equal
volume of Matrigel (Corning) and injected s.c. into the flanks of 8-week-old female BALB/c mice. Mice
were injected i.p. with normal rat IgG, MY-1, or P84 (each at 200 μg) 3 times a week beginning immediately
after tumor cell injection or with these same Abs (each at 400 μg) 3 times a week beginning when the tumor
volume had achieved an average of 100 mm3. Raji cells (3 × 106 in 50 μl of PBS) were mixed with an equal
volume of Matrigel and injected s.c. into the flanks of 6-week-old female NOD/SCID mice. The mice were
injected i.p. with normal rat IgG (100 μg), MY-1 (100 μg), or P84 (100 μg), each with or without rituximab
(40 μg), twice a week beginning when the tumors became palpable (on day 7), or with normal rat IgG (200
μg) or MY-1 (200 μg), each with or without rituximab (150 μg), twice a week beginning after tumor volume
had achieved an average of 150 to 200 mm3. CT26 cells (5 × 105 in 50 μl of PBS) were injected s.c. into
the flanks of 8-week-old female BALB/c mice. The mice were injected i.p. with normal rat IgG (200 μg)
or MY-1 (200 μg), each with or without anti–PD-1 Ab (100 μg), twice a week beginning when tumors had
achieved an average size of 100 mm3. Tumors were measured with digital calipers, and tumor volume was
calculated as: a × b2/2, where a is the largest diameter and b the smallest diameter. For the B16BL6 model,
female C57BL/6J mice at 8 weeks of age were injected i.v. with B16BL6 cells (5 × 104 in 100 μl of PBS) as
described previously (12) and then injected i.p. with normal rat IgG, MY-1, or P84 (each at 200 μg) 3 times
a week. Mice were sacrificed on day 14, and the number of tumor colonies formed in the lungs was counted
with the use of a dissection microscope (MZ9.5; Leica).
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

13
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
Hematologic and blood biochemical analyses. Female C57BL/6J mice at 8 weeks of age were injected i.p.
with PBS or with normal rat IgG or MY-1 (each at 100 μg) 3 times a week. On day 14, hematologic and
blood biochemical parameters were analyzed with an ADVIA 2120 Hematology Analyzer (Siemens) or an
Auto Analyzer 7070 (Hitachi), respectively.
ADCP assay. For preparation of BMDMs, BM cells were isolated from the femur and tibia of mice
using a syringe fitted with a 27-gauge needle as described previously (13), with slight modifications. The
cells (1 × 106/ml) were seeded on culture plates in Iscove’s modified Dulbecco’s medium (Nacalai Tesque)
supplemented with recombinant murine macrophage CSF (10 ng/ml; PeproTech) and 10% FBS in order
to obtain BMDMs. For ADCP assays, BMDMs were plated at a density of 1 × 105 per well in 6-well plates
and allowed to adhere overnight. Target cells (4 × 105 RENCA cells) were labeled with CFSE, added to
the BMDMs (effector cells), and incubated for 4 hours in the presence of Abs (10 μg/ml). Alternatively,
CFSE-labeled target cells and effector cells, both of which had been preincubated with either Abs (10 μg/
ml) or F(ab′)2 fragments (10 μg/ml) for 30 minutes, were washed with PBS and mixed with effector cells
and CFSE-labeled target cells, respectively, and incubated for 4 hours. Cells were then harvested, stained
for F4/80 as well as PI, and analyzed by flow cytometry. The percentage of phagocytosis by BMDMs was
calculated as: 100 × F4/80+CFSE+PI– cells/F4/80+CFSE+PI– cells + F4/80+CFSE–PI– cells.
RNAi. RNAi for endogenous mouse SIRPα was performed with the siRNA sequence 5′-CAAGCAU-
UGAGACAGGCAATT-3′ (Sirpa siRNA). The MISSION siRNA universal negative control (Sigma-
Aldrich) was also used. RENCA cells were transfected with siRNAs using Lipofectamine RNAiMAX
(Thermo Fisher Scientific).
