Abstract

   Tuning immune‐cold tumor hot has largely attracted attention to improve
   cancer treatment, including immunotherapy and antibody‐drug conjugates
   (ADCs). Utilizing multiomic analyses and experimental validation, this
   work identifies a pivotal role for the USP10/B7‐H4 proteolytic axis in
   mediating the interplay between tumor immune responses and ADC
   efficacy, particularly for sacituzumab govitecan (SG) in treating
   triple negative breast cancers (TNBCs). Mechanistically, the inhibition
   of autocrine motility factor receptor (AMFR)‐mediated ubiquitylation of
   B7‐H4 by the deubiquitinase USP10 leads to the stabilization of B7‐H4,
   which suppresses tumor immune activity and reduces SG treatment
   effectiveness. Pharmacological inhibition of USP10 promotes the
   degradation of B7‐H4, enhancing tumor immunogenicity and consequently
   improving the tumor‐killing efficacy of SG. In preclinical TNBC models,
   suppression of USP10/B7‐H4 proteolytic axis is effective in increasing
   SG killing efficacy and reducing tumor growth, especially for the
   tumors with the USP10^high/B7‐H7^high signature. Collectively, these
   findings uncover a novel strategy for targeting the immunosuppressive
   molecule B7‐H4 for cancer therapy.

   Keywords: B7‐H4, deubiquitylation, tumor immunogenicity and ADC
   efficacy, USP10
     __________________________________________________________________

   In immune‐cold triple‐negative breast cancers, elevated B7‐H4
   expression diminishes immune responses and impedes the efficacy of
   sacituzumab govitecan. The regulation of B7‐H4 is a dynamic tug‐of‐war
   between degradation by E3 ligase AMFR and stabilization by
   deubiquitinase USP10. Inhibiting USP10 shifts the balance, restores
   B7‐H4 turnover, revives tumor immunogenicity, and enhances the
   therapeutic impact of sacituzumab govitecan.

   graphic file with name ADVS-11-2400757-g008.jpg

1. Introduction

   How to tune immune‐cold tumor hot has greatly attracted our attention
   to improve cancer treatment.^[ [44]^1 , [45]^2 , [46]^3 , [47]^4 ^]
   Emergence of antibody‐drug conjugates (ADCs) shed light on a new
   therapeutic approach accompanied by immunotherapy in treating
   triple‐negative breast cancers (TNBCs).^[ [48]^5 , [49]^6 ^] TNBCs
   present unique therapeutic challenges due to their lack of hormone
   receptors and HER2 protein, which are common targets in other breast
   cancer subtypes.^[ [50]^7 ^] Although ADCs such as sacituzumab
   govitecan (SG) have benefited TNBC patients, certain subsets encounter
   therapeutic challenges due to de novo and acquired drug resistance.^[
   [51]^8 , [52]^9 , [53]^10 ^] Our recent big data analyses and
   pathological studies with various breast cancer cohorts suggest the
   status of immune‐cold tumors to be a potentially critical barrier
   affecting the tumor response to ADC drug treatment. Nevertheless,
   mechanisms underlying the interplay between tumor immunogenicity and
   ADC drug response remain largely unclear. Our endeavor on multiomic
   analyses coupled with experimental validation has drawn attention to
   the USP10/B7‐H4 proteolytic axis as a possible pivotal target to
   potentiate tumor immune response and augment ADC efficacy in treating
   immune‐cold TNBCs.

   The cytotoxic payload of ADCs is the cornerstone in the battle against
   tumor cells, yet its effectiveness is not solely dependent on the
   agents' cytotoxicity to target tumor cells.^[ [54]^11 ^] It also
   requires a conducive immune environment to optimize its therapeutic
   potential.^[ [55]^12 , [56]^13 ^] ADCs operate at the intersection of
   cancer cell targeting and immune activation, employing mechanisms like
   immunogenic cell death and antibody‐dependent cell‐mediated
   cytotoxicity, as well as initiating dendritic cell maturation. These
   processes not only strike directly at cancer cells but also modify the
   tumor microenvironment to support an immune‐mediated response. However,
   this efficacy can be significantly undermined in what is described as a
   cold‐immune environment, leading to diminished therapeutic responses.^[
   [57]^13 ^] In this regard, understanding the interplay between ADCs and
   the immune environment is essential for enhancing tumor‐specific
   adaptive immunity that promotes a more robust and sustained immune
   response against tumors with ADCs treatment.

   In this study, multiomic analyses identified B7‐H4 as a pivotal factor
   in establishing a suppressive tumor microenvironment conducive to an
   “immune‐cold” response. B7‐H4, also known as V‐set domain containing
   T‐cell activation inhibitor 1 (VTCN1), is a member of the B7 family of
   immune checkpoint ligands. As a transmembrane protein, B7‐H4 acts as a
   co‐inhibitory ligand, regulating T cells, neutrophils, and macrophages,
   which helps cancer cells in evading immune detection.^[ [58]^14 ,
   [59]^15 , [60]^16 , [61]^17 ^] Notably, B7‐H4 is found not only on the
   cell membrane but also within the cytoplasm, where it regulates the
   trafficking of damage‐associated molecular patterns (DAMPs) from ER
   lumen to tumor cell membrane, consequently disrupting the phagocytosis
   of cancer cells by dendritic cells.^[ [62]^3 ^] Furthermore, B7‐H4 is
   upregulated in a range of cancers, with a notable increase especially
   in breast cancer.^[ [63]^18 , [64]^19 , [65]^20 , [66]^21 ^] An
   intriguing observation in breast cancer is the near mutual exclusivity
   of B7‐H4 and PD‐L1 expression, indicating that certain breast cancer
   cases may utilize B7‐H4 as an alternative immune checkpoint pathway.^[
   [67]^3 ^] This dependency on B7‐H4 for immune evasion suggests that
   these cancers may be less responsive to anti‐PD‐L1 therapies,
   highlighting the need for targeted treatments against the B7‐H4
   pathway.

   With the reality that the immune system's engagement can be profoundly
   influential in ADCs combating TNBC, our protein purification coupled
   with mass spectrometry analysis has identified USP10 as a crucial
   deubiquitinating enzyme, stabilizing this immune‐suppressive protein
   and facilitating immune evasion that resist to ADCs‐induced tumor
   eradication. Interestingly, ADC treatments seem to trigger a feedback
   loop that boosts USP10 expression, which, in turn, increases B7‐H4
   levels, further compromising the immune system's ability to monitor and
   attack tumor cells. This leads to diminished T cell activity,
   proliferation, and cytokine production. To overcome this therapeutic
   challenge, we have adopted a pharmacological approach to inhibit USP10,
   thereby diminishing B7‐H4 stability and rejuvenating immunogenicity of
   TNBC cells, increasing their susceptibility to the cytotoxic effects of
   ADCs. Our findings also reveal that suppression of USP10‐B7‐H4
   proteolytic axis was effective in resensitizing SG killing efficacy and
   reducing tumor growth, especially for the tumors with the
   USP10^high/B7‐H7^high signature. This study uncovers a novel strategy
   for targeting the immunosuppressive molecule B7‐H4 for anti‐TNBC
   therapy.

2. Results

2.1. Elevated B7‐H4 Expression Correlates with Immune‐Cold Feature of Breast
Cancers and Poor Clinical Outcome for Antibody‐Drug Conjugate (ADC) Treatment

   To explore novel therapeutic targets that could enhance the immunogenic
   cytotoxic response and overcome resistance in immune‐cold breast
   cancers, we undertook a comprehensive big data analysis of proteomics
   from The National Cancer Institute's Clinical Proteomic Tumor Analysis
   Consortium (CPTAC). This analysis specifically targeted over 100
   proteins implicated in modulating the tumor immune microenvironment,
   aiming to identify pivotal molecular players that could potentiate the
   efficacy of existing and novel therapeutic agents.^[ [68]^3 ^] Our
   analysis revealed elevated expressions of multiple tumor immune
   response‐related proteins, including B7‐H4 (VTCN1), PD‐L1 (CD274), and
   Galectin9 (LGLAS9), in the TNBC compared to other subtypes (Figure
   [69]1A). Hierarchical cluster analysis further stratifies TNBC patients
   into three distinct cohorts based on unique immune‐related protein
   expression profiles (Figure [70]1B). Significantly, elevated B7‐H4
   expression identified via hierarchical clustering (Figure [71]1B),
   volcano plots (Figures [72]1C,D), and density plots (Figures [73]S1A,B,
   Supporting Information) suggests a potential mechanism for immune
   evasion specifically leveraged by a subset of TNBCs, which may
   contribute to the observed reduced efficacy of immunotherapies and
   further hamper the effectiveness of cytotoxic agents’ combinational
   strategies in patients with high B7‐H4 expression. This relationship is
   further substantiated by the informatic analysis in Figure [74]S1C,
   Supporting Information, which indicates a negative correlation between
   elevated B7‐H4 levels and the induction of endoplasmic reticulum
   stress, preventing the release of DAMPs in TNBC patients.

Figure 1.

   Figure 1
   [75]Open in a new tab

   Accumulation of B7‐H4 is strongly associated with an unfavorable tumor
   immune response and prognosis in immune‐cold breast cancers. A)
   Proteomic analysis of 59 immune‐related proteins in TCGA samples of
   PAM50‐defined intrinsic and hormone receptor subtypes, including 19
   TNBC, 13 luminal A, 17 luminal B, and 10 Her2+ breast tumors. The genes
   (rows) are sorted by according to protein expression difference between
   the average level in each subtype and the average level in the other
   cancer types. B) Heatmap and hierarchical clustering is based on the
   expression levels of the 44 immune related proteins differentially
   expressed in TNBC breast cancer types. Each column represents a sample;
   each row represents a protein. The log2 relative protein expression
   scale is depicted on the top left. C,D) Volcano plots showing the
   immune‐related protein expression changes in three clustered TNBC
   patient cohorts. Each circle represents one protein. The log2 fold
   change is represented on the x‐axis. The y‐axis shows the ‐log10 of the
   p value. A p value of 0.05 and a fold change of 1 are indicated by red
   and green lines. E,F) Heatmap and hierarchical clustering is based on
   the expression levels of TROP2 and other 33 immune modulating
   proteins‐Pearson correlation between TROP2 expression and immune
   related proteins were calculated (E). The correlation between TROP2 and
   B7‐H4 is depicted (F). G) Expression of TROP2 and B7‐H4 in various TNBC
   cells detected by immunoblotting. H) HCC1937, MDA‐MB‐468, HCC1599 and
   MDA‐MB‐231 were treated with 0.5 µg mL^−1 SG for 48 h, clonogenic
   survival were determined. I) MDA‐MB‐468 cells were cocultured with
   human PBMCs at Effector (E) to target (T) ratio (3:1) and treated with
   SG for 96 h, cell proliferation rate was measured. J,K) Breast cancer
   specimens TMA stained for B7‐H4 and CD8 (J) and quantified staining
   intensity correlation between B7‐H4 and CD8 (K). ***p < 0.001, and
   ****p < 0.0001. Data (mean ± SEM) are representative of at least three
   independent experiments.

