Abstract

   Mechanisms related to tumor evasion from NK cell-mediated immune
   surveillance remain enigmatic. Dickkopf-1 (DKK1) is a Wnt/β-catenin
   inhibitor, whose levels correlate with breast cancer progression. We
   find DKK1 to be expressed by tumor cells and cancer-associated
   fibroblasts (CAFs) in patient samples and orthotopic breast tumors, and
   in bone. By using genetic approaches, we find that bone-derived DKK1
   contributes to the systemic DKK1 elevation in tumor-bearing female
   mice, while CAFs contribute to DKK1 at primary tumor site. Systemic and
   bone-specific DKK1 targeting reduce tumor growth. Intriguingly,
   deletion of CAF-derived DKK1 also limits breast cancer progression,
   without affecting its levels in circulation, and regardless of DKK1
   expression in the tumor cells. While not directly supporting tumor
   proliferation, stromal-DKK1 suppresses NK cell activation and
   cytotoxicity by downregulating AKT/ERK/S6 phosphorylation. Importantly,
   increased DKK1 levels and reduced cytotoxic NK cells are detected in
   women with progressive breast cancer. Our findings indicate that DKK1
   represents a barrier to anti-tumor immunity through suppression of NK
   cells.

   Subject terms: Breast cancer, Cancer microenvironment, Tumour
   immunology
     __________________________________________________________________

   Elevated levels of Dickkopf-1 (DKK1), a soluble inhibitor of the
   Wnt/β-catenin pathway, are associated with poor prognosis in patients
   with breast cancer. Here the authors show that bone- and
   cancer-associated fibroblast-derived DKK1 contribute to breast cancer
   progression by suppressing NK cell cytotoxicity

Introduction

   Breast cancer is one of the most frequently diagnosed malignancies in
   women^[51]1. Despite early diagnosis, approximately 20-30% of patients
   experience metastatic relapse within 5-10 years of curative-intent
   therapy^[52]2. Hence, the need to find long-term and highly effective
   therapeutic approaches to prevent recurrence.

   Immune checkpoint blockade (ICB) has been FDA-approved to treat a
   variety of cancers previously considered incurable. Pembrolizumab
   (anti-PD-1) in combination with chemotherapy is now a standard-of-care
   for early and advanced triple-negative breast cancer
   (TNBC)^[53]3,[54]4. However, over 30% of TNBC patients do not benefit
   from ICB, and ICB is also not effective for ER^+ breast cancer^[55]5.
   Key factors responsible for the poor therapeutic responses to ICB in
   breast cancer include low tumor mutational burden^[56]6, recruitment of
   suppressive immune cells, and exclusion of lymphocytes at the tumor
   sites^[57]7. A recent analysis of circulating immune cell populations
   in breast cancer patients refractory to neoadjuvant chemotherapy
   revealed an apparent increase in the proportion of dysfunctional
   CD56^dim/CD16^− NK cells compared to patients achieving a pathological
   Complete Response (pCR)^[58]8. NK cells are innate immune populations
   offering a first line of defense against incipient tumors^[59]9. NK
   cell intratumoral and peritumoral abundance correlate with an elevated
   pCR rate in breast cancer patients undergoing neoadjuvant
   chemotherapy^[60]10. Furthermore, preclinical evidence suggests that
   activated NK cells could potentiate the response to ICB in MHC class I
   low tumors^[61]11. Thus, understanding the mechanisms that drive
   systemic immunosuppression, including NK cell inactivation, could
   improve response to ICB in breast cancer patients.

   Dickkopf-1 (DKK1) is a soluble inhibitor of the Wnt/β-catenin pathway,
   primarily recognized for its role in bone homeostasis and
   cancer-induced osteolytic bone disease^[62]12,[63]13. Elevated levels
   of DKK1 in circulation and/or in tumor tissues correlate with poor
   prognosis in numerous cancer types^[64]14–[65]19. Direct effects of
   DKK1 on tumor cells have been described in head and neck cancer,
   pancreatic ductal adenocarcinoma, and esophageal squamous cell
   carcinoma^[66]20–[67]22. Our lab and others have also demonstrated that
   DKK1 drives the accumulation of suppressive myeloid populations in
   melanoma, lung carcinoma, prostate, and gastric
   cancer^[68]18,[69]23–[70]27, ultimately reducing T cell and NK cell
   responses. Thus, DKK1 targeting has been successfully used in
   combination with ICB in mouse models of melanoma and gastric
   cancer^[71]24,[72]26,[73]28. In breast cancer, high levels of DKK1 have
   been associated with poor prognosis and dissemination to bone^[74]15.
   The pro-tumorigenic effects of DKK1 in breast cancer have been mainly
   attributed to its ability to increase bone resorption, thus creating a
   favorable environment for tumor dissemination to bone^[75]15,[76]29.
   However, whether DKK1 exerts systemic immune suppressive effects and/or
   generates a local immune suppressive environment at the primary tumor
   site has never been reported.

   In this work, we demonstrate the direct inhibitory effects of DKK1 on
   NK cell activation and cytotoxicity during breast cancer progression.
   We show that cancer-associated fibroblasts (CAFs) contribute to the
   production of DKK1 in the tumor microenvironment (TME), and bone cells
   contribute to systemic DKK1 elevation. Both bone- and CAF-derived DKK1
   suppress NK cell cytotoxicity, and their targeting reduces tumor
   progression. Finally, increased DKK1 levels and reduced cytotoxic NK
   cells are also detected in breast cancer patients with progressive bone
   disease. Our work positions DKK1 as a negative modulator of anti-tumor
   immunity via suppression of NK cell cytotoxicity and raises the
   importance of DKK1 targeting to improve NK cell-directed therapies.

Results

DKK1 augments breast cancer progression

   To determine the role of DKK1 in breast cancer progression, we first
   measured DKK1 serum levels in C57BL/6 mice orthotopically injected with
   a luminal B, ER^+, hormone-resistant PyMT-BO1 cell line into the
   mammary fat pad (MFP). DKK1 levels were significantly increased in
   tumor-bearing mice (Fig. [77]1A), recapitulating findings in breast
   cancer patients^[78]17. Next, we administered the DKK1-neutralizing
   monoclonal antibody mDKN01, previously validated for its anti-tumor
   effects in melanoma, gastric, and gynecologic mouse tumor
   models^[79]24,[80]26,[81]30–[82]32 and currently under investigation in
   clinical trials for gastric and endometrial cancers ([83]NCT04363801,
   [84]NCT05761951, [85]NCT04681248). Administration of mDKN01 (10 mg/kg)
   every other day following tumor inoculation, led to a significant
   reduction in primary tumor growth compared to IgG (Fig. [86]1B). To
   further investigate the role of DKK1 during tumor dissemination, we
   injected the firefly luciferase-conjugated PyMT-BO1 cell line into
   albino C57BL/6 mice either intracardiacally (i.c.) to study
   dissemination to various organs including the bone, or directly into
   the tibias (i.t.) to study tumor growth in bone, followed by treatment
   with mDKN01. Bioluminescence imaging (BLI) showed significantly reduced
   tumor growth at all sites (Fig. [87]1C–F). To assess the therapeutic
   potential of mDKN01 on established tumors, we initiated treatment 7
   days post-tumor inoculation (Fig. [88]1G) and observed a significant
   reduction in tumor burden compared to IgG controls (Fig. [89]1H).
   Similarly, DKK1 neutralization significantly reduced the growth of the
   luminal B, ER^+/PR^+, hormone-sensitive E0771 breast cancer cell line
   (Supplementary Fig. [90]1A).

Fig. 1. DKK1 augments breast cancer progression.

   [91]Fig. 1
   [92]Open in a new tab

   A DKK1 serum levels were measured by ELISA in no tumor-bearing (NTB)
   6–8 weeks old C57BL/6 WT female mice (n = 5) or 2 weeks after the
   inoculation of PyMT-BO1 breast cancer cells into the mammary fat pad
   (MFP; 10^5 cells, n = 7). B Tumor growth in the MFP was determined by
   caliper measurements in WT mice inoculated with PyMT-BO1 (n = 5
   mice/group) receiving mDKN01 (10 mg/kg) or control IgG antibody i.p.
   every other day. C–F Tumor progression was determined by BLI in mice
   inoculated intracardiacally with PyMT-BO1 cells (i.c.; 10^4 cells,
   albino C57BL/6, n = 9 for IgG, n = 6 for mDKN01) (C, D) or
   intratibially (i.t.; 10^4 cells, C57BL/6, n = 6 for IgG, n = 7 for
   mDKN01) (E, F) followed by administration of mDKN01 or control IgG
   antibody every other day. G Schematic representation of the therapeutic
   administration of IgG and mDKN01. H Tumor growth in the MFP was
   determined by caliper measurements in WT mice inoculated with PyMT-BO1
   (n = 5 mice/group) receiving mDKN01 (10 mg/kg) or control IgG antibody
   i.p. every three days starting 7 days post-tumor inoculation. I DKK1
   serum levels were measured by ELISA in 6–8 weeks old female BALB/c WT
   mice with no tumors (NTB n = 7) or 2 weeks after the inoculation of 4T1
   breast cancer cells into the MFP (n = 6). J Primary tumor growth was
   evaluated by caliper measurements in WT mice inoculated with 4T1 cells
   (n = 4 mice/group) into the MFP receiving mDKN01 (10 mg/kg) or control
   IgG antibody i.p. every other day. K, L Tumor progression was
   determined by BLI in mice inoculated intratibially with 4T1 cells
   (i.t.; 10^4 cells, BALB/c, n = 5/group) followed by administration of
   mDKN01 or control IgG antibody every other day. Results are shown as
   mean ± SEM. An unpaired t-test with a two-tailed P-value (A, C, E, I,
   K), and two-way ANOVA followed by Bonferroni multiple-comparison test
   (B, H, J) were used to determine significance. *P < 0.05, **P < 0.01,
   ***P < 0.001, ****P < 0.0001. Source data are provided as a Source Data
   file.

   To assess whether DKK1 levels were also increased in TNBC, we used
   BALB/c mice orthotopically injected with the 4T1 tumor line and found
   increased DKK1 in circulation (Fig. [93]1I). Administration of mDKN01
   (10 mg/kg) led to a significant reduction in 4T1 tumor growth at the
   primary site and in the bone compared to IgG control (Fig. [94]1J–L).
   These results demonstrate the involvement of DKK1 in supporting tumor
   progression and the therapeutic benefit of DKK1 targeting in various
   breast cancer subtypes.