Depletion of macrophages, NK cells, CD4+ T cells, and CD8+ T cells in vivo. Depletion of macrophages in
8-week-old female BALB/c mice was performed as described previously (60), with minor modifications. In
brief, mice were injected i.v. with 200 μl of either clodronate liposomes or PBS liposomes (Formu Max) 1
day before injection of tumor cells as well as with 100 μl of the respective liposomes every 3 days thereafter.
For NK cell depletion, 8-week-old female BALB/c mice were injected i.p. with pAbs against asialo-GM1
(50 μl) 1 day before and on the day of tumor cell injection and then every 4 days thereafter. For depletion
of CD4+ or CD8+ T cells, 8-week-old female BALB/c mice were injected i.p. with a mAb against CD4
(GK1.5, 400 μg) or against CD8α (2.43, 400 μg) 1 day before injection of tumor cells and then every 5 days
thereafter. For depletion of NK cells or CD8+ T cells in mice with established tumors, 8-week-old female
BALB/c mice were injected i.p. with pAbs against asialo-GM1 (50 μl) and a mAb against CD8α (400 μg) 3
days after injection of tumor cells and then every 4 and 5 days thereafter, respectively. The effectiveness of
macrophage, NK cell, CD4+ T cell, or CD8+ T cell depletion was determined by flow cytometric analysis
of CD45+F4/80+CD11b+, CD45+CD3ε–CD49b+, CD45+CD3ε+CD4+, or CD45+CD3ε+CD8α+ cells among
splenocytes or tumor-infiltrating cells from the treated animals.
Statistics. Data are presented as the mean ± SEM and were analyzed by a 2-tailed Student’s t test,
ANOVA followed by Tukey’s test, or a long-rank test. A P value of less than 0.05 was considered statisti-
cally significant. Analysis was performed using GraphPad Prism 6.0 (GraphPad Software).
Study approval. All animal experiments were performed according to the guidelines of the Animal Care
and Experimentation Committee of Kobe University. All patients included in this study provided informed
consent. The study was also approved by the IRB of Gunma University and the Ethics Committee of the
National Cancer Center and was performed in accordance with the tenets of the Declaration of Helsinki.
Author contributions
TY, YM, and TM designed research studies. TY, YM, DT, SM, DH, and EA conducted experiments,
acquired data, and analyzed data. EWD, KW, and NVG performed research studies. YS, T. Kotani, HO,
MM, YK, PAO, OI, T. Komori, and TM analyzed data. TY, YM, and TM wrote the manuscript.
Acknowledgments
We thank C.F. Lagenaur for the rat mAbs against mouse SIRPα, M. Herlyn for WM239a cells, K. Takeda
and K. Okumura for B16BL6 cells, S. Shirahata for CHO-Ras cells, and N. Honma for the plasmid encod-
ing human SIRPαv2 protein and CHO-Ras cells stably expressing human or mouse SIRPα or mouse CD47.
We also thank S. Hara (Kobe University, Kobe, Japan) and T. Shibata (Kyowa Hakko Kirin) for their
technical advice and helpful discussions. This work was supported in part by a Grant-in-Aid for Scientific
Research (B) from the Japan Society for the Promotion of Science (JSPS); the Program for Development of
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

14
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
Innovative Research on Cancer Therapeutics (P-Direct) from the Japan Agency for Medical Research and
Development (AMED); the Program for Promotion of Fundamental Studies in Health Sciences (10-43)
of the National Institute of Biochemical Innovation (NiBio); and by grants from the Princess Takamatsu
Cancer Research Fund (14-24626), the Suzuken Memorial Foundation, and the Hyogo Science and Tech-
nology Association. The National Cancer Center Biobank is supported by the National Cancer Center
Research and Development Fund (26-A-1).
Address correspondence to: Takashi Matozaki or Yoji Murata, Division of Molecular and Cellular Signal-
ing, Department of Biochemistry and Molecular Biology, Kobe University Graduate School of Medicine,
7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650-0017, Japan. Phone: 81.78.382.5600. E-mail: matozaki@med.
kobe-u.ac.jp (T. Matozaki); ymurata@med.kobe-u.ac.jp (Y. Murata).