   We also investigated whether increased B7‐H4 protein levels affect the
   cytotoxic efficacy of ADCs such as SG, which depends on a supportive
   immune environment for optimal and sustained therapeutic effect. SG is
   a novel antibody‐drug conjugate composed of SN‐38, a potent
   topoisomerase I inhibitor, linked to an anti‐TROP‐2 monoclonal
   antibody. Our analysis found that TNBC patients in CPTAC cohorts with
   high TROP2 expression exhibited a significant positive correlation with
   B7‐H4 expression (Figure [76]1E,F), indicating that patients eligible
   for SG treatment often have an immunosuppressive environment.
   Additionally, we evaluated the baseline expression levels of TROP2 and
   B7‐H4 in various TNBC cell lines, including HCC1937 (high TROP2, low
   B7‐H4), MDA‐MB‐468 (high TROP2, high B7‐H4), HCC1599 (high TROP2, low
   B7‐H4), and MDA‐MB‐231 (low TROP2, low B7‐H4) (Figure [77]1G). Based on
   clonogenic survival assays, we observed that among the cell lines with
   high TROP2, MDA‐MB‐468 cells with high B7‐H4 expression levels showed
   markedly reduced sensitivity to SG treatment in comparison with HCC1599
   cells and HCC1937 cells with low B7‐H4 (Figure [78]1H, Figure [79]S1D,
   Supporting Information). To investigate whether a B7‐H4‐enriched
   immunosuppressive environment contributes to resistance against SG
   efficacy, we treated MDA‐MB‐468 and HCC1937 cells co‐cultured with
   peripheral blood mononuclear cells (PBMCs) with varying doses of SG.
   The results in Figure [80]1I and Figure [81]S1E, Supporting Information
   indicated that, while PBMC co‐culture did not enhance SG‐mediated
   cytotoxicity in B7‐H4‐high MDA‐MB‐468 cells (Figure [82]1I), it
   significantly potentiated tumor cell eradication in HCC1937 cells
   (Figure [83]S1E, Supporting Information). To confirm the observations
   in Figure [84]1A–I and determine if the increased expression of B7‐H4
   can function as a co‐inhibitory ligand to inhibit T‐cell function,
   which in turn facilitates tumor evasion and lowers immunogenic cell
   death, we performed the T cell infiltration correlation analysis and
   ssGSEA pathway analysis. Figure [85]S1F,G, Supporting Information
   demonstrated that increased B7‐H4 expression is associated with
   decreased CD8+ and CD4+ T cell infiltration in TNBCs. Additionally, the
   ssGSEA MsigDB‐based pathway analysis confirmed that high B7‐H4
   expression correlates positively with decreased T cell‐mediated
   immunity, cytokine production, and immune effector processes (Figure
   [86]S1H–P, Supporting Information). Next, we analyzed a tissue
   microarray containing 110 unique breast tissue samples via
   immunohistochemistry. In most TNBC samples, we observed a significant
   increase in B7‐H4 protein expression coinciding with a reduction in the
   CD8+ T cell population (Figure [87]1J,K), further confirming the effect
   of high B7‐H4 levels in enhancing immune evasion. In addition, high
   B7‐H4 expression is linked with lower overall survival as compared to
   other BC cancer patients using TCGA database analysis (Figure [88]S1Q,
   Supporting Information).^[ [89]^22 ^] These results collectively
   establish a significant association between elevated B7‐H4 expression
   and poorer clinical outcomes in TNBC patients potentially receiving SG
   therapy, highlighting the importance of B7‐H4 as a potential
   therapeutic target to enhance the immune system's effectiveness with
   ADC treatment.

2.2. Identification of USP10 as a Deubiquitinase that Stabilizes B7‐H4
through Counteracting B7‐H4 Ubiquitylation in Immune‐Evasive Triple Negative
Breast Cancers (TNBCs)

   To identify the mechanism that governs B7‐H4 protein turnover and
   stability in immune‐cold TNBCs, we employed Tandem Affinity
   Purification (TAP) and mass spectrometry to analyze the B7‐H4
   interactome in MDA‐MB‐468 cells stably expressing FLAG‐B7‐H4. The
   Flag‐tagged B7‐H4 protein was overexpressed, affinity‐purified using an
   empty vector as a control and subjected to mass spectrometry.^[ [90]^3
   ^] This led to identification of USP10 as a key regulator potentially
   responsible for the elevated levels of B7‐H4 in TNBC. Concurrently,
   AMFR was identified as the ubiquitin E3 ligase that regulates B7‐H4 for
   ubiquitylation and subsequent proteasomal degradation (Figure [91]2A).
   To characterize the biochemical implications of the USP10/B7‐H4
   interaction and its potential physiological significance, we conducted
   a series of co‐immunoprecipitation assays in MDA‐MB‐468 cells. Results,
   as shown in Figure [92]2B,C, demonstrated a co‐interaction wherein
   endogenous USP10 was co‐immunoprecipitated with FLAG‐tagged B7‐H4, and
   conversely, endogenous B7‐H4 was present in the c‐Myc‐tagged USP10
   immune complexes. To determine the subcellular localization of this
   interaction, the specific intracellular interaction regions between
   B7‐H4 and USP10 were further elucidated using immunofluorescence
   staining (Figure [93]2D) and validated by a Duolink proximity ligation
   assay (Figure [94]2E). The results confirmed that the B7‐H4 and USP10
   interaction occurs within the cytosolic compartment. To examine the
   direct role of USP10 on B7‐H4 turnover via deubiquitinase activity,
   MDA‐MB‐468 human breast cancer cells were genetically modified to
   stably overexpress or knockdown USP10. Figure [95]2F illustrates that
   enhanced USP10 levels corresponded with an increase in B7‐H4 protein
   levels. Similar increases were also observed in additional cell lines
   such as SKBR3 and HEK‐293T (Figure [96]S2A,B, Supporting Information).
   Conversely, CRISPR/Cas9‐mediated knockdown of USP10 resulted in a
   reduction in B7‐H4 protein expression but not mRNA level across both
   MDA‐MB‐468 cells (Figure [97]2G–J) and SKBR3 breast cancer cells
   (Figure [98]S2C, Supporting Information), further confirming the
   regulatory role of USP10 in counteracting B7‐H4 ubiquitylation
   resulting in B7‐H4 stabilization. To investigate whether USP10 mediates
   the cytosolic deubiquitylation of B7‐H4 and its subsequent
   translocation to the plasma membrane, we utilized flow cytometry with a
   fluorophore conjugated anti‐B7‐H4 antibody to measure B7‐H4 presence on
   the surface of USP10 knockdown (KD) MDA‐MB‐468 cells. The fluorescence
   intensity of cell membrane‐bound B7‐H4 was notably diminished in
   USP10‐deficient MDA‐MB‐468 cells (Figure [99]2K) compared to the
   control, with analogous observations in SKBR3 cells (Figure [100]S2D,
   Supporting Information). Moreover, treatment with spautin‐1, an
   established USP10 inhibitor, resulted in a decrease in B7‐H4 protein
   levels (Figure [101]2L and Figure [102]S2E, Supporting Information) and
   increased B7‐H4 turnover (Figure [103]2M), underscoring the potential
   of pharmacological inhibition of USP10 as a strategy to modulate B7‐H4
   levels. Given that spautin‐1 was initially identified as an inhibitor
   of both USP10 and USP13 with similar IC50 values,^[ [104]^23 ^] we
   evaluated whether spautin‐1‐induced B7‐H4 downregulation could also
   occur through USP13 inhibition. Treatment of MDA‐MB‐468 cells with
   spautin‐1 for 48 h significantly reduced USP10 and B7‐H4 protein levels
   but did not affect USP13 protein levels (Figure [105]S2F–I, Supporting
   Information). Additionally, USP13 KD in MDA‐MB‐468 cells did not alter
   B7‐H4 protein expression (Figure [106]S2J–M, Supporting Information).
   To further validate our findings, we used Wu‐5, another specific USP10
   inhibitor,^[ [107]^24 ^] which also resulted in a noticeable reduction
   in USP10 and B7‐H4 (Figure [108]S2N–Q, Supporting Information). These
   results collectively underscore the specific role of USP10 in
   regulating B7‐H4 expression. Subsequent MsigDB database pathway
   analysis also observed the elevation of B7‐H4 expression in TNBC
   cohorts, which appears to be closely associated with an upregulated
   deubiquitylation pathway (Figure [109]S3A,B, Supporting Information).
   Notably, this analysis paralleled findings in Figure [110]1B,
   identifying a subgroup within TNBC with concurrent upregulation of
   USP10 and B7‐H4 protein levels (Figure [111]S3C, Supporting
   Information) and reinforcing the potential interplay between USP10
   activity and B7‐H4 stability in this breast cancer subset.

Figure 2.

   Figure 2
   [112]Open in a new tab

   Identification of USP10 as a binding partner of B7‐H4 that upregulates
   B7‐H4 protein levels in immune‐cold TNBC breast tumors. A) Coomassie
   blue staining of B7‐H4 immunoprecipitated from MDA‐MB‐468 cells
   expressing FLAG‐tagged B7‐H4 reveals USP10 as a binding partner by mass
   spectrometry analysis. B,C) Confirmation of the B7‐H4/USP10 interaction
   via co‐immunoprecipitation by coimmunoprecipitation of endogenous USP10
   (B) and coimmunoprecipitation of endogenous B7‐H4 (C). D)
   Immunofluorescence staining of MDA‐MB‐468 cells expressing FLAG‐B7‐H4
   shows colocalization with USP10. Scale bar = 30 µm. E)
   MDA‐MB‐468‐FLAG‐B7‐H4 cells were subjected to Duolink in Situ PLA assay
   with FLAG mouse antibody and USP10 rabbit antibody. Red dots indicate
   the interaction of the two proteins. Scale bar = 30 µm. F)
   Overexpression of USP10 in MDA‐MB‐468 cells enhances B7‐H4 protein
   levels, evidenced by immunoblotting (left panel) and subsequent
   quantification (right panel). G–J) USP10 knockdown results in
   downregulation of B7‐H4 protein but not mRNA in MDA‐MB‐468 cells. Three
   USP10 knockdown cell lines were established with different sgRNAs with
   CRISPR/Cas9 system. Empty vector CRISPRv2 used as the control. (G)
   Representative images of immunoblotting of USP10 and B7‐H4. (H)
   Quantification of USP10 and (I) B7‐H4 protein normalized with β‐actin
   amount in biological triplicates. (J) Quantification of B7‐H4 mRNA
   levels with qPCR. K) Decreased membrane and total B7‐H4 in USP10
   knockdown cells. MDA‐MB‐468 cells were stained with PE conjugated
   anti‐human B7‐H4 antibody without or with permeabilization for surface
   B7‐H4 and total B7‐H4, respectively and subjected to flow cytometry
   analysis. Representative images are shown. L) Inhibition of USP10 with
   spautin‐1 leads to decreased B7‐H4 protein in MDA‐MB‐468 cells,
   validated by western blot (left panel) and quantified intensity (right
   panel). M) Spautin‐1 treatment increases B7‐H4 turnover. MDA‐MB‐468
   cells were treated with DMSO or 10 µM spautin‐1 for 24 h and then
   treated with 100 µg mL^−1 cycloheximide (CHX) at the indicated time
   points. Representative images of immunoblotting of B7‐H4 (left panel).
   Quantification of B7‐H4 (40 kDa) intensity normalized to time 0 (right
   panel). *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Data
   (mean ± SEM) are representative of at least three independent
   experiments.