DKK1 is expressed in the tumor microenvironment of breast cancer patient
tissues

   To assess DKK1 expression at the tumor site, we analyzed a microarray
   dataset ([95]GSE3744^[96]33) composed of healthy and malignant breast
   tissues from ER^+, HER2^+, and TNBC patients. DKK1 expression was
   increased in triple-negative (TNBC, Fig. [97]2A) and HER2^+ breast
   cancer (Fig. [98]2C) compared to healthy tissues, whereas no
   significant differences were observed in ER^+ breast cancer
   (Fig. [99]2E). To further identify which cell populations express DKK1,
   we analyzed a single-cell RNAseq dataset from 26 primary human breast
   tumors, including 11 ER^+, 5 HER2^+, and 10 TNBC
   ([100]GSE176078^[101]34). The cells were annotated and clustered by
   using canonical and signature-based markers^[102]35 (Supplementary
   Fig. [103]2A, B). Confirming the microarray analysis, DKK1 was
   expressed in the cancer epithelial cells in HER2^+ and TNBC, while it
   was detected at much lower levels in ER^+ and in the normal epithelial
   cells (Supplementary Fig. [104]2C). No DKK1 expression was detected in
   the immune cell cluster.

Fig. 2. DKK1 is expressed in the tumor microenvironment.

   [105]Fig. 2
   [106]Open in a new tab

   A, C, E Normalized gene expression of DKK1 in healthy human breast
   tissues (n = 7) and triple-negative breast cancer (TNBC) (A, n = 18),
   HER2^+ breast cancer (C, n = 4), ER^+ breast cancer (E, n = 11)
   ([107]GSE3744). B, D, F Representative images of multiplex
   immunohistochemistry (IHC) of TNBC (B, n = 20), HER2^+ (D, n = 12), and
   ER^+ (F, n = 8) human breast cancer subtypes stained for DKK1 (red),
   αSMA (green), PDGFRα (blue), and panCK (cyan). The green inset
   highlights the stromal area and the cyan inset highlights the tumor
   area. G Normalized gene expression of DKK1 in the stroma derived from
   normal breast tissue or invasive ductal carcinoma (IDC) ([108]GSE8977).
   Results are shown as mean ± SEM. H Representative images of multiplex
   IHC of human ductal carcinoma in situ (DCIS) samples (n = 13) stained
   for DKK1 (red), αSMA (green), PDGFRα (blue), COL14a1 (white), and panCK
   (cyan) from patients who did not develop ipsilateral breast cancer
   (blue box) versus patients who developed ipsilateral breast cancer (red
   box). I–K Representative images of DKK1 staining by IHC (brown) in
   orthotopic PyMT tumors (I, n = 5), spontaneous MMTV-PyMT breast tumors
   (J, n = 5), and orthotopic 4T1 tumors (K, n = 4). L, M Representative
   images of multiplex IHC of spontaneous MMTV-PyMT breast tumors (L,
   n = 5), and orthotopic 4T1 tumors (M, n = 4) stained for DKK1 (red),
   αSMA (green), COL1a1 (white), and EpCAM (MMTV-PyMT tumor cells, cyan)
   or GFP (4T1 tumor cells, cyan). An unpaired t-test with a two-tailed
   P-value (A, C, E, G) was used to determine significance. Source data
   are provided as a Source Data file.

   Since DKK1 levels are increased in circulation in ER^+
   patients^[109]36, we sought to examine whether DKK1 was expressed by
   additional cell types in the tumor microenvironment (TME).
   Interestingly, we detected DKK1 expression, albeit limited, in the
   stromal compartment within CAF clusters (Supplementary Fig. [110]2D).
   Further subtype analysis based on established breast cancer CAF
   classification^[111]37–[112]39, showed DKK1 expression in
   ACTA2^+COL1a1^highPDGFRα^+ myofibroblasts (myCAF, Supplementary
   Fig. [113]2E–G).

   To confirm these findings, we performed automated multiplex
   immunohistochemistry (IHC) using TNBC, HER2^+, and ER^+ human breast
   cancer tissue microarrays (TMAs). For each antibody used, images were
   given an arbitrary color with the deconvolution algorithm provided by
   HALO software under the supervision of a trained pathologist, which
   allowed us to determine the populations expressing DKK1. The
   specificity and titration of the anti-DKK1 antibody were determined
   using the placenta as a positive control and benign breast epithelium
   as a negative control (Supplementary Fig. [114]2H). Placenta and normal
   breast tissues were included in the TMA to set up the threshold signal
   in HALO and eliminate any background noise (Supplementary Fig. [115]2I,
   J). As expected, DKK1 staining was observed in TNBC and HER2^+ cancer
   epithelial cells (Fig. [116]2B, D), identified by the expression of the
   pan-cytokeratin (PanCK) marker, but very limited expression was found
   in ER^+ breast cancer (Fig. [117]2F). Interestingly, the epithelial
   cancer cells displayed a strong DKK1 nuclear localization, as recently
   reported^[118]40, rather than a cytoplasmic staining where secreted
   proteins typically reside. Although DKK1 is expected to be secreted, we
   were unable to detect a robust extracellular signal, possibly due to
   the technical limitations of using multiplex IHC. Diffused
   intracellular DKK1 staining was detected in subsets of PDGFRα^+ and
   αSMA^+ cells in TNBC and HER2^+ subtypes (Fig. [119]2B, D), but it was
   barely measurable in the stroma of ER^+ tumors (Fig. [120]2F), in
   agreement with a previous report^[121]41. We also observed that DKK1
   was increased in a microarray dataset from invasive ductal carcinoma
   (IDC) compared to normal breast tissue (Fig. [122]2G;
   [123]GSE8977^[124]42). Next, we performed multiplex IHC on breast
   biopsies from 13 patients diagnosed with ductal carcinoma in situ
   (DCIS, stage 0 breast cancer) with clinical annotations, of which 3
   patients had no recurrence and 10 patients developed ipsilateral breast
   cancer after partial mastectomy. Interestingly, we observed DKK1
   staining only in cells expressing the fibroblast markers PDGFRα and
   COL14a1, but not in the epithelial cells, in 9 out of 10 in DCIS
   tissues from patients who recurred (Fig. [125]2H Recurrence). DKK1
   protein expression was not detected in any of the DCIS patients who did
   not recur (Fig. [126]2H No Recurrence).

DKK1 is expressed in cancer-associated fibroblasts and bone in murine breast
cancer models

   Next, we investigated Dkk1 expression in the PyMT, 4T1, and E0771
   murine tumors. While Dkk1 transcripts were only detected in the 4T1
   tumor cell line in vitro (Supplementary Fig. [127]1B), Dkk1 was found
   in the tumor mass of PyMT and E0771 models (Supplementary Fig. [128]1C,
   D). IHC showed DKK1 staining in stromal cells with an elongated,
   fibroblast-like morphology in the PyMT (Fig. [129]2I), spontaneous
   MMTV-PyMT (Fig. [130]2J), and 4T1 orthotopic tumors (Fig. [131]2K).
   Further, co-staining with CAF and epithelial markers confirmed DKK1 in
   elongated αSMA^+ cells and partial colocalization with COL1a1
   (Fig. [132]2L, M, Supplementary Fig. [133]2K). As expected, DKK1 was
   also detected in the 4T1 tumor cells, but not in the PyMT cancer cells.

   Because DKK1 levels in circulation are increased in ER^+ patients with
   bone metastases^[134]36 and Dkk1 is highly expressed in bone in
   homeostatic conditions^[135]43, we evaluated the expression of
   bone-derived Dkk1 in mice bearing orthotopic breast tumors and found
   increased expression compared to no tumor controls (Supplementary
   Fig. [136]1E, F). These results indicate local production of Dkk1 at
   the tumor site by either tumor cells and/or CAFs, and distal production
   of Dkk1 by bone cells.

Bone and CAF-derived DKK1 contribute to systemic and local increases in DKK1
levels during tumor progression

   To determine the role of bone versus CAF-derived DKK1 during tumor
   progression, we generated mouse models with targeted deletion of Dkk1
   in osteoblasts and fibroblasts. To specifically delete Dkk1 from
   osteoblasts, we crossed Dkk1^fl/fl mice with the
   doxycycline-repressible Sp7 Cre line (herein referred to as
   Sp7-Dkk1cKO)^[137]44. Because Dkk1 deletion leads to embryonic
   lethality, moms and pups were fed a doxycycline-containing diet until
   weaning to suppress the transgene activation, and mice were
   orthotopically injected with the PyMT tumor cells when they reached 6–8
   weeks of age. Strikingly, Sp7-Dkk1cKO mice showed a significant
   reduction in primary tumor growth (Fig. [138]3A). While Dkk1 expression
   at the tumor site was not reduced compared to littermate controls, DKK1
   levels in circulation were drastically reduced, indicating that
   bone-derived DKK1 contributes to systemic elevation of DKK1 during
   tumor progression (Fig. [139]3B-D).

Fig. 3. Bone and CAF-derived DKK1 contribute to systemic and local increases
in DKK1 levels during tumor progression.

   [140]Fig. 3
   [141]Open in a new tab

   A Tumor growth was determined by caliper measurements in 6–8 weeks old
   Sp7-Dkk1WT (control) and Sp7-Dkk1cKO female mice (n = 4 mice/group)
   inoculated with PyMT in the MFP. B, C qRT-PCR for Dkk1 expression in
   bone and primary tumor (TM). D DKK1 serum levels measured by ELISA. E
   Tumor growth was determined by caliper measurements in 10–12 weeks old
   control and αSMA-Dkk1cKO (n = 5 mice/group). F, G qRT-PCR for Dkk1
   expression in bone and primary tumors (TM). H DKK1 serum levels
   measured by ELISA. I Multiplex immunohistochemistry (IHC) of orthotopic
   PyMT tumors in αSMA-Dkk1WT (top, n = 6) or αSMA-Dkk1cKO mice (bottom,
   n = 6) stained for DKK1 (red), αSMA (green), tdTomato (white) and
   hematoxylin (blue). J qRT-PCR for Dkk1 expression in sorted tdT^+ CAFs
   from primary tumors in αSMA-Dkk1WT-tdT and αSMA-Dkk1cKO-tdT mice
   (n = 3/group). K Tumor growth was determined by caliper measurements in
   WT mice co-injected with 10^5 PyMT cells and 10^5 tdT^+ CAFs isolated
   from αSMA-Dkk1WT-tdT or αSMA-Dkk1cKO-tdT (n = 8/group). Mice injected
   with tumor cells alone (n = 4) were used as control. L Representative
   immunofluorescence images of MDA-MB-231 cells stained for DKK1 (green)
   and DAPI (blue) (n = 3). M Tumor growth was determined by caliper
   measurements in Nude mice co-injected with 10^5 MDA-MB-231 cells and
   10^5 tdT^+ CAFs isolated from αSMA-Dkk1WT-tdT or αSMA-Dkk1cKO-tdT
   (n = 6/group). Mice injected with tumor cells alone (n = 6) were used
   as control. Results are shown as mean ± SEM. An unpaired t-test with a
   two-tailed P-value (B–D, F–H, J), two-way ANOVA followed by Bonferroni
   multiple-comparison test (A, E, K, M) were used to determine
   significance. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
   Source data are provided as a Source Data file.