NVG’s present address is: Laboratorio de Inmunología y Biología Molecular,Facultad de Medicina Veteri-
naria y Zootecnia,Universidad del Tolima, Ibagué, Columbia.
1. Hanahan D, Coussens LM. Accessories to the crime: functions of cells recruited to the tumor microenvironment. Cancer Cell.
2012;21(3):309–322.
2. Quail DF, Joyce JA. Microenvironmental regulation of tumor progression and metastasis. Nat Med. 2013;19(11):1423–1437.
3. Munn DH, Bronte V. Immune suppressive mechanisms in the tumor microenvironment. Curr Opin Immunol. 2016;39:1–6.
4. Okazaki T, Chikuma S, Iwai Y, Fagarasan S, Honjo T. A rheostat for immune responses: the unique properties of PD-1 and
their advantages for clinical application. Nat Immunol. 2013;14(12):1212–1218.
5. Kohrt HE, et al. Immunodynamics: a cancer immunotherapy trials network review of immune monitoring in immuno-oncology
clinical trials. J Immunother Cancer. 2016;4:15.
6. Groh V, Wu J, Yee C, Spies T. Tumour-derived soluble MIC ligands impair expression of NKG2D and T-cell activation. Nature.
2002;419(6908):734–738.
7. Raffaghello L, et al. Downregulation and/or release of NKG2D ligands as immune evasion strategy of human neuroblastoma.
Neoplasia. 2004;6(5):558–568.
8. Matozaki T, Murata Y, Okazawa H, Ohnishi H. Functions and molecular mechanisms of the CD47-SIRPα signalling pathway.
Trends Cell Biol. 2009;19(2):72–80.
9. Barclay AN, Van den Berg TK. The interaction between signal regulatory protein alpha (SIRPα) and CD47: structure, function,
and therapeutic target. Annu Rev Immunol. 2014;32:25–50.
10. Seiffert M, et al. Signal-regulatory protein alpha (SIRPα) but not SIRPβ is involved in T-cell activation, binds to CD47 with high
affinity, and is expressed on immature CD34+CD38– hematopoietic cells. Blood. 2001;97(9):2741–2749.
11. Ishikawa-Sekigami T, et al. SHPS-1 promotes the survival of circulating erythrocytes through inhibition of phagocytosis by
splenic macrophages. Blood. 2006;107(1):341–348.
12. Okajo J, et al. Regulation by Src homology 2 domain-containing protein tyrosine phosphatase substrate-1 of α-galactosylceramide-
induced antimetastatic activity and Th1 and Th2 responses of NKT cells. J Immunol. 2007;178(10):6164–6172.
13. Saito Y, et al. Regulation by SIRPα of dendritic cell homeostasis in lymphoid tissues. Blood. 2010;116(18):3517–3525.
14. Oldenborg PA. CD47: A cell surface glycoprotein which regulates multiple functions of hematopoietic cells in health and dis-
ease. ISRN Hematol. 2013;2013:614619.
15. Okazawa H, et al. Negative regulation of phagocytosis in macrophages by the CD47-SHPS-1 system. J Immunol.
2005;174(4):2004–2011.
16. Oldenborg PA, Zheleznyak A, Fang YF, Lagenaur CF, Gresham HD, Lindberg FP. Role of CD47 as a marker of self on red
blood cells. Science. 2000;288(5473):2051–2054.
17. Zhao XW, et al. CD47-signal regulatory protein-α (SIRPα) interactions form a barrier for antibody-mediated tumor cell destruc-
tion. Proc Natl Acad Sci U S A. 2011;108(45):18342–18347.
18. Chao MP, et al. Anti-CD47 antibody synergizes with rituximab to promote phagocytosis and eradicate non-Hodgkin lympho-
ma. Cell. 2010;142(5):699–713.
19. Willingham SB, et al. The CD47-signal regulatory protein alpha (SIRPa) interaction is a therapeutic target for human solid
tumors. Proc Natl Acad Sci U S A. 2012;109(17):6662–6667.