   To determine if USP10 directly affects B7‐H4 turnover dynamics, we
   conducted a pulse‐chase assay in MDA‐MB‐468 cells with
   CRISPR/Cas9‐mediated USP10 knockdown. Our results showed that USP10
   depletion significantly accelerated B7‐H4 protein turnover (Figure
   [113]3A,B), reducing the B7‐H4 protein half‐life from 1.25 to 0.70–0.85
   h. To examine the role of USP10 on B7‐H4 degradation through catalyzing
   B7‐H4 deubiquitylation, ubiquitylation assays were conducted in
   HEK293T‐FLAG‐B7‐H4 cells with enforced USP10 expression.
   Immunoprecipitation of endogenous B7‐H4 followed by subsequent
   immunoblotting for ubiquitin revealed that USP10 overexpression
   attenuates B7‐H4 ubiquitylation (Figure [114]3C). Conversely, knocking
   down USP10 or treatment of spautin‐1 resulted in an appreciable
   increase in B7‐H4 ubiquitylation, reinforcing the role of USP10 in the
   preservation of B7‐H4 protein levels (Figure [115]3D and Figure
   [116]S3D, Supporting Information). To further confirm the specific role
   of USP10 in regulating B7‐H4 turnover and to determine if other USPs
   may exhibit similar regulatory effects in a TNBC setting, we aligned
   the protein sequences of all human USPs and identified those with high
   sequence similarity to USP10. Based on the phylogenetic analysis, we
   identified 10 USPs closely related to USP10: USP36, USP17, USP42,
   USP30, USP49, USP44, USP3, USP28, USP25 (Figure [117]S4A, Supporting
   Information). We designed sgRNAs to knock down these USPs and examined
   B7‐H4 expression in breast cancer cells. None of these USPs
   significantly reduced B7‐H4 levels (Figure [118]S4B,C, Supporting
   Information). Additionally, results from a cycloheximide (CHX)‐based
   chase assay demonstrated that other USPs do not affect B7‐H4 protein
   turnover (Figure [119]S4D,E, Supporting Information). To investigate
   the potential role of USP10 in facilitating immune escape by cancer
   cells, we utilized a co‐culture system with cancer cells and human
   PBMCs, activating cytotoxic T‐cells with anti‐human CD3 and CD28.2
   antibodies. Following co‐culture, cancer cells were assessed for
   apoptosis using Annexin V/Propidium Iodide staining and analyzed via
   flow cytometry. Results revealed a significant increase in apoptotic
   cell count in USP10 knockdown populations compared to controls,
   indicating that loss of USP10 sensitizes cancer cells to T‐cell
   mediated cytotoxicity (Figure [120]3E–G). In contrast, USP10 knockdown
   did not induce apoptosis in the absence of PBMCs (Figure [121]S4F–H,
   Supporting Information). Furthermore, the depletion of USP10 was
   associated with enhanced cytotoxic T‐cell activity, as evidenced by
   increased Ki67 expression, indicating higher T‐cell proliferation, and
   elevated TNFα+ CD8+ T cell population (Figure [122]3H,I). Conversely,
   overexpression of USP10 in MDA‐MB‐468 cells led to a marked reduction
   of apoptosis accompanied with decreased TNFα+ and IFN‐γ+ CD8+ T cell
   population (Figure [123]3J–M and Figure [124]S4I–K, Supporting
   Information). These findings collectively suggest that the
   deubiquitinase USP10 is instrumental in maintaining high levels of
   B7‐H4, which in turn promotes an immune‐suppressive microenvironment by
   impairing T‐cell efficacy in immune‐cold TNBC settings.

Figure 3.

   Figure 3
   [125]Open in a new tab

   USP10 is a pivotal regulator to stabilize B7‐H4 via deubiquitylation
   and modulate T cell function. A,B) USP10 knockdown leads to increased
   B7‐H4 turnover in MDA‐MB‐468. A) Pulse‐chase analysis for the control
   CRISPR v2 and USP10 knockdown MDA‐MB‐468 cells with 100 µg mL^−1
   cycloheximide. B) Quantification of the amount of B7‐H4 (40 kDa)
   normalized to the initial time point. The half‐life t[1/2] is estimated
   based on the plot. C) USP10 overexpression reduces B7‐H4 ubiquitylation
   in HEK‐293T stable cell line expressing FLAG‐B7‐H4. FLAG‐B7‐H4 was
   immunoprecipitated followed by immunoblotting with anti‐ubiquitin. D)
   USP10 knockdown enhances B7‐H4 ubiquitylation in MDA‐MB‐468 cells.
   USP10 was knocked down in MDA‐MB‐468 cells stably expressing HA‐tagged
   B7‐H4. Following immunoprecipitation of HA‐B7‐H4, ubiquitylation levels
   were assessed via immunoblotting with an anti‐ubiquitin antibody. E–G)
   Flow cytometry analysis of MDA‐MB‐468 cells cocultured with PBMCs shows
   increased apoptosis in USP10 knockdown cells (E). Quantitative analysis
   of the percentage of total apoptosis (Annexin V+) (F) and late
   apoptosis (Annexin V+ and PI+) cancer cells (G). H,I) USP10 knockdown
   or B7‐H4 knockdown in MDA‐MB‐468 releases the suppression of the
   function of cytotoxic T cells by cancer cells as determined by the
   percentage of TNFα positive cytotoxic T cells (H) and the percentage of
   Ki67+ cytotoxic T cells (I). J,K) Flow cytometry analysis of MDA‐MB‐468
   cells cocultured with PBMCs shows decreased apoptosis in USP10
   overexpression cells (J). Quantitative analysis of the percentage of
   late apoptosis (Annexin V+ and PI+) cancer cells (K). L,M) USP10
   overexpression in MDA‐MB‐468 further suppressed the function of
   cytotoxic T cells by cancer cells as determined by the percentage of
   TNFα+cytotoxic T cells (L) and the percentage of IFN‐γ+ cytotoxic T
   cells (M). *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 by
   the one‐way ANOVA test or two‐way ANOVA test, Data (mean ± SEM) are
   representative of at least three independent experiments.

2.3. Mapping and Structural Modeling of the Interaction between USP10 and
B7‐H4

   To uncover the mechanism by which B7‐H4 is recognized by USP10 allowing
   the removal of its attached ubiquitin‐conjugates, we engineered a set
   of B7‐H4 deletion fragments with Flag tag and fragments of USP10 with
   c‐Myc tag, which were then co‐expressed in HEK‐293T cells for
   subsequent co‐immunoprecipitation (co‐IP) analyses. We crafted and
   expressed a suite of c‐Myc‐tagged USP10 fragments (Figure [126]4A)
   alongside Flag‐tagged B7‐H4 in HEK‐293T cells. The resulting co‐IP
   assays pinpointed the USP domain stretching from amino residue 415–798
   at the C‐terminus of USP10 as integral for binding B7‐H4
   (Figure [127]4B,C). To further identify the exact site responsible for
   USP10‐B7‐H4 interaction, a series of USP domain deletion mutants were
   engineered, subsequently co‐IP assay demonstrated that the mid‐segment
   of the USP domain (amino acid 516–615) emerged as key in mediating this
   interaction, as its removal negated the binding (Figure [128]4D,E). To
   identify the interacting domain on B7‐H4, different B7‐H4 fragments
   were co‐transfected with Flag‐tagged USP10 in HEK‐293T cells. The co‐IP
   assay indicated the Ig‐like V‐type domain of B7‐H4 as the primary site
   for USP10 interaction (Figure [129]4F,G), with amino acid residues
   24–153 of B7‐H4 being essential for their interplay. We have recently
   explored and characterized the role of AMFR acting as an E3 ligase
   responsible for the ubiquitin‐proteasomal degradation of B7‐H4. To
   comprehensively address the mechanism by which the interplay of
   AMFR‐mediated ubiquitylation and USP10‐guided deubiquitylation in
   regulating B7‐H4 proteolysis, we conducted structural modeling of the
   interaction between B7‐H4, AMFR, and USP10 by integrated docking
   simulations. We generated in silico structural models for the complexes
   B7‐H4‐AMFR and B7‐H4‐USP10, consistent with experimental data for the
   interfaces formed between B7‐H4 and the two respective enzymes AMFR and
   USP10 (Figure [130]4H). Subsequently, we engaged in comprehensive
   molecular docking simulations to generate molecular models of the
   interaction between amino acids 516–615 on USP10 and B7‐H4's Ig‐like
   V‐type domain. Docking simulations have explored the potential binding
   sites and energetics of the B7‐H4‐USP10 interaction with lowest energy
   −654.5 kJ mol^−1, and revealed amino acid residues (Lue122, Glu144
   Lys146 and Thr147) of B7‐H4 (yellow) that forming covalent interaction
   with USP10 (green) residue (Lys518, Ser516, Glu522, Tyr527). The
   interface of B7‐H4‐USP10 is further stabilized by other hydrophobic
   interactions between (His35, Val39, Ala43, Asn142) and (Ile531, Pro601,
   Ile602, Phe606), respectively (Figure [131]4I).

Figure 4.