   Based on the expression of DKK1 in αSMA^+ cells in the TME, next, we
   crossed Dkk1^fl/fl mice with the inducible αSMACreER^T2 line (referred
   to as αSMA-Dkk1cKO). Cre activation was induced by 5 consecutive
   intraperitoneal (i.p.) injections of tamoxifen (100 mg/kg per dose) to
   10–12 weeks old αSMA-Dkk1cKO and αSMA-Dkk1WT mice. As a control, we
   crossed the αSMACreER^T2 mice with the Rosa26-LSL-tdTomato line
   (referred to as αSMA-tdT mice) and injected PyMT cells into their MFP
   one day after the first tamoxifen injection to confirm the presence of
   αSMA-tdT^+ cells exclusively at the tumor site, but not in the bone
   (Supplementary Fig. [142]3A). Intriguingly, αSMA-Dkk1cKO mice also
   showed a significant reduction in primary tumor growth compared to
   their littermate controls (Fig. [143]3E), despite displaying no changes
   in DKK1 levels in the bone and circulation but achieving efficient
   deletion in the tumor mass (Fig. [144]3F–H). Multiplex IHC further
   confirmed expression of DKK1 or lack thereof in αSMA^+ cells
   (Fig. [145]3I). Confirming the estrogen insensitivity of the PyMT line,
   tamoxifen did not directly affect tumor cell proliferation in vitro
   (Supplementary Fig. [146]3B) and PyMT subcutaneous (SQ) tumor growth
   was reduced in tamoxifen-treated αSMA-Dkk1cKO male mice compared to
   controls (Supplementary Fig. [147]3C).

   Because αSMA is also expressed in myoepithelial and endothelial cells,
   to further determine the relevance of CAF-derived DKK1 during tumor
   progression, as a complementary approach we crossed Dkk1^fl/fl mice
   with the FSP1 Cre line (referred to as FSP1-Dkk1cKO) to allow deletion
   of Dkk1 in fibroblasts. Similarly to the αSMA-Dkk1cKO mice, we observed
   a significant reduction in primary tumor growth in this model
   (Supplementary Fig. [148]3D). Efficient deletion of Dkk1 in the TME was
   confirmed via qRT-PCR (Supplementary Fig. [149]3E).

   Next, to further understand the local effects of DKK1 in the TME, we
   isolated tdT^+ CAFs from primary tumors in αSMA-Dkk1cKO-tdT and
   αSMA-Dkk1WT-tdT mice and co-injected them with PyMT tumor cells (1:1
   ratio) into the MFP of naïve WT recipient mice. Expression or deletion
   of Dkk1 in the sorted tdT^+ CAFs was confirmed by qRT-PCR
   (Fig. [150]3J). Mice injected with tumor cells alone were used as
   controls. Highlighting the importance of local production of DKK1, mice
   co-injected with Dkk1-deficient CAFs showed smaller tumor size compared
   to mice co-injected with Dkk1-sufficient CAFs (Fig. [151]3K).

   Because DKK1 is also expressed by cancer epithelial cells in TNBC
   patient samples, next we assessed the relative importance of
   CAF-derived DKK1 versus tumor cell-derived DKK1. We used the MDA-MB-231
   breast cancer cell line, which expresses endogenous DKK1^[152]45.
   Similarly to the TNBC TMAs, we observed a strong DKK1 staining in the
   nucleus of MDA-MB-231 cells (Fig. [153]3L). Next, we co-injected
   MDA-MB-231 cells with either Dkk1-sufficient CAFs (isolated from tumors
   in αSMA-Dkk1WT-tdT mice) or Dkk1-deficient CAFs (isolated from tumors
   in αSMA-Dkk1cKO-tdT mice) at 1:1 ratio into the MFP of Nude recipient
   mice. Strikingly, we found a significant decrease in tumor growth in
   mice co-injected with Dkk1 null CAFs compared to those co-injected with
   Dkk1-sufficient CAFs (Fig. [154]3M). These findings highlight the
   importance of local DKK1 production by CAFs regardless of DKK1
   expression in the tumor cells.

DKK1 has immunomodulatory effects

   To assess whether DKK1 exerts direct effects on tumor cell
   proliferation, we cultured the PyMT, 4T1, and E0771 tumor cells in the
   presence of recombinant DKK1 (rDKK1) and performed an MTT assay. For
   all three cell lines, rDKK1 did not increase cell density compared to
   unstimulated cells at all time points and doses tested (Supplementary
   Fig. [155]4A–C), nor it induced any significant changes in cell cycle
   and survival (Supplementary Fig. [156]4D, E).

   To better understand how DKK1 promotes tumor progression in vivo, we
   performed bulk RNA sequencing using GFP-H2B-mApple-Thy1.1^+ PyMT-BO1
   cells isolated from orthotopic tumors in WT mice receiving IgG or
   mDKN01, after exclusion of Ter119^+ erythrocytes and CD45^+ immune
   cells (Fig. [157]4A, Supplementary Fig. [158]4F). Only 134 genes were
   differentially expressed (DEGs, p < 0.05 and |fold change | >2) between
   the two groups (Fig. [159]4B). KEGG pathway enrichment analysis
   confirmed no differences in pathways related to cell viability or cell
   cycle but rather showed changes in pathways related to immune responses
   (Fig. [160]4C). Gene set enrichment analysis (GSEA) further showed
   hallmarks of anti-tumor immune responses being upregulated in the
   mDKN01-treated tumors compared to IgG, including interferon-gamma
   response, interferon-alpha response, TNFα signaling via NFκB, and
   IL-2/STAT5 signaling (Fig. [161]4D). These results suggest that
   expression of DKK1 at tumor site might contribute to create an immune
   suppressive environment, rather than directly affecting tumor growth.

Fig. 4. DKK1 has immunomodulatory effects.

   [162]Fig. 4
   [163]Open in a new tab

   A Schematic representation of PyMT tumor cell isolation from IgG or
   mDKN01-treated mice (n = 3/group) and analysis of transcriptome via
   bulk RNA sequencing. B Venn diagram depicting uniquely and commonly
   expressed genes in PyMT cells isolated from orthotopic tumors injected
   into WT mice treated with IgG or mDKN01. C KEGG pathway enrichment
   analysis on differentially expressed genes (DEGs, p < 0.05, |fold
   change | >2). D GSEA analysis of hallmarks upregulated in PyMT tumor
   cells isolated from mDKN01-treated mice compared to IgG-treated mice.
   Normalized enrichment score (NES) and nominal P-value were calculated
   as previously described^[164]59.

Local production of DKK1 at tumor site affects tumor immune infiltration

   To determine if DKK1 modulates the immune landscape of the TME, we
   profiled PyMT tumor-infiltrating immune populations from IgG or
   mDKN01-treated mice. We found a significant increase in the number of
   CD45^+ cells per gram of tumor following mDKN01 administration, with
   CD4^+ and CD8^+ T cells, F4/80^+ macrophages, and NK cells being the
   most increased subsets (Fig. [165]5A, Supplementary Fig. [166]5A, B).
   Similar results were found in orthotopic breast tumors isolated from
   αSMA-Dkk1cKO mice compared to αSMA-Dkk1WT mice (Supplementary
   Fig. [167]5C)

Fig. 5. DKK1 supports tumor progression by targeting NK cells.

   [168]Fig. 5
   [169]Open in a new tab

   A Tumor-infiltrating CD45^+ immune cells, NK cells, T cells, and
   myeloid subsets per gram of PyMT tumor mass from WT mice treated with
   IgG (n = 4) or mDKN01 (n = 5). B–D Deconvoluted IHC images from
   orthotopic PyMT tumors stained for CD45 (red) and hematoxylin (gray)
   isolated from WT mice treated with IgG or mDKN01 (B, n = 5/group),
   αSMA-Dkk1WT and αSMA-Dkk1cKO mice (C, n = 5/group), and mice
   co-injected with tumor cells and tdT^+ CAFs from tumors in
   αSMA-Dkk1WT-tdT or αSMA-Dkk1cKO-tdT mice (D, n = 8/group). E Tumor
   growth by caliper measurements in 6–8 weeks old NSG immune-compromised
   mice (n = 6 mice/group) inoculated with PyMT into the MFP. F, G PyMT
   orthotopic growth determined by caliper measurements in 6–8 weeks WT
   mice treated with IgG (n = 8) or mDKN01 (10 mg/kg, n = 8) every other
   day along with anti-CD4 and anti-CD8 (F, n = 4 and n = 9) or anti-NK1.1
   (G, n = 4 and n = 8). H Tumor growth was determined by caliper
   measurements in Prf1^−^/^− mice inoculated with PyMT into the MFP and
   treated i.p. with mDKN01 (10 mg/kg) or control IgG antibody every other
   day (n = 4 mice/group). Results are shown as mean ± SEM. An unpaired
   t-test with a two-tailed P-value (A), two-way ANOVA followed by
   Bonferroni multiple-comparison test (E, H), ordinary one-way ANOVA
   followed by Dunnett’s multiple-comparison test (F, G) were used to
   determine significance. *P < 0.05, **P < 0.01, ***P < 0.001,
   ****P < 0.0001. Source data are provided as a Source Data file.

   IHC further indicated the presence of CD45^+ immune populations in the
   central regions of the tumor mass in PyMT tumors treated with mDKN01
   (Fig. [170]5B), αSMA-Dkk1cKO animals (Fig. [171]5C) and from WT mice
   co-injected with tumor cells plus αSMA-Dkk1cKO CAFs (Fig. [172]5D). In
   contrast, peripheral CD45^+ cell localization was observed in all
   control groups. These results suggest that CAF-derived DKK1 can limit
   the infiltration of immune cells at the tumor site, regardless of the
   elevated levels of DKK1 in circulation.