20. Zhao XW, Kuijpers TW, van den Berg TK. Is targeting of CD47-SIRPα enough for treating hematopoietic malignancy? Blood.
2012;119(18):4333–4; author reply 4334.
21. Uhlén M, et al. A human protein atlas for normal and cancer tissues based on antibody proteomics. Mol Cell Proteomics.
2005;4(12):1920–1932.
22. Barretina J, et al. The Cancer Cell Line Encyclopedia enables predictive modelling of anticancer drug sensitivity. Nature.
2012;483(7391):603–607.
23. Arai E, et al. Alterations of the spindle checkpoint pathway in clinicopathologically aggressive CpG island methylator pheno-
type clear cell renal cell carcinomas. Int J Cancer. 2015;137(11):2589–2606.
24. Ohsie SJ, Sarantopoulos GP, Cochran AJ, Binder SW. Immunohistochemical characteristics of melanoma. J Cutan Pathol.
2008;35(5):433–444.
25. van Beek EM, Cochrane F, Barclay AN, van den Berg TK. Signal regulatory proteins in the immune system. J Immunol.
2005;175(12):7781–7787.
26. Verjan Garcia N, et al. SIRPα/CD172a regulates eosinophil homeostasis. J Immunol. 2011;187(5):2268–2277.
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140

15
insight.jci.org doi:10.1172/jci.insight.89140
RESEARCH ARTICLE
27. Chuang W, Lagenaur CF. Central nervous system antigen P84 can serve as a substrate for neurite outgrowth. Dev Biol.
1990;137(2):219–232.
28. Ohnishi H, et al. Differential localization of Src homology 2 domain-containing protein tyrosine phosphatase substrate-1 and
CD47 and its molecular mechanisms in cultured hippocampal neurons. J Neurosci. 2005;25(10):2702–2711.
29. Hayashi A, et al. Positive regulation of phagocytosis by SIRPβ and its signaling mechanism in macrophages. J Biol Chem.
2004;279(28):29450–29460.
30. Mills CD, Lenz LL, Harris RA. A Breakthrough: macrophage-directed cancer immunotherapy. Cancer Res. 2016;76(3):513–516.
31. Guiducci C, Vicari AP, Sangaletti S, Trinchieri G, Colombo MP. Redirecting in vivo elicited tumor infiltrating macrophages and
dendritic cells towards tumor rejection. Cancer Res. 2005;65(8):3437–3446.
32. Bronte V, et al. Recommendations for myeloid-derived suppressor cell nomenclature and characterization standards. Nat Com-
mun. 2016;7:12150.
33. Liu C, Workman CJ, Vignali DA. Targeting regulatory T cells in tumors. FEBS J. 2016;283(14):2731–2748.
34. Kasai M, Yoneda T, Habu S, Maruyama Y, Okumura K, Tokunaga T. In vivo effect of anti-asialo GM1 antibody on natural
killer activity. Nature. 1981;291(5813):334–335.
35. Sriram S, Carroll L. In vivo depletion of Lyt-2 cells fails to alter acute and relapsing EAE. J Neuroimmunol. 1988;17(2):147–157.
36. Takenaka K, et al. Polymorphism in Sirpa modulates engraftment of human hematopoietic stem cells. Nat Immunol.
2007;8(12):1313–1323.
37. Reff ME, et al. Depletion of B cells in vivo by a chimeric mouse human monoclonal antibody to CD20. Blood. 1994;83(2):435–445.
38. Iwai Y, Terawaki S, Honjo T. PD-1 blockade inhibits hematogenous spread of poorly immunogenic tumor cells by enhanced
recruitment of effector T cells. Int Immunol. 2005;17(2):133–144.
39. Li B, VanRoey M, Wang C, Chen TH, Korman A, Jooss K. Anti-programmed death-1 synergizes with granulocyte macrophage
colony-stimulating factor-secreting tumor cell immunotherapy providing therapeutic benefit to mice with established tumors.
Clin Cancer Res. 2009;15(5):1623–1634.