   Figure 4
   [132]Open in a new tab

   Molecular mapping the domains on USP10 and B4‐H4 that facilitates the
   interaction between USP10 and B7‐H4. A–E) Schematic diagram of human
   USP10 domains and strategy to engineer a series of USP10 fragments (A)
   and USP10 deletion mutants (D). Mapping of the molecular domain on
   USP10 involving in the interaction with B7‐H4. The interactions between
   B7‐H4 and USP10 fragments were examined by co‐IP experiments in
   HEK‐293T cells. Amino acid stretches from 516 to 615 amino acid on
   USP10 (B‐E) were identified as the region that mediates its interaction
   with B7‐H4. F) Schematic diagram of human B7‐H4 domains and strategy to
   engineer a series of B7‐H4 deletion fragments. G) The interactions
   between USP10 and B7‐H4 fragments were examined by co‐IP experiments in
   HEK‐293T cells. Amino acid stretches from 24 to 146 (including Ig like
   V‐type domain) on B7‐H4 were identified as the region that mediates the
   interaction between USP10 and B7‐H4. H) Molecular docking model of the
   interaction between USP10‐B7‐H4‐AMFR complex. AMFR is the E3 ligase for
   B7‐H4. The ring motif of AMFR is in green; the Ig‐like V‐type and
   Ig‐like‐C2 type domain of B7‐H4 is in red and 516–615 residues of USP10
   is in blue. The docking model shows that USP10 binds to the Ig‐like
   V‐type domain of B7‐H4 and the deubiquitinase AMFR binds to the
   Ig‐like‐C2 type domain of B7‐H4, where the interacting residues are
   highlighted in red and green line. I) The Structural model for the
   interaction of USP10 (sky blue) with B7‐H4 (red) investigated through
   docking studies (left panel). Docking simulations have explored the
   potential binding sites and energetics of the B7‐H4‐USP10 interaction
   with lowest energy −654.5 kJ mol^−1, and also revealed amino acid
   residues (Lue122, Glu144 Lys146 and Thr147) of B7‐H4 (yellow) that
   forming conventional hydrogen bond interaction with USP10 (green)
   residue (Lys518, Ser516, Glu522, Tyr527) and further stabilized by
   other hydrophobic interactions between (Ile531, Pro601, Ile602, Phe606)
   and (His35, Val39, Ala43, Asn142), respectively. The zoomed view of
   interface indicating the close interaction between B7‐H4 (stick yellow)
   with USP10 (stick green) shows in (right panel).

   To determine the critical amino acid residues essential for the
   interaction between USP10 and B7‐H4, we utilized in silico mutation
   scans to assess the impact of specific mutations on this protein
   interaction. Our findings indicate that mutations of B7‐H4 residues
   Glu144 and Lys146 to Arg144 and Asp146, respectively, which yielded the
   lowest binding energy of −793.8 kJ mol^−1, significantly disrupted the
   interaction interface with USP10, highlighting their crucial role in
   facilitating their binding (Figure [133]5A). To experimentally validate
   the direct involvement of the Glu144 (E144) and Lys146 (K146) residues
   of B7‐H4 in mediating its association with USP10, we engineered mutant
   FLAG‐B7‐H4^E144R/K146D constructs and co‐transfected them with
   c‐Myc‐USP10 into HEK‐293T cells. Following co‐immunoprecipitation
   assays, the results illustrated in Figure [134]5B suggest a substantial
   reduction of the interaction between the mutant B7‐H4 and USP10. This
   alteration was also associated with increased ubiquitin linkage
   (Figure [135]5C) and accelerated degradation of the mutated B7‐H4
   protein (Figure [136]5D). Moreover, in the context of a cancer cell and
   PBMC coculture system, MDA‐MB‐468, SKBR3, and HEK‐293T cells expressing
   the B7‐H4^E144R/K146D mutant demonstrated a differential apoptotic
   response. Compared to cells with vector control, the mutant cells
   showed reduced apoptosis, while compared to cells expressing B7‐H4^WT,
   they exhibited increased apoptosis (Figure [137]5E–H and Figure
   [138]S5A,B, Supporting Information). The above observation suggests
   that the B7‐H4^E144R/K146D mutant, despite retaining some immune
   suppressive function, is compromised in its efficacy due to a
   deficiency in deubiquitylation. These results collectively underscore
   the critical role of specific amino acid residues in the functional
   interaction between B7‐H4 and USP10, with implications for the
   stability and immune regulatory function of B7‐H4.

Figure 5.

   Figure 5
   [139]Open in a new tab

   Structural modeling identified the critical role of E144/K146 on B7‐H4
   in mediating USP10‐mediated stabilization of B7‐H4. A) The structure
   model of USP10‐B7‐H4^E144R/K146D. Mutation of B7‐H4 (red) residue
   Glu144 and Lys146 to Arg144 and Asp146 highlighted as sticks and
   colored in yellow. The double mutation disrupts the interaction: Glu144
   (yellow)‐Tyr527(green) interaction, and Lys146(yellow)‐Glu522(green) in
   the previous interface with USP10 (sky blue) and shows long distance
   (63.6 and 64.1 Å) to each other. USP10 binds to a region other than the
   critical residues of B7‐H4 (left). Zoom in view of the mutations in the
   B7‐H4 and its hydrophobic interactions with USP10. B,C) Validation of
   interaction between USP10 with B7‐H4 and B7‐H4^E144R/K146D by
   coimmunoprecipitation of ectopic Flag‐tagged B7‐H4^WT and Flag‐tagged
   B7‐H4^E144R/K146D in HEK‐293T cells. The ubiquitylation status of
   B7‐H4^WT and B7‐H4^E144R/K146D were also determined using an antibody
   against ubiquitin. D) MDA‐MB‐468‐B7‐H4^WT and
   MDA‐MB‐468‐B7‐H4^E144R/K146D cells were treated with 100 µg mL^−1 CHX.
   Cell lysates were collected at indicated time points and followed by
   measuring B7‐H4 protein turnover. Quantification of B7‐H4 turnover were
   illustrated in the bottom panel. E–H) Annexin V/Propidium Iodide
   apoptosis assay shows that in comparison with WT, E121R&K123D mutation
   leads to increased cancer cell apoptosis in the coculture with human
   PBMC. MDA‐MB‐468 (E.F) and SKBR3 (G,H) cells expressing either B7‐H4^WT
   and B7‐H4^E144R/K146D mutant were cocultured with activated human
   PBMCs. Representative images of flow cytometry (E, G). Quantitative
   data show the percentage of cells in late apoptosis (Annexin V+ and
   PI+) stage (F, H). *p < 0.05 and ****p < 0.0001. Data (mean ± SEM) are
   representative of at least three independent experiments.

2.4. Unexpected Elevation of USP10/B7‐H4 Impedes the Effectiveness of
Sacituzumab Govitecan through its Immunosuppressive Role in Immune‐Cold
Breast Tumors

   Tumor‐intrinsic genomic instability can be further enhanced by TOP1
   inhibitors, such as SN‐38, the payload of SG, to instigate immune
   responses via T cell activation. Our previous study has shown that
   increased B7‐H4 expression block chemo agents induced immunogenic cell
   death.^[ [140]^3 ^] In addition, pathway analysis from MsigDB suggests
   a significant association between USP10 expression and the DNA damage
   response,^[ [141]^25 , [142]^26 ^] as depicted in Figure [143]S6A,
   Supporting Information. This observation raises the question of whether
   DNA damage stimulus, like the payload of SG, impacts the
   deubiquitylation of B7‐H4 orchestrated by USP10. Pathway correlation
   analysis (Figure [144]6A) has revealed that robust DNA repair activity
   is often accompanied by a negative adaptive immune response and reduced
   T cell cytotoxicity. Moreover, gene correlation analysis has
   underscored a strong association between DNA topoisomerases and USP10
   levels (Figure [145]6B and Figure [146]S6B,C, Supporting Information),
   reinforcing the possible interaction between DNA damage repair
   mechanisms and immune evasion strategies. In our experimental setup,
   MDA‐MB‐468 breast cancer cells exposed to SG in addition to other DNA
   damaging agents showed increased levels of both USP10 and B7‐H4
   (Figure [147]6C and Figure [148]S6D, Supporting Information),
   suggesting that SG and other DNA‐damaging agents, including those used
   in systemic chemotherapy, may contribute to an immune‐suppressive
   milieu governed by B7‐H4. Further investigation into the effects of SG
   on USP10‐mediated deubiquitylation showed that SG treatment
   significantly reduced B7‐H4 ubiquitylation (Figure [149]S6E, Supporting
   Information) and stabilized B7‐H4 protein expression (Figure [150]6D).
   To further elucidate the role of USP10 in mediating SG‐induced B7‐H4
   upregulation, we treated MDA‐MB‐468 cells with a stable knockdown of
   USP10 with 100 ng mL^−1 SG. As shown in Figure [151]S6F–H, Supporting
   Information and Figure [152]6D, depletion of USP10 led to unchanged
   B7‐H4 levels and protein turnover post‐SG treatment, highlighting the
   critical role of SG in maintaining B7‐H4 stability through USP10. To
   investigate the mechanism through which SG treatment modulates USP10
   expression and subsequently impacts B7‐H4 levels, we assessed both
   USP10 mRNA levels and protein turnover rate. As shown in Figure
   [153]S6I,J, Supporting Information, while the SG treatment did not
   alter the USP10 turnover rate, we observed a significant increase in
   mRNA levels, suggesting that SG induces USP10 upregulation
   predominantly through transcriptional activation.

Figure 6.

   Figure 6
   [154]Open in a new tab

   Stabilization of B7‐H4 by USP10 impedes the effectiveness of
   Sacituzumab govitecan through its immunosuppressive role in immune‐cold
   breast tumors. A) The MSigDB‐based pathway correlation analysis using
   [155]GSE72362 database shows that tumor cell related adaptive immune
   response pathways are negatively correlated with SG payload induced DNA
   damage response. B) Spearman's rank gene correlation analysis using
   [156]GSE72362 database shows that USP10 expression is highly positively
   correlated with DNA topoisomerase expression. C) MDA‐MB‐468 cells were
   treated with 0.5 µg mL^−1 SG for 24 h. USP10 and B7‐H4 protein levels
   were determined by immunoblotting (left panel). Quantification of USP10
   and B7‐H4 protein normalized with β‐actin amount in biological
   triplicates (right panel). D) MDA‐MB‐468 cells either with or without
   USP10 knockdown were treated with either PBS (control) or 100 ng mL^−1
   SG for 24 h, followed by treatment with 100 µg mL^−1 cycloheximide
   (CHX) at specified time points (hours). The left panel shows
   representative immunoblotting images for B7‐H4. The right panel
   provides the quantification of B7‐H4 (40 kDa) intensity at 2 h
   normalized to the level at time 0. E) MDA‐MB‐468‐B7‐H4 KD and
   MDA‐MB‐468‐USP10 KD cells were cocultured with human PBMCs at Effector
   (E) to Target (T) ratio (3:1) and treated with SG (0.1 µg mL^−1), cell
   proliferation rate was measured by the cell confluence normalized to
   the corresponding vehicle control of each cell line (%). F–J)
   MDA‐MB‐468‐B7‐H4 KD and MDA‐MB‐468‐USP10 KD cells were treated with SG
   (0.5 µg mL^−1), surface expression of damage‐associated molecular
   patterns (DAMPs) markers: HSP70, HSP90 and Calreticulin (CRT) was
   measured by flow cytometry. (F, H) representative flow cytometry
   images. G, I, J) Quantification of the percentage of positive cells.
   K,L) The control CRISPR v2 and USP10 knockdown (KD) MDA‐MB‐468 cells
   were pretreated with SG (100 ng mL^−1) for 24 h and then cocultured
   with purified human dendritic cells at a 1:1 ratio. The cancer cells
   and purified human dendritic cells were stained with cell tracker green
   CMFDA and red CMTPX, respectively, and subjected to fluorescence
   imaging after a 3‐h coculture. K) Representative microscopic images.
   Scale bar = 500 µm. L) Quantification of phagocytosis based on
   fluorescence microscopic images. The phagocytosis percentage is the
   number of dendritic cells with engulfed cancer cells, indicated by the
   colocalization of green and red signals, over the total number of
   dendritic cells per view. M,N) SG (100 ng mL^−1) pretreated control
   CRISPR v2 and USP10 knockdown (KD) MDA‐MB‐468 cells were stained with
   cell tracker green CMFDA, seeded on a plate, and cultured with human
   dendritic cells for 2 h. The dendritic cells were then collected and
   subjected to immunostaining with anti‐CD11c‐APC and analyzed by flow
   cytometry. M) Representative flow cytometry images. N) Quantification
   of the percentage of phagocytosis, indicated as the percentage of
   CMFDA+ dendritic cells. O,P) SG (100 ng mL^−1) pretreated control
   CRISPR v2 and USP10 knockdown (KD) MDA‐MB‐468 cells were cultured with
   the human dendritic cells for 2 h. Dendritic cells were collected and
   subjected to immunostaining with anti‐MHC‐I‐PE‐Dazzle‐594 and analyzed
   by flow cytometry. O) Representative flow cytometry image showing the
   distribution of surface MHC‐I expression levels in dendritic cells. P)
   Quantification of the median membrane staining of MHC‐I in dendritic
   cells cocultured with the corresponding cancer cells. *p < 0.05, **p <
   0.01, and ***p < 0.001, ****p < 0.0001 by the one‐way or two‐way ANOVA
   test. Data (mean ± SEM) are representative of at least three
   independent experiments.