DKK1 targets NK cells to support tumor progression

   To identify the immune populations targeted by DKK1, we injected PyMT
   cells into the MFP of female NSG mice, which lack T, B, and NK cells,
   and administered IgG or mDKN01. Strikingly, the anti-tumor effects of
   mDKN01 were fully abrogated in this mouse model (Fig. [173]5E). Next,
   we adopted a selective immune cell depletion approach. Mice were either
   depleted of T or NK cells, or treated with vehicle control and
   randomized to receive IgG or mDKN01. As expected, the depletion of T
   cells only slightly increased tumor burden compared to IgG controls
   (Fig. [174]5F, Supplementary Fig. [175]5D), confirming the limited
   involvement of T cells in the PyMT tumor model^[176]46. Furthermore,
   mDKN01 anti-tumor effects were not affected by T cell depletion. In
   contrast, depletion of NK cells significantly increased tumor burden
   compared to control mice and completely abrogated the anti-tumor
   effects of mDKN01 (Fig. [177]5G, Supplementary Fig. [178]5E). Similar
   findings were observed in mice lacking perforin (Prf1^−^/^−), a key
   effector molecule utilized by T and NK cells to form pores enabling the
   transfer of cytotoxic granzymes to induce the specific killing of
   target cells (Fig. [179]5H). Since T cell depletion did not alter the
   anti-tumor effects of mDKN01, findings in Prf1^−^/^− mice suggest that
   DKK1 supports tumor progression through suppression of NK cell
   function.

DKK1 directly suppresses NK cell cytotoxicity

   NK cells must reside near their target cells to exert cytotoxic
   effects. We observed that NK cells exposed to rDKK1 (200 ng/ml) for 3 h
   were rarely found in close contact with mCherry^+ PyMT tumor cells, as
   shown by high magnification confocal microscopy of F-actin-stained
   cells (Fig. [180]6A rDKK1). In contrast, NK cells not exposed to rDKK1
   were often in direct association with the tumor cells (Fig. [181]6A
   control). Based on these observations, we aimed to determine whether
   DKK1 directly affects NK cell cytotoxicity, and quantified NK
   cell-mediated killing of the cell trace violet (CTV)-labeled PyMT tumor
   cells by assessing the expression of 7-AAD via flow cytometry
   (Fig. [182]6B, Supplementary Fig. [183]6A). rDKK1 significantly
   suppressed NK cell-mediated killing at all tested effector (NK cells)
   to target (PyMT) ratios (Fig. [184]6C). Similar results were observed
   using a 48 h  IncuCyte Live Cell assay with a 2:1 effector (NK cells)
   to target (H2B-mApple-Thy1.1^+ PyMT-BO1 cells) ratio (Supplementary
   Fig. [185]6B).

Fig. 6. DKK1 suppresses NK cell cytotoxicity.

   [186]Fig. 6
   [187]Open in a new tab

   A Representative immunofluorescence images of mCherry^+ PyMT cells
   (red) cultured with murine NK cells in the presence of PBS (n = 11) or
   rDKK1 (n = 9) for 3 h prior to fixation and staining for F-actin
   (green). Small, mCherry^− cells visualized by arrows or by image
   contrast in insets represent NK cells. Images in the insets are
   enlarged 1.5 times. B Schematic representation of NK cell isolation
   from the spleen of Poly I:C treated WT mice (n = 4) and incubation with
   PyMT target cells in the presence of PBS or rDKK1. C Analysis of
   percent specific killing measured by 7-AAD^+ PyMT target cells after
   4 h incubation with NK cells. (n of wells per condition = 3). D–G
   Schematic representation of CAF isolation from orthotopic PyMT tumors
   in αSMA-DKK1WT-tdT mice (D, n = 5), osteoblast precursor (pre-OB)
   isolation from the bone marrow of tumor-bearing WT mice (F, n = 3) and
   isolation of NK cells from the spleen of Poly I:C injected mice
   (n = 4). NK cells and increasing numbers of CAFs/pre-OBs were incubated
   with PyMT target cells in the presence of mDKN01 or IgG (50 mg/ml) and
   percent specific killing measured by 7-AAD^+ PyMT target cells after
   4 h incubation with NK cells (2:1 NK cell to PyMT ratio) (E, G n of
   wells per condition = 3). Results are shown as mean ± SEM. Experiments
   in (C, E, G) were performed in triplicates. Two-way ANOVA followed by
   Bonferroni multiple-comparison test was used to determine significance
   (C, E, G). **P < 0.01, ***P < 0.001, ****P < 0.0001. Source data are
   provided as a Source Data file.

   To investigate whether CAF and bone-derived DKK1 exert similar NK
   suppressive effects, we co-cultured different numbers of CAFs or
   osteoblast precursors (pre-OBs), as a source of bone-derived DKK1,
   together with NK cells and CTV-labeled PyMT tumor cells (2:1 effector
   to target ratio) and assessed their ability to kill the tumor cells
   with or without mDKN01. TdT^+ CAFs, isolated from orthotopic tumors in
   αSMA-Dkk1WT-tdT mice (Fig. [188]6D), reduced the NK-mediated tumor
   killing, and DKK1 neutralization restored NK cell functionality
   (Fig. [189]6E). Similarly, pre-OBs isolated from the bone marrow of
   tumor-bearing mice reduced the NK-mediated killing of PyMT cells
   (Fig. [190]6F, G). These inhibitory effects were restored by mDKN01.
   Collectively, these results demonstrate that both bone-derived and
   CAF-derived DKK1 significantly suppress NK cell tumoricidal activities.

DKK1 downregulates PI3K/AKT/mTOR and MAPK/ERK pathways in NK cells

   To determine how DKK1 impacts NK cells, we performed bulk RNA
   sequencing of NK cells sorted from the spleen of poly I:C treated mice
   using the CD45^+CD3^−NK1.1^+ markers and stimulated ex vivo with rDKK1
   for 4 h. Out of 14,197 genes detected, 327 genes were differentially
   expressed in the rDKK1-stimulated versus unstimulated NK cells
   (Fig. [191]7A). KEGG pathway analysis showed enrichment in signaling
   pathways related to NK cell development (Notch pathway) and function
   (HIF-1, Rap1, mTOR, and pathways involved in actin cytoskeleton
   regulation) (Fig. [192]7B). Furthermore, GSEA analysis showed
   reductions in genes related to anti-tumor immune responses including
   the IL-2/STAT5, the interferon-gamma and the PI3K/AKT/mTOR signaling
   (Fig. [193]7C). Accordingly, the rDKK1-exposed NK cells showed
   decreased gene expression of cytokine receptors Ifngr1, Ifnar1, Il2r,
   Il12rb1, Il15ra and Il18r1, adhesion molecules involved in the
   maintenance of the immunological synapse Itgal, Cd244a, and Cd226 and
   activating receptors Ncr1, Klrk1, Klrb1c, and Cd226 (Fig. [194]7D). We
   also observed that the expression of Prf1 was decreased in the NK cells
   exposed to rDKK1.

Fig. 7. DKK1 downregulates PI3K/AKT/mTOR and MAPK/ERK signaling pathways in
NK cells.

   [195]Fig. 7
   [196]Open in a new tab

   A Venn diagram indicating numbers of uniquely and commonly expressed
   genes in NK cells isolated spleen of Poly I:C treated WT mice (n = 3)
   and stimulated with rDKK1 (200 ng/ml) or PBS as a control
   (n = 3/group). B KEGG pathway enrichment analysis showing
   differentially expressed genes (DEGs, p < 0.05, |fold change | >2). C
   GSEA analysis showing hallmarks downregulated in rDKK1 stimulated NK
   cells. Normalized enrichment score (NES) and nominal P-value were
   calculated as previously described^[197]59. D Heatmap of genes related
   to NK cell cytotoxicity in PBS or rDKK1 stimulated NK cells. E Fold
   changes from baseline of mean fluorescence intensity (MFI) measurements
   of phosphorylated AKT, S6, ERK1/2, and STAT5 in NK1.1^+ NK cells from
   the spleens of Rag1^−^/^− mice (n = 4) following stimulation with rDKK1
   (200 ng/ml) for indicated times. F, G Fold changes in the percentage of
   NK cells expressing CD107a (F) or IFNγ (G) following ex vivo
   stimulation with anti-NK1.1 antibody (F) or a cytokine cocktail of
   IL-12 + IL-15 compared to unstimulated cells were evaluated in the PyMT
   tumor mass and bone marrow of αSMA-Dkk1WT (n = 4) or αSMA-Dkk1cKO mice
   (n = 6). Results are shown as mean ± SEM. Ordinary one-way ANOVA
   followed by Dunnett’s multiple-comparison test (E), and an unpaired
   t-test with a two-tailed P-value (F, G) were used to determine
   significance. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.
   Source data are provided as a Source Data file.

   Given the importance of PI3K^[198]47, MAPK^[199]48, and
   JAK/STAT^[200]49 in regulating NK cell activation, we investigated
   whether these signaling pathways were directly modulated by DKK1. We
   analyzed the phosphorylation of AKT (downstream of PI3K), ribosomal S6
   protein (downstream of mTOR), ERK1/2 (downstream of MAPK), and STAT5 in
   NK cells from the spleens of Rag1^−^/^− mice, which lack T and B cells,
   thereby making the NK cell gating and intracellular FACS analysis more
   efficient. Splenocytes were stimulated ex vivo with rDKK1 and NK cells
   were gated based on NK1.1 expression. We observed reduced
   phosphorylation of AKT, ribosomal S6 protein, and ERK1/2 in response to
   rDKK1, but no changes in STAT5 phosphorylation (Fig. [201]7E,
   Supplementary Fig. [202]6C). Interestingly, we did not observe any
   activation of AKT/ERK/S6 pathways when NK cells were stimulated with
   the canonical Wnt ligand rWnt3a (Supplementary Fig. [203]6D). However,
   when NK cells were stimulated with rWnt3a plus rDKK1 we observed a
   significant reduction in AKT/ERK/S6 phosphorylation, similar to rDKK1
   stimulation alone (Supplementary Fig. [204]6E). These results suggest
   that rDKK1 modulates NK cell activation independent of canonical Wnt
   signaling.

   To validate these findings in vivo and assess the local effects of DKK1
   on NK cell activation, we examined NK cell activation in orthotopic
   tumors or the bone marrow from αSMA-Dkk1WT and αSMA-Dkk1cKO mice. The
   single-cell suspensions of the tumor mass or bone marrow were
   stimulated for 4 h with the NK1.1 antibody or a cytokine cocktail
   containing IL-12 and IL-15, followed by analysis of NK cell
   degranulation via CD107a surface expression and NK cell activation by
   IFNγ staining. We found a significant increase in CD107a^+ and IFNγ^+
   NK cells in the tumor mass of αSMA-Dkk1cKO compared to αSMA-Dkk1WT, but
   no differences in the bone marrow, where DKK1 levels are similar
   between αSMA-Dkk1cKO and αSMA-Dkk1WT mice (Fig. [205]7F, G,
   Supplementary Fig. [206]6F). These results demonstrate the local
   effects of CAF-derived DKK1 in suppressing NK cell activation in the
   TME.