40. Motzer RJ, et al. Overall survival and updated results for sunitinib compared with interferon alfa in patients with metastatic
renal cell carcinoma. J Clin Oncol. 2009;27(22):3584–3590.
41. Motzer RJ, Mazumdar M, Bacik J, Berg W, Amsterdam A, Ferrara J. Survival and prognostic stratification of 670 patients with
advanced renal cell carcinoma. J Clin Oncol. 1999;17(8):2530–2540.
42. Finn L, Markovic SN, Joseph RW. Therapy for metastatic melanoma: the past, present, and future. BMC Med. 2012;10:23.
43. Bedke J, Kruck S, Gakis G, Stenzl A, Goebell PJ. Checkpoint modulation- a new way to direct the immune system against renal
cell carcinoma. Hum Vaccin Immunother. 2015;11(5):1201–1208.
44. Davey RJ, van der Westhuizen A, Bowden NA. Metastatic melanoma treatment: combining old and new therapies. Crit Rev
Oncol Hematol. 2016;98:242–253.
45. Komohara Y, Niino D, Ohnishi K, Ohshima K, Takeya M. Role of tumor-associated macrophages in hematological malignan-
cies. Pathol Int. 2015;65(4):170–176.
46. Pan YF, et al. Signal regulatory protein α is associated with tumor-polarized macrophages phenotype switch and plays a pivotal
role in tumor progression. Hepatology. 2013;58(2):680–691.
47. Leidi M, et al. M2 macrophages phagocytose rituximab-opsonized leukemic targets more efficiently than M1 cells in vitro.
J Immunol. 2009;182(7):4415–4422.
48. Liu X, et al. CD47 blockade triggers T cell-mediated destruction of immunogenic tumors. Nat Med. 2015;21(10):1209–1215.
49. Sockolosky JT, et al. Durable antitumor responses to CD47 blockade require adaptive immune stimulation. Proc Natl Acad Sci
USA. 2016;113(19):E2646–E2654.
50. Soto-Pantoja DR, et al. CD47 in the tumor microenvironment limits cooperation between antitumor T-cell immunity and radio-
therapy. Cancer Res. 2014;74(23):6771–6783.
51. Mattiola I, et al. Priming of human resting NK cells by autologous M1 macrophages via the engagement of IL-1β, IFN-β, and
IL-15 pathways. J Immunol. 2015;195(6):2818–2828.
52. Bodduluru LN, Kasala ER, Madhana RM, Sriram CS. Natural killer cells: the journey from puzzles in biology to treatment of
cancer. Cancer Lett. 2015;357(2):454–467.
53. Chiba S, et al. Recognition of tumor cells by Dectin-1 orchestrates innate immune cells for anti-tumor responses. Elife.
2014;3:e04177.
54. Asano K, et al. CD169-positive macrophages dominate antitumor immunity by crosspresenting dead cell-associated antigens.
Immunity. 2011;34(1):85–95.
55. Chao MP, Weissman IL, Majeti R. The CD47-SIRPα pathway in cancer immune evasion and potential therapeutic implica-
tions. Curr Opin Immunol. 2012;24(2):225–232.
56. Liu J, et al. Pre-clinical development of a humanized anti-CD47 antibody with anti-cancer therapeutic potential. PLoS One.
2015;10(9):e0137345.
57. Stenberg Å, et al. Signal regulatory protein ɑ is present in several neutrophil granule populations and is rapidly mobilized to the
cell surface to negatively fine-tune neutrophil accumulation in inflammation. J Innate Immun. 2014;6(4):553–560.
58. Motegi S, et al. Essential roles of SHPS-1 in induction of contact hypersensitivity of skin. Immunol Lett. 2008;121(1):52–60.
59. Motegi S, et al. Role of the CD47-SHPS-1 system in regulation of cell migration. EMBO J. 2003;22(11):2634–2644.
60. Weisser SB, van Rooijen N, Sly LM. Depletion and reconstitution of macrophages in mice. J Vis Exp. 2012;(66):4105.
Downloaded from http://insight.jci.org on January 24, 2017. https://doi.org/10.1172/jci.insight.89140