   To assess the potential of targeting the USP10/B7‐H4 axis to sensitize
   breast cancer cells to SG, we conducted coculture assays with
   MDA‐MB‐468 and HCC70 where B7‐H4 or USP10 was knocked down, followed by
   SG treatment. The results, particularly illustrated in Figure [157]6E
   and Figure [158]S6K–N, Supporting Information, demonstrate that
   disruption of this axis significantly enhances the anti‐proliferative
   effects of SG by promoting apoptosis in cancer cells, even at a low
   effector‐to‐target (E: T) ratio of 3:1. Conversely, restoring B7‐H4
   expression in USP10 knockdown (KD) cells significantly mitigated the
   cytotoxic effects of SG, particularly in the presence of PBMC coculture
   (Figure [159]S7A–E, Supporting Information). This confirms the crucial
   role of the USP10‐B7‐H4 axis in modulating the tumor immune response to
   SG treatment. Furthermore, the increased lethality of B7‐H4 and USP10
   depletion tumor cells upon SG treatment appears to be linked to the
   induction of endoplasmic reticulum (ER) stress and the upregulation of
   DAMPs, such as calreticulin (CALR), heat shock proteins 70 and 90
   (HSP70 and HSP90), as evidenced by flow cytometry analysis
   (Figure [160]6F–J). To further characterize how downregulation of the
   USP10/B7‐H4 axis‐induced DAMPs potentiates immunogenic cell death via
   enhanced T cell activation, we investigated the role of the USP10/B7‐H4
   axis in dendritic cell phagocytosis and subsequent antigen
   presentation. Purified human dendritic cells were cocultured with
   either control or USP10 knockdown MDA‐MB‐468 cancer cells pretreated
   with 100 ng mL^−1 SG. Real‐time live cell imaging revealed an increased
   percentage of dendritic cell phagocytosis of cancer cells in the
   coculture with USP10 KD cells (Figure [161]6K,L). Consistently,
   quantitative analysis of dendritic cell phagocytosis via flow cytometry
   confirmed a similar increase in phagocytosis of USP10 KD MDA‐MB‐468
   cancer cells after SG treatment (Figure [162]6M,N). To assess whether
   this enhanced phagocytosis was accompanied by improved antigen
   presentation, potentially leading to T cell priming and activation, we
   evaluated the membrane expression of MHC‐I on dendritic cells. A
   significant elevation in MHC‐I membrane expression was detected after
   coculturing with USP10 KD MDA‐MB‐468 cancer cells compared to controls
   (Figure [163]6O,P). These observations suggest that elevated expression
   of B7‐H4 and USP10 may be a barrier to the effectiveness of SG
   treatment. Restoring B7‐H4 turnover by inhibiting USP10 could
   potentially convert a cold tumor immune environment to a more
   responsive one, thereby increasing SG effectiveness.

2.5. Suppression of USP10/B7‐H4 Proteolytic Axis Reinvigorates Therapeutic
Efficacy of ADCs through Tuning Immune‐Cold Breast Tumor Hot

   To investigate the influence of USP10 on tumor progression through the
   regulation of B7‐H4 in vivo, we engineered murine 4T1 breast cancer
   cell lines to alter USP10 expression. We established USP10
   overexpression (USP10‐OE) and knockdown (USP10‐KD) in 4T1 cells (Figure
   [164]S8A,B, Supporting Information), using cells with an empty vector
   as the control group. The control 4T1 cells and 4T1 cells
   overexpressing USP10 were then injected subcutaneously into the mammary
   fat pads of female BALB/c mice. Tumor volumes were periodically
   measured with calipers and calculated with the formula V = (W^2 × L)/2,
   where V represents the tumor volume, W is the width, and L is the
   length of the tumor. After 21 days post‐injection, mice injected with
   USP10‐OE cells exhibited a substantial increase in tumor growth rate
   and final tumor weight (Figure [165]7A–D). Additionally,
   immunohistochemical staining of tumor tissues demonstrated that
   upregulation of the USP10‐B7‐H4 axis dampened the tumor immune
   response, evidenced by a decreased CD8+ T cell population
   (Figure [166]7E). In contrast, mice injected with USP10‐KD cells
   displayed significantly inhibited tumor growth and reduced tumor weight
   compared to the control group (Figure [167]7F,G). To determine whether
   alteration of USP10 could directly affect tumor cell proliferation
   independently of immune system interactions, we conducted experiments
   using nude mice bearing 4T1‐USP10‐OE or 4T1‐USP10‐KD xenografts. Our
   results demonstrated that neither USP10 overexpression nor USP10
   knockdown exhibited significant differences in tumor growth in the 4T1
   xenograft mouse model (Figure [168]S8C–H, Supporting Information).

Figure 7.

   Figure 7
   [169]Open in a new tab

   Suppression of USP10/B7‐H4 proteolytic axis in vivo reinvigorates
   therapeutic efficacy of ADCs through regulating tumor immune response.
   A–D) 4T1 control and 4T1‐USP10 overexpression breast cancer cells were
   orthotopically injected into the right fourth mammary gland of the
   BALB/c wild‐type (WT) mice (A). Phosphate‐buffered saline (PBS) was
   used in the control group. Tumor growth curve was plotted (B) and tumor
   was harvested at 21 days after tumor challenge, weight was imaged and
   measured (C,D). E) Representative images of IHC staining of mouse
   USP10, mouse B7‐H4 and mouse CD8+ cells level of control and USP10 OE
   mouse tumors. F,G) 4T1 control and 4T1‐USP10 KD tumors were harvested
   21 days after tumor challenge and analyzed. Tumor growth curve was
   plotted (F) and tumor weight (G) was measured at the end point. H)
   Schematic diagram of 4T1‐hTROP2, where the basal mouse TROP2 was
   replaced with a human counterpart, and 4T1‐hTROP2 cells were
   orthotopically injected into the left fourth mammary fat pad and allow
   to grow ≈100 mm^3, followed by injection of spautin‐1 (20 mg kg^−1,
   i.p.) for two times per week, and SG (10 mg kg^−1, i.p.) for three
   times. PBS and IgG control were used in control groups. I–K) 4T1‐hTROP2
   and 4T1‐hTROP2 mB7‐H4 OE breast cancer cells were orthotopically
   injected into the left fourth mammary fat pad followed by injection of
   SG (10 mg kg^−1, i.p.) for three times. Tumor growth curve was plotted
   (I) and tumor was harvested after tumor challenge, weight was imaged
   (J) and measured (K). L,M) 4T1‐hTROP2 cells were orthotopically
   injected into the left fourth mammary fat pad and allow to grow ≈100
   mm^3, followed by injection of spautin‐1 (20 mg kg^−1, i.p.) for two
   times per week, and SG (10 mg kg^−1, i.p.) for three times. PBS and IgG
   control were used in control groups. The tumor growth was monitored
   twice per week. Tumor growth (L) and survival curve (M) of the mice
   were plotted. N) Representative images of IHC staining of mouse CD8,
   granzyme B and TOX level of 4T1‐hTrop2 mouse tumors after the treatment
   of SG treatment or SG + Spautin‐1 combination. Scale bar = 20 µm. O,P).
   USP10 expression level and B7‐H4 expression level were analyzed and
   quantified by IHC using TMA consisting of TNBC samples. Representative
   IHC images (K). Quantification of USP10 and B7‐H4 expression (L). *p <
   0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001 by the two‐way ANOVA
   test. Data (means ± SEM) are representative of at least two independent
   experiments with five to ten independently analyzed mice per group.

   To evaluate the impact of elevated B7‐H4 on the efficacy of SG, we
   overexpressed B7‐H4 in 4T1 cells engineered to express human TROP2
   (4T1‐hTROP2) (Figure [170]S8I, Supporting Information and
   Figure [171]7H–K). Results shown in Figure [172]7H–K indicate that 4T1
   cells with increased B7‐H4 levels exhibited significantly reduced
   sensitivity to SG, leading to diminished tumor growth inhibition
   compared to the parental 4T1‐hTROP2 cells. Similarly, 4T1‐hTROP2 cells
   with USP10 knockdown and subsequent reintroduction of B7‐H4 also showed
   notable increases in tumor growth and enhanced resistance to SG (Figure
   [173]S8J–M, Supporting Information), confirming the axis's critical
   role in TNBC progression and response to SG therapy. Furthermore, to
   explore the therapeutic potential of targeting the USP10/B7‐H4 axis in
   TNBC, we investigated the anti‐tumor effects of the USP10 inhibitor
   spautin‐1, both alone and in combination with the Trop‐2 directed
   antibody‐drug conjugate, SG, in a TNBC 4T1 model engineered to express
   human TROP2. The findings, presented in Figure [174]7L,M, demonstrate
   that the combination of spautin‐1 with SG significantly inhibited tumor
   growth and extended the median survival from 43 to 55 days compared to
   each agent alone. Additionally, at the end of the 60‐day experimental
   monitoring phase, there were still 5 mice alive in the combination
   group. To evaluate the impact of inhibiting the USP10‐B7H4 axis on T
   cell function, we analyzed tissue samples from 4T1‐hTROP2 TNBC tumors.
   The findings, presented in Figure [175]7N, showed that the combination
   of SG and spautin‐1 markedly increased CD8+ T cell infiltration and
   activation. Additionally, we observed a moderate decrease in T cell
   exhaustion, as evidenced by reduced Tox staining, when compared to
   treatment with SG alone. These outcomes suggest the therapeutic
   potential of targeting the USP10‐B7‐H4 axis pharmacologically to
   enhance immunogenic cell death in response to the TOP1 inhibitor, the
   payload of SG, particularly in TNBC subtypes characterized by high
   B7‐H4 expression and low immunogenicity.