DKK1 directly suppresses human NK cell functions

   To investigate whether DKK1 suppresses human NK cells (hNK), we
   isolated hNK cells from healthy donor PBMCs (Fig. [207]8A) and found
   that recombinant human DKK1 (rhDKK1) significantly decreased hNK cell
   cytotoxicity against the MDA-MB-231 TNBC cell line (Fig. [208]8B), the
   T47D ER^+ breast cancer cell line (Fig. [209]8C) and the NK-sensitive
   K562 cell line (Fig. [210]8D). Furthermore, rhDKK1 led to a significant
   decrease in hNK cell activating receptors NKG2D, NKp30, and NKp46
   (Fig. [211]8E, Supplementary Fig. [212]7A). Interestingly, we did not
   observe changes in the expression of NK cell activating and/or
   inhibitory ligands on MDA-MD-231 breast cancer cells following rhDKK1
   stimulation (Supplementary Fig. [213]7B–D). These results were also
   confirmed using a microarray dataset from human breast cancer samples
   ([214]GSE3744^[215]33, Supplementary Fig. [216]7E–G). Collectively,
   these findings suggest that in human breast cancer, DKK1 directly
   suppresses NK cells, rather than modulating the tumor cells.

Fig. 8. DKK1 levels correlate with reduced therapeutic responses and
cytotoxic NK cells.

   [217]Fig. 8
   [218]Open in a new tab

   A Schematic representation of NK cell isolation from healthy donor
   human PBMCs and incubation with target cells in the presence of PBS or
   rhDKK1 for 4 h. B–D Analysis of percent specific killing measured by
   7-AAD^+ MDA-MB-231 (B, n = 7 NK donors), T47D (C, n = 3 NK donors), and
   K562 (D, n = 3 NK donors) target tumor cells after 4 h incubation with
   human NK cells in the presence of rhDKK1 or PBS as a control. E FACS
   analysis of NK cell activating receptors on human NK cells from
   different donors (n = 11) stimulated with rhDKK1 for 24 h or PBS. F
   Schematic representation of blood sample collection of advanced breast
   cancer patients at the time of diagnosis of bone metastases and at
   15–18 months follow-up visit receiving standard-of-care and
   antiresorptive treatments. G, H DKK1 plasma levels, and percent of NK
   cell subsets in patients with regressive/stable (blue, n = 7) versus
   progressive bone metastases (red, n = 8) from initial diagnosis
   (abbreviated as I) and follow-up visits (abbreviated as F). Results are
   shown as mean ± SEM (B–D). Two-way ANOVA followed by Bonferroni
   multiple-comparison test (B–D), and a paired t-test with a two-tailed
   P-value (E, G, H) were used to determine significance. *P < 0.05,
   **P < 0.01, ****P < 0.0001. Source data are provided as a Source Data
   file.

DKK1 levels correlate with metastatic progression and reduced cytotoxic NK
cells in breast cancer patients

   To determine whether DKK1 levels correlate with tumor progression and
   immune suppression in breast cancer patients, we analyzed DKK1 plasma
   levels and the activation status of NK cells in the blood of 15
   patients with stage IV, HER2^−, ER^+ breast cancer and skeletal
   disease, at time of diagnosis and after 15–18 months of
   standard-of-care endocrine therapy plus Denosumab as antiresorptive
   therapy. 7 patients were classified as stable as they had no
   radiographic evidence of skeletal metastatic progression during the
   study period. 8 patients were classified as progressive as they
   demonstrated radiographic evidence of new or progressive skeletal
   lesions (Fig. [219]8F, Table [220]1). Although DKK1 levels in
   circulation were not significantly different in the stable and
   progressive patients at baseline, patients with progressive skeletal
   metastases demonstrated a significant increase in DKK1 compared to the
   time of diagnosis (Fig. [221]8G).

Table 1.

   Characteristics of breast cancer patients including time of sample
   collection (initial diagnosis of metastatic dissemination to bone and
   at 15–18-month follow-up visit), skeletal and visceral metastasis (met)
   status, and treatments received at the indicated time. All patients
   received subcutaneous (SQ) Denosumab 120 mg every 4–12 weeks as an
   antiresorptive therapy at the time of diagnosis of skeletal metastases
   Bone met status Timeline Bone metastasis Visceral metastasis Treatments
   Stable

   Initial

   Follow up

   Sternum, Sacrum, Ilium, Lumbar Spine

   No changes

   Nodal disease

   Increase R axillary nodes
   Letrozole

   Initial

   Follow up

   Spine, Sternum, Clavicle, Pelvis

   No Changes

   None

   None
   Anastrozole

   Initial

   Follow up

   Spine

   No changes

   Lung

   No changes

   1. Fulvestrant

   2. Zoladex

   Initial

   Follow up

   Sternum, Ilium

   No changes

   None

   None
   Anastrozole

   Initial

   Follow up

   Sternum

   No changes

   Lung, Liver, Lymph node

   Progressive Liver mets

   1. Everolimus + Exemestane

   2. Capecitabine

   Initial

   Follow up

   Glenoid

   No changes

   None

   None
   Fulvestrant

   Initial

   Follow up

   Sternum, Pelvis

   No changes

   None

   None
   Letrozole
   Progressive

   Initial

   Follow up

   Spine, Sternum, Clavicle, Pelvis, Scapula

   Progression of mets in the axial skeleton

   Lymph node, Pleural mass, soft tissue

   Progessive hepatic mets

   1. Tamoxifen

   2. Letrozole + Palbociclib

   Initial

   Follow up

   Spine, Sacrum

   Increase T4-T6 Spine Mets

   Lung

   Progessive Lung nodules
   Letrozole + Palbociclib

   Initial

   Follow up

   Spine

   New humerus met

   None

   None
   Anastrozole

   Initial

   Follow up

   Ribs, Spine, Sternum

   Progression of osseous mets

   Liver

   Increase Liver mets

   1. Everolimus + Exemestane

   2. Tamoxifen

   Initial

   Follow up

   Spine, Ribs, Ilium, Sacrum

   Progessive mets, New Spine, Femur, Ischium mets

   None

   None
   Aromasin

   Initial

   Follow up

   Vertebra, Pelvis

   New Humerus met

   Lymph node, Liver

   Progressive nodal met

   1. Fulvestrant

   2. Letrozole + Palbociclib

   Initial

   Follow up

   Diffuse multifocal osseous mets

   Progression of multifocal osseous mets

   None

   None
   Palbociclib + Tamoxifen

   Initial

   Follow up

   Diffuse multifocal osseous mets

   Progression of multifocal osseous mets

   Lung, Liver

   Progression of Lung/Liver mets
   Tamoxifen + AKT inhibitor
   [222]Open in a new tab

   To assess whether patients with increased DKK1 had reduced and/or
   dysfunctional NK cells, we measured the number of CD3^−CD56^+ NK cells
   in circulation and found no changes between the stable and progressive
   patients (Supplementary Fig. [223]8A, B). However, the percentage of
   cytotoxic CD16^+CD56^dim NK cells^[224]50 and expression of perforin
   (PRF1) and granzyme B (GZMB) were reduced in patients with progressive
   skeletal metastases (Fig. [225]8H) compared to baseline.

Discussion

   In this study, we demonstrate that high levels of DKK1 in breast cancer
   create an immune suppressive microenvironment through direct inhibition
   of NK cell effector functions, and targeting DKK1 has significant
   anti-tumor effects in the primary and metastatic settings. Moreover, we
   report a previously unappreciated role for DKK1 in inducing systemic
   and local immune suppression. Bone cells contribute to the elevated
   levels of DKK1 in circulation, while CAFs contribute to DKK1 production
   in the TME and do not affect DKK1 systemically. Intriguingly, deletion
   of either source of DKK1 results in a significant anti-tumor effect.
   Our findings indicate the importance of monitoring both systemic and
   tumor DKK1 levels in breast cancer patients and warrant the exploration
   of neutralizing DKK1 to achieve an efficient therapeutic response.
   These results position DKK1 and its inhibitory effects on NK cells as a
   barrier to immunotherapies and an important driver of breast cancer
   progression.

   Breast cancer is poorly immunogenic, which accounts for the limited
   response to immune-based anti-tumor therapies. Local production of DKK1
   at the tumor site limits tumor immune infiltration, and DKK1
   neutralization or its deletion in CAFs reverses these effects.
   Differently from the B16F10 melanoma model dependent on T
   cells^[226]23, T cell depletion in our breast cancer models does not
   limit the therapeutic efficacy of mDKN01. These results are in line
   with the inconclusive clinical trials in ER^+ breast cancer showing
   limited effects of T cell-based therapies^[227]5. Instead, we identify
   NK cells as important modulators of the anti-tumor effects of mDKN01.
   Our findings are in line with the observation that the intratumoral
   abundance of NK cells correlates with increased pCR rate in breast
   cancer patients undergoing neoadjuvant chemotherapy^[228]10 and
   findings in the B16F0 melanoma model dependent on NK cells^[229]24.

   Direct effects of DKK1 on NK cells have never been reported. We
   demonstrate that DKK1 suppresses NK cell activation and cytotoxicity.
   DKK1 directly reduces the phosphorylation of AKT, S6, and ERK1/2, which
   are key modulators of NK cell activation. These observations are
   consistent with the transcriptional downregulation of IFNγ and
   PI3K/AKT/mTOR pathways in NK cells exposed to DKK1. Subsequently, NK
   cells isolated from a DKK1-deficient TME show improved responses to
   IL-12, IL-15, and NK1.1 stimulation. The lower levels of perforin,
   adhesion molecules, and pathways related to cytoskeletal
   reorganization, together with reduced surface expression of CD107a
   following NK1.1 engagement, further suggest defects in the formation of
   immunological synapses with their target cells. In support of this
   hypothesis, NK cells are not found in close contact with the target
   tumor cells when exposed to DKK1. We acknowledge that there are
   additional potential mechanisms whereby DKK1 may indirectly impact NK
   cell responses. A prior study identified that DKK1 could induce
   downregulation of NKG2D and DNAM-1 ligands in a metastatic latency
   breast cancer model^[230]51. In addition, there was downregulation of
   death receptors (FAS and TRAILR), which may protect against this mode
   of NK cell killing. This idea is further supported by the evaluation of
   NKG2D-CAR T cells in gastric cancer, whereby DKK1 suppressed NKG2D
   ligands, reducing NKG2D-based recognition^[231]52. Other studies have
   identified that DKK1-induced macrophages and immature myeloid
   populations become immunosuppressive, thereby inhibiting effector CD8^+
   T cells and NK cells in gastric cancer^[232]23,[233]26. In future
   studies, these mechanisms will be investigated, as well more general
   impact on the tumor microenvironment. Importantly, DKK1 levels should
   be monitored in future NK cell-directed clinical trials due to its NK
   suppressive effects that were not previously appreciated.