   To determine the clinical relevance of B7‐H4 proteolytic regulation on
   immune response and prognosis in TNBC, we conducted a differential gene
   expression analysis comparing tumor tissues with adjacent normal
   tissues. This analysis revealed a significant upregulation of USP10 in
   breast cancer cases (Figure [176]S9A, Supporting Information). We
   stratified TNBC patients into TROP2‐high and TROP2‐low groups and
   observed a positive correlation between high TROP2 expression and
   increased USP10 expression (Figure [177]S9B, Supporting Information).
   This suggests that USP10 may play a pivotal role in mediating the
   increase in B7‐H4 expression in the context of high TROP2 levels and SG
   treatment. Additionally, immunohistochemistry (IHC) analysis on a
   tissue microarray comprising 110 distinct breast tissue specimens
   further confirmed a positive correlation between USP10 and B7‐H4
   expression levels (Figure [178]7O,P). Furthermore, immune profiling
   using CIBERSORT, as shown in Figure [179]S9C–E, Supporting Information,
   indicated that tumors with increased USP10 expression exhibited a
   reduced CD8+ T cell population. This was associated with lower overall
   lymphocyte infiltration and heightened tumor proliferation rates.
   Moreover, an analysis of TNBC patient data from the TCGA database
   demonstrated that high USP10 expression was associated with worse
   survival outcomes, even in cases with a high presence of cytotoxic T
   lymphocytes (Figure [180]S9F–I, Supporting Information).

   Overall, our results demonstrate that USP10 plays a critical role in
   regulating B7‐H4, which, in turn orchestrates tumor immune evasion by
   stabilizing protein expression through deubiquitylation. Suppression of
   USP10/B7‐H4 proteolytic axis could be a novel strategy to reduce B7‐H4
   and improve immune response in breast cancer treatment (Figure [181]8
   ).

Figure 8.

   Figure 8
   [182]Open in a new tab

   Proposed working model. In immune‐cold triple‐negative breast cancers,
   elevated B7‐H4 expression is a barrier that attenuates tumor immune
   responses and impedes the efficacy of targeted therapies such as
   sacituzumab govitecan. B7‐H4 levels is governed by a “yin” and “yang”
   proteolytic regulation through E3 ligase AMFR and deubiquitinase USP10,
   creating a delicate balance between degradation and stabilization. The
   stabilization of B7‐H4 by USP10 leads to diminished immune activity and
   lessened response to ADCs. Pharmacological inhibition of USP10 restore
   B7‐H4 turnover, reviving tumor immunogenicity and enhancing the
   therapeutic impact of Sacituzumab govitecan.

3. Discussion

   Based on the integration of unbiased big data analyses, mass
   spectrometry and experimental validation including preclinical models,
   we identified the impact of the USP10/B7‐H4 proteolytic axis in
   orchestrating tumor immunogenicity and the therapeutic efficacy of
   ADCs. We demonstrated the cellular protein levels of B7‐H4 are dictated
   by the proteolytic machineries USP10 and AMFR, whose deregulation
   dampens tumor immune activity and cytotoxic effectiveness of ADCs.
   Pharmacological inhibition of B7‐H4 deubiquitylation restored B7‐H4
   degradation, thereby transforming immune‐cold TNBC tumors into
   immune‐hot tumors. This further promoted the onset of ADC‐induced
   immunogenic cell death and revived the tumor immune response
   (Figure [183]8). This study provides a new paradigm to develop
   strategies to treat immune‐cold TNBC patients.

3.1. “Immune‐Cold” Status Impedes Tumor Immunogenicity and Cytotoxic Drug
Response

   The dedicated interplay between tumor immunology and the therapeutic
   efficacy of cytotoxic drugs is a pivotal aspect of contemporary breast
   cancer research. This dynamic is critically influenced by the presence
   of an “immune cold” tumor microenvironment, characterized by diminished
   immunogenicity and a weakened cytotoxic response to therapeutic
   agents.^[ [184]^27 , [185]^28 ^] This environment severely hampers the
   effectiveness of cytotoxic and chemotherapeutic strategies, as the lack
   of immune activity results in deficient inflammatory signaling and
   antigen presentation, which are essential for the recognition and
   destruction of tumor cells by effector immune cells.^[ [186]^29 ^] As a
   result, tumors fortify themselves against the intended effects of
   cytotoxic drugs. This immune evasion leads to a resilience that enables
   the tumor to withstand cytotoxic therapeutic interventions,
   underscoring the urgency to unravel the underlying mechanisms that
   govern this evasion.

   The “cold” nature of a tumor microenvironment is not a passive state
   but rather an active orchestration by the tumor to evade immune
   detection, which effectively protects the tumor by shielding it from
   the immune system's surveillance.^[ [187]^30 , [188]^31 ^] ADCs provide
   an approach to overcome the challenge and re‐educate the immune system
   to detect the tumor again by engineering and delivering potent
   cytotoxic agents directly to cancer cells with remarkable
   specificity.^[ [189]^13 , [190]^32 ^] ADCs offer a clear advantage over
   traditional chemotherapeutics by utilizing a targeted mechanism that
   limits systemic toxicity. The multifaceted action of ADCs goes beyond
   simply inducing apoptosis in cancer cells through their toxic payloads.
   They also impede receptor‐mediated signaling pathways via the attached
   antibody and induce a “bystander effect,” attracting immune effector
   cells to the tumor. This process reignites antitumor immunity and
   transforms a “cold” microenvironment into a “hot” one, which is
   beneficial for overcoming therapeutic resistance in TNBC.^[ [191]^11 ,
   [192]^33 ^] ADCs induced immunogenic cell death can promote antigen
   uptake and migration of tumor‐resident dendritic cells to the
   tumor‐draining lymph nodes, enhancing T cell priming and expansion.^[
   [193]^34 , [194]^35 ^]

   However, the function of these immune cells can be hindered by the
   engagement of immune checkpoint molecules such as PD‐L1/PD1 and
   B7‐H4.^[ [195]^3 ^] While the immune checkpoint inhibitors (ICIs) can
   unleash and sustain the T‐cell response against the tumor, limited
   fraction of TNBC patients with PD‐L1 expression and inherent resistance
   is a barrier affecting its therapeutic effectiveness.^[ [196]^36 ^]
   Therefore, identifying the impact of B7‐H4 and its proteolytic
   regulatory mechanisms in TNBC offers a potential targeting strategy to
   sensitize immune‐cold TNBC tumors to ICIs and ADCs.

3.2. Unexpected Elevation of B7‐H4 Protein Level Due to its Impaired
Proteolysis Attenuates Tumor Immune Response and Susceptibility to
Anti‐Cancer Drugs

   Our multiomic analyses pinpointed the role of B7‐H4 in orchestrating
   the immune “cold” status of tumors in a subset of TNBC patients. IHC
   analyses with various breast cancer cohorts further unveiled the
   unexpected accumulation of B7‐H4 protein in TNBC patient samples,
   indicating a correlation with dysregulated B7‐H4 turnover regulation
   and poor prognosis. Overexpression of the immunosuppressive protein
   B7‐H4 inhibits T‐cell activation and tumor adaptive immune response.^[
   [197]^14 , [198]^15 , [199]^16 , [200]^17 ^] Intriguingly, our studies
   also revealed the concurrent upregulation of TROP2, the target receptor
   for the antibody‐drug conjugate sacituzumab govitecan, in patients with
   elevated levels of B7‐H4, suggesting the potential connection between
   B7‐H4 with the therapeutic effectiveness of SG. Our recent studies also
   explored the impotence of B7‐H4 in regulating immunogenic cell death,
   where B7‐H4 interferes with the PERK‐mediated phosphorylation of eIF2α,
   a step essential for the surface exposure of CALR, acting as a “danger”
   or “eat‐me” signal for phagocytic cells.^[ [201]^3 ^]

   Consequently, B7‐H4 undermines the cascade of events that would
   typically lead to the release of DAMPs associated with immunogenic cell
   death (ICD), preventing the phagocytosis of tumor cells by dendritic
   cells and ultimately obstructing T cell recruitment to the tumor
   site.^[ [202]^3 ^] In this regard, the elevated B7‐H4 expression may
   pose a significant barrier for a subgroup of TNBC patients receiving SG
   treatment which was also validated in this study, where TNBC cells with
   high B7‐H4 exhibited significantly lower response to SG treatment.
   Given this context, the upregulation of B7‐H4 could represent a
   significant obstacle for a subset of TNBC patients undergoing SG
   therapy. Therefore, the destruction of B7‐H4 could boost tumor
   immunogenicity and promote ICD‐induced by SG, especially for the subset
   of B7‐H4^high TNBC patients.

3.3. Stabilization of B7‐H4 by USP10 Limits the Effectiveness of Sacituzumab
Govitecan in Breast Cancer Treatment

   We demonstrated the unexpected accumulation of B7‐H4 in TNBCs is
   attributed to its dysregulated turnover regulation. To maintain
   homeostasis, B7‐H4 needs to be precisely regulated by posttranslational
   modifications. While degradation of B7‐H4 is governed by ubiquitin E3
   ligase AMFR, stabilization can be achieved at two levels, including
   B7‐H4 glycosylation and deubiquitylation. Previously, we showed that
   glycosylation of B7‐H4 alters its protein conformation, which changes
   the recognition and interaction of the E3 ligase with B7‐H4.^[ [203]^3
   ^] In the present study, the mass spectrometry identified the USP10 as
   a deubiquitinase that upregulates B7‐H4 through counteracts the E3
   ligase‐mediated ubiquitylation and degradation, resulting in inhibition
   of tumor immune response and cellular response to ADC treatment.
   Stabilization of B7‐H4 by USP10 enhances the “cold” tumor environment
   and potentially reduces the therapeutic efficacy of SG. Moreover, our
   data suggest that treatment with SG may promote a feed‐forward loop
   that upregulates USP10 expression, thereby leading to further elevation
   of B7‐H4 level, and potentially reinforcing the immune evasion
   facilitated by this protein. Our work established the regulatory
   circuity for “ying” and “yang” regulation of B7‐H4 and revealed that
   deregulated USP10 in the context of etiological status is a key for the
   accumulation of B7‐H4 in TNBCs observed by IHC and cohort studies.
   Pharmacological inhibition of the USP10/B7‐H4 axis could be an
   effective strategy to convert “cold” TNBC tumors into more immunogenic
   ones and augment the effectiveness of SG in treating TNBC patients.

4. Experimental Section

Bioinformatics Analysis

   Public cancer patient datasets were analyzed with established
   bioinformatics methods.^[ [204]^37 ^] This work analyzed several public
   datasets from TCGA and CPTAC related to breast cancer using established
   bioinformatics methods^[ [205]^38 ^] and grouped into different
   subtypes according to PAM50 signature. The genes were clustered
   according to the difference between the average expression of each
   subtype and the average expression of other subtypes. Protein
   expression levels corresponding to the listed immune‐related proteins
   were analyzed and heatmap were generated. This work evaluated
   Spearman's correlation coefficients to assess the significance of the
   correlations between gene expression values in TNBC samples. For the
   hierarchical clustering of enriched terms, this work employed the
   average linkage method and visualized the correlation coefficients in a
   heatmap. Patient survival analysis and expression analysis were
   conducted in Python based on the TCGA PacCancer Atlas data sets
   available through the NCI GDC Data Portal. The immune infiltration
   analysis, immunogenicity analysis and pathway enrichment analysis were
   done by using TIMER2.0, TIDE or CAMOIP web‐based analysis servers.^[
   [206]^22 , [207]^39 , [208]^40 ^] Molecular Signature Database
   (MSigDB),^[ [209]^41 ^] KEGG pathway,^[ [210]^42 ^] GO,^[ [211]^43 ^]
   and Reactome gene sets^[ [212]^44 ^] were used for pathway and
   functional enrichment analyses. All software used in the present study
   are available in the Python Package Index (PyPI)
   ([213]https://pypi.org/) and/or other relevant Python repositories.

Cell Culture and Chemical Reagents

   HEK‐293T, MDA‐MB‐468, MDA‐MB‐231, SKBR3, 4T1, HCC1599, HCC70, and
   HCC1937 were obtained from American Type Culture Collection (ATCC)
   (Manassas, VA). 4T1‐hTROP2 cells, where the basal mouse TROP2 was
   knocked out and replaced with human TROP2. HEK‐293T, MDA‐MB‐468,
   MDA‐MB‐231, SKBR3, HCC70 & 4T1 were cultured in Dulbecco's Modified
   Eagle Medium (Corning, 10‐017‐cv), and HCC1599 & HCC1937 were cultured
   in RPMI‐1640 Medium (Corning, 10‐040‐CV), respectively supplemented
   with 10% fetal bovine serum (FBS) and 1% of streptomycin (100 U mL^−1)
   and penicillin (100 U mL^−1) at a humidity incubator at 37 °C with 5%
   CO[2]. Other chemicals used in cell culture: Cycloheximide (Cayman,
   14 126), Sacituzumab Govitecan (TRODELVY), Spautin‐1 (Selleckchem,
   S7888), Olaparib (Sigma‐Aldrich, SML3705), Doxorubicin (Sigma‐Aldrich,
   D1515) and Wu‐5 (AOBIOUS, [214]AOB11528).

Plasmids Information

   The N‐FLAG tag human B7‐H4 (HG10738‐NF) were ordered from Sino
   Biological Inc. LentiCRISPRv2 puro (Plasmid #98 290) was purchased from
   Addgene and used for the vector control for CRISPR Knock‐down. For
   CRISPR knockdown of B7‐H4 and USP10, sgRNA guide sequences were cloned
   into LentiCRISPRv2 puro following a published PCR free method.^[
   [215]^45 ^] The sgRNA sequence targeting human B7‐H4:
   5′‐TCGGGGAGCCTCACACCGCA‐3′ and 5′‐ATTGGAGACTTCCGAGAAGT‐3′. The sgRNA
   sequence targeting human USP10: 5′‐CTTACCTCAACTGAAGATCG‐3′ and
   5′‐CATGGCCCTCCACAGCCCGC‐3′. The sgRNA sequence targeting USP36:
   5′‐CGGGCTACAGCTGCCATGC‐3′; USP17: 5′‐TCAATCGAAATACAAGTGT‐3′; USP42:
   5′‐TTCCCGTTAGGACTCGACA‐3′; USP30: 5′‐GCCAAGGGAACGTTGAACG‐3′;USP49:
   5′‐GGCGTGGTCCTCAATATAG‐3′; USP44: 5′‐TGTTGGGCGTAACCACGAA‐3′; USP3:
   5′‐TCTACTAACAAATCGTAAG‐3′; USP28: 5′‐ATTTCAGCATGTCCGGGAG‐3′; USP25:
   5′‐AAGGGGAATCGTTTGGGAC‐3′; and USP13: 5′‐ ATCCACCCTCTCCGTGTAG‐3′ and
   5′‐CGCTTATGAACTAACGAGA‐3′; The successful cloning of sgRNA guide
   sequences was validated by sequencing with the primer
   5′‐GAGGGCCTATTTCCCATGATT‐3′. The N‐cMyc tag human USP10 (#184 034) were
   purchased from Addgene and then corresponding cMyc‐USP10 coding
   sequence was cloned into lentiviral vector Phage‐PGK‐Flag‐via Gibson
   assembly (ClonExpress II One Step Cloning Kit C112 from Vazyme) and
   validated by sequencing. For USP10 and B7‐H4 fragments, the indicated
   coding sequence was amplified by PCR, cloned into pIRES and pCI vector,
   respectively and validated by restriction enzyme digestion and
   sequencing. For B7‐H4 mutagenesis, the corresponding mutations were
   introduced with primers by PCR amplification, ligated via Gibson
   assembly and validated by restriction enzyme digestion and sequencing.
   For mouse USP10 knockdown, CRISPRV v2 sgRNA for mouse ups10‐1 and
   ups10‐3 were purchased from Sigma Aldrich. For mouse USP10
   overexpression, mouse usp10 coding sequence in pcDNA3 vector was
   purchased from Sigma Aldrich, cloned into lentiviral vector
   Phage‐PGK‐HA and validated by sequencing.

Cell Transfection

   Cells were transfected with PEI STAR (TOCRIS) Transfection Reagent
   following the user manual. Generally, cells were seeded the day before
   transfection to reach 50–80% confluence on the day of transfection. DNA
   and PEI STAR were mixed in Opti‐MEM medium and added to the cells after
   10 min‐incubation at room temperature. Cells were collected for
   analysis 2–3 days after transfection.

Stable Cell Line Construction with Lentivirus Transduction

   Lentiviral particles were generated by co‐transfection of the
   expression plasmid and lentiviral package plasmids (pVSV‐G, pRRE, and
   pRSV‐REV) into HEK‐293T with PEI STAR (TOCRIS). Culture media were
   harvested 48 h after transfection, filtered through 0.45‐µm filters,
   mixed with polybrene and added to the targeting cell lines for
   transduction. 2–3 days after transduction, the stable cell lines were
   established by puromycin or hygromycin selection.

Purification of B7‐H4 Complex and Mass Spectrometry

   MDA‐MB‐468 cells with stable expression of FLAG‐B7‐H4 were lysed with
   NP40 buffer (1% NP‐40, 10% glycerol, 25 mmol L^−1 Tris‐HCl (pH 7.9) and
   protease inhibitor cocktails. B7‐H4‐interactome was purified by
   immunoprecipitating FLAG‐B7‐H4 with anti‐FLAG M2 agarose beads (Sigma,
   A2220) and washed four times with TBST buffer [137 mmol L^−1 NaC1,
   20 mmol L^−1 Tris‐HCl (pH 7.6), 0.1% Tween‐20]. The interactome was
   eluted with 3× FLAG peptide (Sigma, F4799) in TBS buffer. The elute was
   then separated on SDS‐ PAGE followed by Coomassie blue staining. The
   elute was sent for mass spectrum analysis.^[ [216]^46 ^]

Immunoblotting

   Cells were lysed with 1X Laemmli sample buffer (Biorad, 1 610 737),
   boiled at 95 °C for 5 min, and centrifuged at 14 000 × g for 15 min at
   4 °C to remove the undissolved cell debris. Protein concentration of
   cell lysate was measured with the BCA assay kit (Thermo Fisher
   Scientific, 23 225). 20 µg protein was resolved on SDS‐PAGE and
   transferred to nitrocellulose membrane. Membranes were blocked with
   PBST containing 5% skim milk for 2 h, incubated with various primary
   antibodies at 4 °C overnight followed with 5 min‐washing with PBST for
   four times and the corresponding horseradish peroxidase (HRP)
   conjugated secondary antibody at room temperature for 2 h followed with
   5 min‐washing with PBST for four times. Chemiluminescence substrate was
   applied using Clarity Western ECL Substrate (Biorad, 1 705 061). Blots
   were imaged using the ChemiDoc Touch Imaging System (Bio‐Rad).
   Semiquantification of Chemiluminescence was performed using ImageLab
   (Bio‐Rad). Primary antibodies used in immunoblotting: hB7‐H4 (Cell
   Signaling Technology, 14 572), mB7‐H4 (Abcam, ab269455), USP10 (Cell
   Signaling Technology, 8501), USP13 (Cell Signaling Technology, 12 577),
   TROP2 (ABclonal, 20 824), FLAG (Cell Signaling Technology, 14 793), HA
   (Cell Signaling Technology, 3724), Ubiquitin (Cell Signaling
   Technology, 43 124), cMyc (Santa Cruz, sc789). Secondary antibodies
   used in immunoblotting: Anti‐Rabbit‐HRP (Promega, w4018) and
   Anti‐Mouse‐HRP (Promega, w4028).

Immunoprecipitation

   Cells were lysed at 4 °C in ice‐cold cell lysis buffer (25 mM Tris‐HCl,
   150 mM NaCl, 1 mM EDTA, 1% NP40, 5% glycerol) with protease inhibitor
   cocktail (SIGMAFAST, S8830) for 1 h. Cell lysates were cleared by
   centrifugation at 13 000 × g for 30 min. Protein concentration in the
   supernatant was determined using the BCA Protein assay Kit (Pierce,
   23 225). For immunoprecipitation, cell lysate was incubated with
   anti‐FLAG M2 agarose affinity gel (Sigma, A2220) or anti‐cMyc agarose
   affinity gel (Sigma, A7470) overnight at 4 °C on a rotator and washed
   with PBS for five times. The complex was eluted with 100 µg mL^−1 3X
   Flag peptides in PBS for 30 mins or boiling in 1X Laemmli sample buffer
   for 5 min.