   Our findings demonstrate the importance of monitoring various sources
   of DKK1 during tumor progression. DKK1 is expressed in cancer
   epithelial cells in TNBC and HER2^+ tumors, in CAFs, and in the bone.
   Surprisingly, we observe a strong nuclear DKK1 staining in the tumor
   cells, which is not typical for a secreted protein, although we cannot
   exclude that DKK1 has been secreted. Nuclear DKK1 has been previously
   associated with chemoresistance and poor outcomes in colorectal cancer,
   where DKK1 binds to specific chromatin sites to regulate the expression
   of genes involved in the detoxification of chemotherapeutic
   agents^[234]40. In contrast, a more diffused DKK1 localization appears
   in CAFs. Co-culture assays demonstrate that DKK1 produced by CAFs
   reduces NK cell cytotoxicity, a process restored by DKK1
   neutralization. Furthermore, the growth of MDA-MB-231 tumor cells,
   which display nuclear DKK1 localization, is reduced when these cells
   are co-injected with Dkk1-deficient CAFs compared to Dkk1-sufficient
   CAFs. These results highlight the importance of CAF-derived DKK1 in
   supporting tumor progression regardless of DKK1 tumoral expression,
   although we cannot exclude that tumor cells also contribute to the
   release of DKK1 in the TME in patients. Interestingly, Dkk1 is also
   highly expressed in bone and its deletion exerts profound anti-tumor
   effects. Like CAFs, DKK1-producing pre-osteoblasts reduce NK cell
   cytotoxicity, which can be restored by DKK1 neutralization. It is still
   unclear why bone and CAF-derived DKK1 do not compensate for each other,
   and that deletion of Dkk1 in one or the other compartment has such
   profound anti-tumor effects. Ex vivo NK cell reactivation assays
   clearly demonstrate that deletion of Dkk1 in CAFs improves NK cell
   responses in the TME but not in the bone marrow. Since bone is the
   major site for hematopoiesis, we cannot exclude that bone-derived DKK1
   might alter NK cell progenitors by reprogramming hematopoiesis, a
   process in part modulated by DKK1^[235]53.

   Bone is a preferred organ for breast cancer metastatic dissemination.
   DKK1 is increased in breast cancer patients with progressive bone
   metastases compared to those with stable disease or visceral
   metastases^[236]15. While the source of DKK1 in these patients is
   unknown, the low expression of DKK1 in the ER^+ tumors by IHC and
   scRNAseq suggests that bone might be the primary site for DKK1
   production in ER^+ breast cancer. The supportive role of DKK1 on
   metastatic dissemination to bone has been associated with the local
   activation of osteoclast-mediated bone resorption^[237]29. However, our
   clinical findings show that high levels of DKK1 correlate with the
   progression of bone metastatic disease despite treatment with
   anti-resorptives. Instead, we find that the percentages of circulating
   cytotoxic NK cells were reduced in the progressive patients, suggesting
   that NK cells might also modulate tumor growth in bone.

   In sum, our findings position DKK1 as a barrier for efficient
   anti-tumor immunity in breast cancer through its suppressive effects on
   NK cell activation and killing efficiency. Monitoring DKK1 levels in
   circulation and in the TME, and DKK1 neutralization should be
   considered to improve the therapeutic response of NK-based strategies
   and/or the limited efficacy of immunotherapies in breast cancer.

Methods

   This study complies with all relevant ethical regulations including
   protocols approved by the Institutional Animal Care and Use Committee
   and guidelines set by the Institutional Review Board of Washington
   University (IRB ID#: 201102244) along with federal and state
   guidelines.

Cell culture

   Polyoma middle tumor-antigen murine mammary tumor cells (PyMT,
   C57BL/6), mCherry-conjugated PyMT (PyMT-mCherry), PyMT-derivative
   PyMT-BO1 conjugated with firefly luciferase (PyMT-BO1-fluc), H2B-mApple
   and Thy1.1 conjugated PyMT-BO1 (PyMT-BO1-GFP-fluc-H2B-mApple-Thy1.1),
   E0771 murine mammary tumor cells (E0771-fluc, C57BL/6), 4T1 murine
   mammary tumor cells (4T1-GFP-fluc, BALB/c), T47D and MDA-MB-231 human
   breast cancer cells were cultured at 37 °C with 5% CO[2] in complete
   media (DMEM supplemented with 100 μg/ml streptomycin, 100 IU/ml
   penicillin, and 1 mM sodium pyruvate) containing 10% FBS. K562 cell
   line was cultured in complete media (RPMI 1640 supplemented with 2 mM
   L-glutamine, 100 μg/ml streptomycin, 100 IU/ml penicillin, 1×
   nonessential amino acids, and 1 mM sodium pyruvate) containing 10% FBS.
   All cell lines tested negative for Mycoplasma contamination during
   routine evaluations. Aliquots for each cell line were used for a
   maximum of 1 month after the initial thaw.

Animal models

   Because the breast cancer cell lines used in this study were obtained
   from female mice, we have restricted our analyses to females. Female
   wild-type (WT) C57BL/6, WT BALB/c, B6(Cg)-Tyr^c^-^2J/J (albino
   C57BL/6), B6.129S7-Rag1^tm1Mom/J (Rag1^−^/^−), Nu/J (Nude),
   NOD.Cg-Prkdc^scidIl2rg^tm1Wjl/SzJ (NSG), C57BL/6-Prf1^tm1Sdz/J
   (Prf1^−^/^−), Sp7 Cre (Sp7-tTA, tetO-EGFP/Cre) mice were purchased from
   The Jackson Laboratory. Mice arrived at 4–6 weeks of age and were
   allowed to recover from shipping stress and acclimatize to the new
   environment for at least 2 weeks before use in experiments.
   αSMACreER^T2 transgenic mice were a generous gift from Dr. Ivo Kalajzic
   (University of Connecticut Health Center, Farmington, CT^[238]54). FSP1
   Cre mice were a generous gift from Dr. Regis J. O’Keefe (Washington
   University, St. Louis, MO^[239]55). Dkk1 floxed mice were a generous
   gift from Seppo J. Vainio (University of Oulu, Finland^[240]56). In
   experiments using conditional KO mice, Cre^+;Dkk1WT or Cre^−;Dkk1^fl/fl
   were used interchangeably as littermate controls. Animals were housed
   in a pathogen-free animal facility at Washington University (St. Louis,
   MO) with a 12-h light/12-h dark cycle and 20~23 °C and 40~60% humidity
   housing conditions. All experiments were performed according to
   protocols approved by the Institutional Animal Care and Use Committee
   at Washington University (Protocol ID: 2022-0315).

   To establish tumors, PyMT, PyMT-BO1-fluc,
   PyMT-BO1-GFP-fluc-H2B-mApple-Thy1.1, E0771-fluc, and 4T1-fluc tumor
   cells were suspended in 1:1 PBS/Matrigel ratio (Corning 354234) and
   10^5 cells injected into the MFP or inoculated subcutaneously in the
   left flank. For models of metastases, 10^4 cells suspended in PBS were
   injected into the right tibia or intracardiacally in female age-matched
   mice. αSMACreER^T2 transgenic mice (αSMA-Dkk1cKO, αSMA-Dkk1WT) were
   intraperitoneally injected with tamoxifen (Sigma, 100 mg/kg)
   resuspended in corn oil starting a day prior to tumor inoculation. For
   MFP and subcutaneous tumors, measurements were performed every other
   day with a caliper, and volumes were calculated using the following
   formula: V = 0.5 (length [mm] × width [mm]^2). The maximal tumor size
   (<2000 mm^3) permitted by the Institutional Animal Care and Use
   Committee was not exceeded. For intratibial and intracardiac tumor
   delivery, growth curves were determined by bioluminescence imaging
   using an IVIS 50 imaging system (PerkinElmer, 1–60 s exposures, binning
   4, 8, or 16, FOV 15 cm, f/stop1, open filter). Mice were i.p. injected
   with D-luciferin (150 mg/kg in PBS; Gold Biotechnology) and imaged
   10 min later under isoflurane anesthesia (2% vaporized in O[2]).
   Bioluminescence photon flux (photons per second) data were analyzed by
   region of interest measurements (fixed region of interest over the
   whole body, or hindlimb) in Living Image 3.2 (Caliper Life Sciences).
   For i.c. injections, mice with extra pleural intrathoracic tumors were
   excluded from the analysis.

In vivo treatments (Neutralizing antibodies)

   The monoclonal anti-mouse anti-DKK1 Ab (Leap therapeutics, mDKN01,
   produced by grafting complementary determining regions of mDKN01
   (IgG4/kappa), with minor modifications, onto a murine IgG2a/kappa
   construct with an FcR incompetent construct D265A substitution to
   abrogate FcR interactions^[241]57) was used to neutralize DKK1.
   Biodistribution analysis of infrared (IR)-dye conjugated mDKN01 or
   IgG2a antibody was evaluated in mice bearing primary tumors to confirm
   mDKN01 delivery at tumor site (Supplementary Fig. [242]9). To assess
   the anti-tumor effects of DKK1 targeting, treatment consisted of
   intraperitoneal injections of mDKN01 and control antibodies at a
   concentration of 10 mg/kg 2–3 times a week.

   The monoclonal anti-mouse anti-CD4 Ab (BioXCell, clone GK1.5),
   anti-mouse anti-CD8a (BioXCell, clone 2.43) were used to deplete T
   cells and PBS was used as control. Treatment consisted of
   intraperitoneal injections of anti-CD4 and anti-CD8a into mice 2 days
   prior to tumor cell implantation using the following regimen: 500μg on
   the first dose and 250μg for the subsequent doses every 4 days of each
   antibody.

   The monoclonal Ab anti-mouse anti-NK1.1 Ab (Leinco Technologies, Inc.,
   clone PK136), was used to deplete NK cells and PBS was used as control.
   Treatment consisted of intraperitoneal injections of anti-NK1.1 into
   mice 2 days prior to tumor cell implantation at a dose of 100μg once a
   week. Effective depletion of T and NK cells was assessed by flow
   cytometry of peripheral blood.

Multiplex IHC

   Freshly isolated mouse primary tumors were fixed in 10%
   neutral-buffered formalin (DiRuscio & Associates, Inc.) for 24 h.
   Tissues were paraffin-embedded and sectioned 5 μm thick by the
   histology core of the Washington University Musculoskeletal Research
   Center.