Structure Modeling and Docking Analysis

   Structure modeling of full length human B7‐H4 and USP10 sequence was
   performed through the SWISS‐MODEL and AlphaFold approach using
   (alphaFold2.ipynb) web‐based server and structural coordinates of AMFR
   downloaded from PDB (PDB ID: 2LXH).^[ [217]^47 ^] All cocrystallized
   hetero atoms, including water molecules and cocrystallized ligand, were
   pulled out from the original coordinates. Further, the structure‐based
   molecular docking and their binding pattern was explored by using the
   molecular docking approach. There are full length variants of B7‐H4 and
   USP10, but the region where the interaction of B7‐H4 and USP10 occurs
   was selected and used for molecular docking simulations. Structural
   coordinates of B7‐H4 fragments covering the Ig‐like V‐type domain
   His‐35 to Gly‐148, and USP10 domain Ser‐516 to Tyr‐615, respectively.
   The docking was performed using the ClusPro 2.0
   ([218]https://cluspro.bu.edu/login.php) and HADDOCK 2.4 webserver,^[
   [219]^48 , [220]^49 ^] a docking program based on the Fast Fourier
   Transform (FFT) correlation approach.^[ [221]^50 ^] The docking output
   was filtered based on the interacting residue and energy values, and
   then top docked conformations (Cluster 3 docking poses) selected and
   further mutate the Lys‐146, Glu‐144 of B7‐H4 and analyzed the
   interaction using PyMOL and Discovery Studio Visualizer for their
   possible interactions.^[ [222]^51 ^]

Apoptosis Assay

   MDA‐MB‐468 or HEK‐293T cells were seeded 1 day before coculture. Human
   PBMCs were obtained from donors were thawed and resuspend with
   1 ng mL^−1 anti‐human CD3 (BioLegend, 300 438) and 1 ng mL^−1
   anti‐human CD28.2 (BioLegend, 302 934) for T‐cell activation and added
   to the seeded cells for at 3:1 ratio. Prepare extra samples for PBMC
   only control, cancer cells only control as well as single stain
   control. After 40 h co‐culture, cells were collected for staining with
   FITC Annexin V Apoptosis detection kit with PI (BioLegend, 640 914) and
   PE anti‐human‐CD‐45 (BioLegend, 304 039) which was used to stain PBMCs.
   Cells were analyzed by flow cytometry with proper machine setting and
   data were analyzed with FlowJo software (BD Biosciences).

Phagocytosis Assay

   Tumor cells were stained with 10 µM CellTracker Green CMFDA Dye
   (ThermoFisher Scientific, C7025) at 37 °C for 30 min, seeded on the
   plates, treated with 100 ng mL^−1 sacituzumab govitecan (SG) for 24 h
   to induce DAMPs, rinsed with PBS twice to remove the drug and then
   cocultured with Human monocyte derived dendritic cells (Zen‐Bio,
   SER‐MODC‐F) at 1:1 ratio for 2–3 h. Dendritic cells were collected,
   stained with Zombie Yellow, blocked with Human TruStain FcX (BioLegend,
   422 302), APC anti‐human CD11c antibody (BioLegend, 980 604) and
   PE/Dazzle 594 anti‐human HLA‐A2 Antibody (BioLegend, 343 334) and
   subjected to flow cytometry. For fluorescence imaging, dendritic cells
   were stained with CellTracker Red CMTPX Dye (ThermoFisher Scientific,
   [223]C34552) at 37 °C for 30 min before the coculture with cancer cells
   and imaged with BioTek LionHeart fluorescence microscope.

T‐Cell Function Analysis with Flow Cytometry

   Breast cancer cells were seeded 1 day before coculture. Human PBMCs
   obtained from donors were thawed and resuspend with 1 ng mL^−1
   anti‐human CD3 (BioLegend, 300 438) and 1 ng mL^−1 anti‐human CD28.2
   (BioLegend, 302 934) for T‐cell activation and added to the seeded
   cells for at 5:1 ratio. 40 h later, PMA/INOmycin/BFA (Biolegend,
   423 303) was added at 1:500 dilution and incubated for 3–4 h before
   cell collection. Cells were collected and stained with Zombie yellow,
   then CD45 (Biolegend, 304 039), CD4 (Biolegend, 317 433), and CD8
   (Biolegend, 344 722), then permeabilized, fixed and stained antibody
   for INF‐γ (Biolegend 502 528), TNF‐α (Biolegend, 502 947), and Ki67
   (Biolegend, 350 515) followed by washing. Cells were analyzed by flow
   cytometry CYTEK AURORA with proper machine setting. Zombie yellow
   negative, CD45+, CD8+ cytotoxic T cells were selected for functional
   analysis with FlowJo software (BD Biosciences).^[ [224]^37 ^]

Protein Measurement by Flow Cytometry

   Single cell suspensions were washed with PBS, then fixed with 4%
   formaldehyde in PBS. Cells were washed with PBS twice and before the
   incubation with the indicated antibodies with fluorophore conjugation.
   For the staining of cell surface protein, the staining was performed
   without membrane permeabilization. For staining of the whole cell,
   cells were permeabilized with 0.1% Triton X‐100. Cells were washed with
   PBS and analyzed by flow cytometry CTTEL AURORA with proper machine
   setting and data were analyzed with FlowJo software (BD Biosciences).
   Antibodies used for protein measurement by flow cytometry: hB7‐H4 (BD
   Pharmingen, 562 507), HSP90 (Cell Signaling Technology, 4877) HSP70
   (Cell Signaling Technology, 4873), Calreticulin (Cell Signaling
   Technology, 12 238).

Immunofluorescence

   Cells grown on coverslip were fixed with 4% paraformaldehyde in PBS for
   20 min and permeabilized by 0.1% Triton X‐100 for 15 mins at room
   temperature. The indicated antibodies diluted in PBST with 0.2% BSA
   with the dilution factor recommended by the manufacture were incubated
   with cells overnight at 4 °C. After being washed with PBST 10 min × 3,
   cells were incubated with Alexa 488 anti‐rabbit/594‐anti‐mouse IgG
   1:500 diluted in PBST with 0.2% BSA for 1 h. Cells were washed, stained
   with 1 ng mL^−1 DAPI in PBS, washed, mounted, and imaged with BioTek
   LionHeart fluorescence microscope. Primary antibodies used in
   immunofluorescence: hB7‐H4 (Sigma, HPA054200) and USP10 (Cell Signaling
   Technology, 8501). Secondary antibodies used in immunofluorescence:
   anti‐Mouse‐ Alexa Fluor 488 (Fisher Scientific, A11029 and anti‐Rabbit‐
   Alexa Fluor Plus 594 (Fisher Scientific, A32740).

Immunohistochemistry

   The human breast cancer tissue array BC081116d were purchased from TMA
   (US Biomax Inc.), including invasive ductal carcinoma cases (grade 2, n
   = 28; grade 3, n = 25) and medullary carcinoma cases (n = 9), multiplex
   IHC staining of USP10, B7‐H4 and CD8 using the Opal multiple color IHC
   kit (Akoya Biosciences) as described previously.^[ [225]^37 ^] Briefly,
   TMA slides were deparaffinized, re‐hydrolyzed, implemented antigen
   retrieval in AR9 retrieval buffer (Akoya Biosciences) followed by
   staining procedures. Briefly, the slides were blocked, incubated with
   primary antibodies, and incubated with the secondary HRP‐ conjugated
   antibodies. Heating in AR6 retrieval buffer (Akoya Biosciences) was
   applied after each staining cycle to release the bounded primary and
   second antibodies. After the staining procedures, the slide was
   counter‐ stained with DAPI. Primary antibodies used in
   immunohistochemistry: hB7‐H4 (Cell Signaling Technology, 14 572),
   mB7‐H4 (Fisher Scientific, AF2154), USP10 (Cell Signaling Technology,
   8501), TOX (Cell Signaling Technology, 73 758), Granzyme B (Cell
   Signaling Technology, 46 890), mCD8 (Cell Signaling Technology, 98 941)
   and hCD8 (Cell Signaling Technology, 85 336).

Duolink Proximity Ligation Assay

   The colocalization of B7‐H4 and USP10 was analyzed by the proximity
   ligation assay by using Duolink In Situ PLA Kit (Sigma‐Aldrich) and
   following the manufacturer's protocol. In brief, cells were plated on
   glass coverslips and fixed with 4% formaldehyde and permeabilized with
   0.1% Triton X‐100. Cells were then washed with PBS, incubated with
   blocking buffer followed by overnight primary antibody incubation at 4
   °C. Cells were then incubated with PLA probes, ligation mix,
   amplification mix and washed twice. The coverslips were mounted using
   the Duolink In Situ Mounting Medium followed by imaging the cells with
   BioTek LionHeart fluorescence microscope.

RNA Extraction and qPCR

   Total RNA was extracted using TRIzol reagent (Invitrogen) according to
   the manufacturer's instructions. The complementary DNA synthesis was
   performed using SuperScript One‐Step RT‐PCR (Invitrogen). Real‐Time
   Quantitative Reverse Transcription PCR was used to quantify genes by
   SYBR Green (Bio‐Rad), and relative abundance of each mRNA was
   normalized to beta‐actin mRNA. The primers for qPCR of B7‐H4:
   5′‐TCTGGGCATCCCAAGTTGAC‐3′ and 5′‐TCCGCCTTTTGATCTCCGATT‐3′. The primers
   for qPCR of USP10: 5′‐ATTGAGTTTGGTGTCGATGAAGT‐3′ and
   5′‐GGAGCCATAGCTTGCTTCTTTAG‐3′. The primers for qPCR of ACTB:
   5′‐AAGATCATTGCTCCTCCTGAGC‐3′ and 5′‐CATACTCCTGCTTGCTGATCCA‐3′.

In Vivo Tumor Challenge

   BALB/C (8‐week‐old female) were purchased from The Jackson Laboratory.
   Engineered 4T1 control, 4T1 USP10 KD, 4T1 USP10 OE and 4T1‐hTROP2
   breast cancer cells (1 × 10^6) were injected subcutaneously into the
   mammary fat pad of female BALB/C mice on day 0. Tumor volumes were
   measured along three orthogonal axes (a, b, and c) and calculated as
   abc/2 twice a week. The administration of drugs was done with
   Intraperitoneal injection. The tumor weight was determined at the end
   point. All animal experiments were approved by the Institutional Animal
   Care and Use Committee of Emory University.

Statistical Analysis

   Unless specified, data were plotted as mean ± SEM and experiments were
   performed at least twice. Statistical analyses were performed using
   GraphPad Prism (version 8, GraphPad Software), or R software (version
   3.3.3). Student t‐test or one‐ or two‐way ANOVA, followed by Tukey
   multiple comparisons test was applied for statistical tests. Log‐rank
   (Mantel–Cox) test was applied for the comparison of survival curves.
   Pearson correlation or Spearman correlation was applied for the
   correlation analysis. p value < 0.05 was considered significant.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   L.Z. and Y.Z. contributed equally to this work. Y.W., L.Z., and Y.Z.
   conceived the project. Y.W. supervised research. L.Z., Y.Z., X.C.,
   J.C., A.U., X.S., Z.Z., and M.D. collected, analyzed, and interpreted
   data. Y.Z. performed bioinformatic and computational analyses. L.Z.,
   Y.Z., and A.U. produced figures with input from Y.W., M.C., and K.K.
   oversaw clinical bioinformatic analyses. L.Z., Y.Z., and Y.W. prepared
   the manuscript. All authors discussed the results and commented on the
   manuscript.

Supporting information

   Supporting Information
   [226]ADVS-11-2400757-s001.pdf^ (7.3MB, pdf)

Acknowledgements