   Human tissue microarrays were prepared by The St. Louis Breast Tissue
   Registry (funded by The Department of Surgery at Washington University
   School of Medicine, St. Louis, MO). Data and tissues were obtained in
   accordance with the guidelines established by the Washington University
   Institutional Review Board (IRB #201102394) and WAIVER of elements of
   Consent per 45 CFR 46.116 (d). All patient information was deidentified
   prior to sharing with investigators.

   Tissues were automatically stained via the Bond Rxm (Leica Biosystems)
   following dewaxing and appropriate epitope retrieval. Immunostaining
   was chromogenically visualized using the Bond Polymer Refine Detection
   (#DS9800, Leica Biosystems) or the Bond Polymer Refine Red Detection
   (#DS9390, Leica Biosystems). Slides were mounted using Xylene-based
   Cytoseal (Thermo Fisher) or Vectamount (Vector Labs) as appropriate,
   scanned via a Zeiss AxioScan 7 microscope and IHC analyses performed
   using the HALO image analysis platform (Indica Labs, Deconvolution
   v1.1.1, Multiplex IHC v.3.2.3 algorithms). The following antibodies
   were used: αSMA (Abcam, Cat# ab5694, 1:1500 (Human), 1:200 (Mouse)),
   COL1a1 (Cell signaling, Cat# 72026, 1:100 (Mouse)), COL14a1 (Cell
   signaling, Cat# 61964, 1:200 (Human)), DKK1 (Proteintech, Cat#
   21112-1-AP, 1:100 (Human), 1:3000 (Mouse)), PDGFRα (Cell signaling,
   Cat# 5241, 1:200 (Human)), PanCK (Novus, Cat# NBP2-29429, 1:1000
   (Human)).

scRNAseq analysis

   For scRNAseq analysis of the human breast cancer dataset, the matrix
   was downloaded from the European Genome-Phenome Archive (EGA)
   EGAS00001005173 and analyzed using Seurat version 4. Seurat object
   generation and further data processing were conducted as
   described^[243]39. Non-tumor (negative for KRT8, 18, 14, 17, and EPCAM)
   non-immune cell populations were re-clustered for in-depth analysis of
   endothelial cells (positive for PECAM1) and CAF subpopulations
   (positive for PDGFRB) as shown in ref. ^[244]39. The identities of CAF
   subpopulations were determined based on the expression of CAF signature
   genes identified by previous literature^[245]37,[246]38. Specifically,
   myCAFs were identified as CAFs positive for PDGFRA, ACTA2, THBS2, and
   POSTN; vascular CAFs (vCAF) were identified as CAFs positive for MCAM,
   ACTA2, DES, and NOTCH3; inflammatory CAFs (iCAF) were identified as
   CAFs positive for PDGFRA, CXCL12, COL14A1, and ALDH1A1, all of which
   have been illustrated in a previous work^[247]39. Visualizations of
   DKK1 expression were performed based on the adaptively-threshold
   low-rank approximation (ALRA) assay of the dataset^[248]58.

Bulk RNAseq analysis

   RNA was isolated from the sorted cells using RNeasy Micro Kit (Qiagen).
   Total RNA integrity was determined using Agilent Bioanalyzer or 4200
   Tapestation. Library preparation was performed with 10 ng of total RNA
   with a Bioanalyzer RIN score greater than 8.0. ds-cDNA was prepared
   using the SMARTer Ultra Low RNA kit for Illumina Sequencing
   (Takara-Clontech) per the manufacturer’s protocol. cDNA was fragmented
   using a Covaris E220 sonicator using peak incident power 18, duty
   factor 20%, and cycles per burst 50 for 120 s. cDNA was blunt-ended,
   had an A base added to the 3′ ends, and then had Illumina sequencing
   adapters ligated to the ends. Ligated fragments were then amplified for
   12–15 cycles using primers incorporating unique dual index tags.
   Fragments were sequenced on an Illumina NovaSeq-6000 using paired-end
   reads extending 150 bases.

   Partek Flow software (Partek Inc., St. Louis, MO) was used for data
   analysis. Briefly, sequenced reads were aligned with STAR 2.7.8a index
   (Mus musculus – mm10 assembly, Whole genome index). Raw read counts
   were obtained by quantitating aligned reads using HTSeq with the
   Ensembl Transcripts release 102 annotation model. Raw read counts were
   then normalized using counts per million (CPM) and an offset of 0.0001
   was added to all normalized read counts. Normalized read counts for
   each mRNA were statistically modeled using Partek Flow’s Gene Specific
   Analysis (GSA) approach. Differentially expressed genes (DEGs) were
   then filtered by using P-values less than or equal to 0.05 and fold
   changes bigger than 2. KEGG pathway enrichment analysis was performed
   with Partek Flow software (Partek Inc., St. Louis, MO). GSEA analysis
   was performed as previously described^[249]59,[250]60.

Multiparametric flow cytometry

   Immediately upon sacrifice, single-cell suspensions were prepared from
   tumors. In brief, tumor tissues were minced, and then digested with
   3.0 mg/ml collagenase A (Roche) and 50 U/ml DNase I (Sigma-Aldrich) in
   serum-free media for 30 min at 37 °C. Cells were filtered through 70 μm
   nylon strainers (Thermo Fisher Scientific) and washed twice in PBS with
   2% FBS. Red blood cells were then removed with red blood cell lysis
   buffer (Sigma-Aldrich). Cells were washed once, blocked with anti-mouse
   CD16/CD32 blocker, and stained in PBS with 0.5% BSA, 2 mM EDTA, and
   0.01% NaN[3] with the anti-mouse antibodies. The acquisition was
   performed on a BD LSRFortessa X-20 Cell Analyzer or CYTEK Northern
   Lights and the dedicated software Diva (BD) or CYTEK SpectroFlo. Data
   were analyzed with FlowJo 10.9.0 software (Tree Star).

   The following antibodies were used: Anti-mouse CD3e-FTIC (Biolegend,
   Cat# 100306, Clone 17A2, 1:400), CD4-APC (BD Pharmingen, Cat# 561091,
   Clone RM4-5, 1:200), CD8a-BUV395 (BD Horizon, Cat# 565968, Clone
   53-6.7, 1:200), CD11b-BUV395 (BD Biosciences, Cat# 565976, Clone M1/70,
   1:400), CD16/32 (Biolegend, Cat# 101302, Clone 93, 1:500), CD45-APC Cy7
   (Biolegend, Cat# 103116, Clone 30-F11, 1:400), CD45-BV605 (Biolegend,
   Cat# 103140, Clone 30-F11, 1:400), CD107a-eFlour660 (Thermo Fisher,
   Cat# 50-1071-82, Clone 1D4B, 1:200), F4/80-BV711 (Biolegend, Cat#
   123147, Clone BM8, 1:100), IFNg-BV750 (Biolegend, Cat# 505865, Clone
   XMG1.2, 1:50), Ly6C-APC (Biolegend, Cat# 128016, Clone HK1.4, 1:400),
   Ly6G-BV421 (BD Biosciences, Cat# 562737, Clone 1A8, 1:400), NK1.1-BV711
   (Biolegend, Cat# 108745, Clone PK136, 1:200), pAKT-BV421 (BD phosflow,
   Cat# 562599, Clone M89-61, 1:20), pERK1/2-PE (BD phosflow, Cat# 612566,
   Clone 20A, 1:10), pS6-eFlour450 (BD phosflow, Cat# 561457, Clone
   N7-548, 1:20), pSTAT5-Alexa Fluor647 (BD phosflow, Cat# 612599, Clone
   47/Stat5, 1:20), Ter119-BV605 (Biolegend, Cat# 116239, Clone TER119,
   1:200), CD90.1- eFlour450 (Thermo Fisher, Cat# 48-0900-82, Clone HIS51,
   1:200).

   Anti-human CD3-ECD (Immunotech, Cat# A07748, Clone UCHT1, 1:50),
   CD16-PerCP Cy5.5 (BD Pharmingen, Cat# 560717, Clone 3G8, 1:200),
   CD45-BV605 (Biolegend, Cat# 304042, Clone 2D1, 1:20), CD56-PE Cy7
   (Immunotech, Cat# A51078, Clone N901, 1:100), CD58-PerCP Cy5.5
   (Biolegend, Cat# 330914, Clone TS2/9, 1:50), B7-H6-PE (R&D systems,
   Cat# FAB7144P, Clone 875001, 1:20), GzmB-AF700 (BD Pharmingen, Cat#
   560213, Clone GB11, 1:50), HLA-E-APC (Biolegend, Cat# 342606, Clone
   3D12, 1:20), NKG2D-APC (Invitrogen, Cat# 17-5878-42, Clone 1D11, 1:20),
   NKp30-BV785 (Biolegend, Cat# 325229, Clone P30-15, 1:20), NKp46-BV421
   (BD Horizon, Cat# 564065, Clone 9E2, 1:20), PRF1-PE (Biolegend, Cat#
   308106, Clone dG9, 1:50).

NK cell killing assays

   Cytotoxicity of NK cells was assessed in a standard 4h flow
   cytometry-based ex vivo killing assay as previously described^[251]61.
   4 × 10^4 tumor cells were plated in 96-well plates and different
   numbers of NK cells were added one hour later. Recombinant mouse DKK1
   (Biolegend) or human DKK1 (Biolegend) was added to a final
   concentration of 200 ng/mL in RPMI 1640 media. For each effector:
   target (E:T) ratio, Percent Specific Killing was calculated as [%
   7-AAD^+ of CTV^+ cells](Effector+Target) well – [% 7-AAD^+ of CTV^+
   cells](Targets only) well.

   Specific killing of control and rDKK1-stimulated NK cells was also
   evaluated using the IncuCyte Live-cell Analysis system (Satorius). A
   total of 5 × 10^3 mApple^+ PyMT tumor cells were incubated for 2 h in a
   96-well plate and imaged prior to the addition of the NK cells at a 2:1
   E:T ratio. Real-time images were captured every 4 h and up to 48 h and
   analyzed using the IncuCyte software. Data are presented as red object
   counts (mApple^+ cells).

   Osteoblast precursors were generated from bone marrow harvested from
   tibias and femurs of PyMT tumor-bearing mice by centrifugation and
   cultured for 7 days on a tissue culture dish in αMEM media at 37 °C.
   Adherent cells were detached with Trypsin LE Express (Gibco) and plated
   at increasing numbers 24 h before plating NK cells (8 × 10^4 Cells) and
   tumor cells (4 × 10^4 Cells) in the presence of IgG or mDKN01
   (50 μg/ml). Analysis of NK tumoricidal activity was performed as
   indicated above.

   CAFs were isolated by plating the single-cell suspension of the tumor
   mass from αSMA-Dkk1WT-tdT mice for 30 min on a tissue culture dish at
   37 °C, followed by replacing the media to remove immune and tumor cells
   in suspension. After 20 h of culture, the adherent cells were carefully
   detached with accutase digestion, stained with anti-CD45, and sorted
   based on tdT expression and exclusion of CD45^+ cells. Isolated CAFs
   were plated at increasing concentrations 24 h before plating NK cells
   (8 × 10^4 Cells), tumor cells (4 × 10^4 Cells) + IgG or mDKN01
   (50μg/ml). Analysis of NK tumoricidal activity was performed as
   indicated above.

Analysis of phosphorylated signaling pathways in NK cells

   Spleens from Rag1^−^/^− mice were mechanically disassociated with a 1
   cc syringe plunger and passed through a 70 μm cell strainer. Red blood
   cells were lysed with red blood cell lysing buffer Hybrid Max
   (Sigma-Aldrich). Splenocytes were washed with PBS and 5 × 10^5 cells
   were plated in 96-well plates in media without IL-15 followed by rDKK1
   (200 ng/ml) stimulation at 37 °C for indicated amount of time. After
   incubation, cells were stained with anti-NK1.1 antibody and fixed with
   1% PFA for 10 min at room temperature and permeabilized with ice-cold
   100% methanol for 30 min at 4 °C. Cells were washed three times with
   the FACS buffer, stained with antibodies against pAKT, pS6, pERK1/2,
   and pSTAT5, followed by acquisition with CYTEK Northern Lights. Mean
   fluorescence intensity (MFI) was analyzed from NK1.1^+ cells with
   FlowJo 10.9.0 software (Tree Star).

Ex vivo NK cell activation

   Single-cell suspensions were prepared from the tumor mass and bone
   marrow from αSMA-Dkk1cKO and αSMA-Dkk1WT mice. Red blood cells were
   lysed with red blood cell lysing buffer Hybrid Max (Sigma-Aldrich).
   Cells were washed with PBS and 1–2 × 10^6 cells were plated in media
   containing 10 ng/mL IL-15 and anti-CD107a flow antibody onto 96-well
   plates coated with anti-NK1.1 antibody (PK136, 5μg/mL), or simulated
   with IL-12 and IL-15 cytokine cocktail (10 ng/ml each). After 1 h at
   37 °C, GolgiPlug (BD biosciences) and GolgiStop (BD biosciences) were
   added for an additional 3 h. Cells were washed with the FACS buffer,
   and stained with anti-CD45, CD3, CD49b, and NKp46 flow antibodies.
   IL-12 and IL-15 stimulated cells are then fixed/permeabilized with BD
   cytofix/cytoperm buffer for 30 min at 4 °C, stained with anti-IFNγ
   antibody, and acquired on CYTEK Northern Lights. Live NK cells
   (CD45^+CD3^−CD49b^+NKp46^+) were analyzed with FlowJo 10.9.0 software
   (Tree Star).

Human NK cell purification and cell culture

   Human platelet apheresis donor PBMCs were obtained by Ficoll
   centrifugation. NK cells were purified using RosetteSep (StemCell
   Technologies, ≥95% CD3^−CD56^+). Cells were plated at 3–5 × 10^6
   cells/mL and cultured in complete RPMI 1640 medium containing 10% human
   AB serum (Sigma-Aldrich) supplemented with rhIL-15 (1 ng/mL) to support
   survival, with 50% of the medium being replaced every other day with
   fresh cytokines.

Immunofluorescence and Confocal microscopy

   MDA-MB-231 tumor cells were plated overnight onto FBS-coated slides,
   fixed with 4% PFA, and permeabilized with 0.3% Triton X-100. Cells were
   stained with DKK1 antibody (Proteintech (21112-1-AP)) followed by
   anti-rabbit-Alexa 488 secondary antibody (Abcam, ab150077, 1:1000) and
   DAPI (Abcam, ab228549) for an hour and mounted with anti-fade
   fluorescence mounting medium (Abcam).

   PyMT-mCherry tumor cells were plated onto the poly-D-lysine (Millipore
   sigma) overnight coated slides and incubated for an hour before adding
   NK cells. NK and tumor cells were cultured together for 3 h prior to
   fixation with 4% PFA and permeabilization with 0.3% Triton X-100. Cells
   were stained with an F-actin antibody conjugated with Alexa 488
   (Invitrogen, A12379, 1:200) for an hour and mounted with an anti-fade
   fluorescence mounting medium (Abcam).

   Confocal images were generated using a Nikon AX-R Confocal Microscope
   (100× objective lens, oil immersion, Numerical Aperture: 1.45,
   Refractive Index: 1.515). The camera setting was set in multi-channel
   detector mode (FITC and mCherry), Galvano unidirectional scanner with
   Band scan mode, 4× line averaging, 1.0 μs dwell time. Images were
   analyzed with NIS-Elements software (Nikon, ver.5.21.00) and
   ImageJ^[252]62.

Real-time PCR analysis

   RNA was isolated from the cell line or tumor mass using RNeasy Mini Kit
   (Qiagen). Purified RNA was then reverse transcribed to cDNA using a
   High Capacity cDNA reverse transcription kit (Applied Biosystems)
   according to the manufacturer’s instructions. The subsequent real-time
   PCR analysis was performed with SYBR Green PCR Master Mix (Applied
   Biosystems) and primers specific for murine DKK1 and cyclophilin were
   used as follows: for Dkk1, CTC ATC AAT TCC AAC GCG ATC A (forward), GCC
   CTC ATA GAG AAC TCC CG (reverse) and for cyclophilin AGC ATA CAG GTC
   CTG GCA TC (forward) and TTC ACC TTC CCA AAG ACC AC (reverse).

   For primers specific for human NK cell activating/ inhibitory receptor
   ligands were used as follows: for HLA-E, TTC CGA GTG AAT CTG CGG AC
   (forward), GTC GTA GGC GAA CTG TTC ATA C (reverse), PVR, GGA CGG CAA
   GAA TGT GAC CT (forward) GGT CGT GCT CCA ATT ATA GCC T (reverse),
   MIC-A, CTT CAG AGT CAT TGG CAG ACA T (forward), TGT GGT CAC TCG TCC CAA
   CT (reverse), PVRL2, CAC TTG CGA GTT TGC CAC C (forward), GCC ACT GTC
   GTA GGG TCC T (reverse), CD58, AGA GCA TTA CAA CAG CCA TCG (forward),
   ATC TGT GTC TTG AAT GAC CGC (reverse).

MTT assay

   5 × 10^2 PyMT or 4T1 cells and 2 × 10^3 E0771 cells were plated for
   24 h before treatment with rDKK1 (50, 100, 200 ng/ml) or tamoxifen (1,
   2, 10 μM) for the indicated amount of time. The MTT assay (Invitrogen)
   and annexin V assay (Thermo Fisher Scientific) were performed per the
   manufacturer’s protocol.

Human study

   All human samples were obtained in accordance with guidelines set by
   the Institutional Review Board of Washington University (IRB ID#:
   201102244) and followed federal and state guidelines. All participants
   gave written informed consent under the IRB-approved protocol prior to
   inclusion in the study, including access to archival tumor tissue for
   research. Samples were deidentified prior to sharing with
   collaborators. All studies were conducted in compliance with the
   Declaration of Helsinki. The clinical information of the enrolled
   patients, including age, timeline of visits, sites of metastatic
   dissemination and treatment histories, was reviewed and documented by
   the breast oncologist C.X.M. Patients were all female with a diagnosis
   of estrogen receptor (ER) positive, HER2 negative breast cancer, stage
   IV, with bone metastases, prior to first-line systemic therapy for
   metastatic breast cancer or had prior therapy for metastatic breast
   cancer but met the following criteria: (i) no prior chemotherapy or
   immune therapy in the past 2 months, (ii) patients currently stable or
   progressing on hormonal therapy or hormonal therapy combination or
   starting hormonal therapy or hormonal therapy combination, and (iii) no
   limitation on the number of prior hormonal therapy or chemotherapy
   treatments. Radiologic tumor assessment was required within 1 month
   prior to or after the collection of the baseline blood sample to serve
   as the baseline tumor assessment.

   Peripheral blood was collected at enrollment and at 15–18 months
   follow-up visits from 15 patients being treated with Denosumab
   (anti-RANKL) and on standard-of-care treatment. To isolate peripheral
   blood mononuclear cells (PBMC), EDTA-treated whole blood was diluted to
   a volume of 20 mL with PBS, transferred to a 50 mL conical tube, and
   underlaid with 15 mL of Ficoll (Atlantal Biologicals). Tubes were
   centrifuged at 400 × g for 30 min. The PBMC fraction was collected at
   the interface layer and washed three times with 40 mL of PBS. After
   counting, a minimum of 5 × 10^6 PBMCs were frozen in 10% [volume for
   volume (v/v)] DMSO (Sigma-Aldrich, catalog no. D5879) in FBS and stored
   in liquid nitrogen for subsequent analysis.

ELISA

   Serum or plasma levels of DKK1 from mice or patients were quantified
   using ELISA kits specific for DKK1 (Mouse Dkk1 Quantikine ELISA kit or
   Human Dkk1 Quantikine ELISA kit, R&D systems) as per the manufacturer’s
   protocols.

Statistical analysis

   In vitro experiments include technical and biological triplicates and
   were performed at least 3 times. In vivo experiments were done with at
   least 4–9 mice per group (the number of mice used for each experiment
   is specified in the figure legends) and at least 3 independent
   experiments were performed. Numerical data are shown as mean ± SEM.
   T-test statistical analysis was used to compare two groups (unpaired
   t-test when comparing murine data sets and paired when comparing human
   cells from single donors or patients). In calculating two-tailed
   significance levels for equality of means, equal variances were assumed
   for the two populations. In some experiments with multiple groups or
   time points, analysis of variance (ANOVA), including Dunnett’s
   multiple-comparison test, or Bonferroni’s multiple-comparison test, was
   used. Results were considered significant at p < 0.05. All statistical
   analyses were performed with GraphPad Prism 10.2.1 software for Windows
   (GraphPad Software).

Reporting summary

   Further information on research design is available in the [253]Nature
   Portfolio Reporting Summary linked to this article.

Supplementary information

   [254]Supplementary Information^ (3.8MB, pdf)
   [255]Reporting Summary^ (5.4MB, pdf)
   [256]Transparent Peer Review file^ (2MB, pdf)

Source data

   [257]Source Data file^ (68KB, xlsx)

Acknowledgements