Abstract

   The functional and pharmacological significance of dopamine receptor D4
   (DRD4) in psychiatric and neurological disorders is well elucidated.
   However, the roles of DRD4 in colorectal cancer (CRC) remain unclear.
   This study observes a significant upregulation of DRD4 expression in
   clinical samples, which is negatively correlated with patient
   prognosis. In vitro, overexpression of DRD4 causes a constitutive
   activation of β‐Arrestin2/PP2A/AKT independent of dopamine.
   Interestingly, this classical signaling pathway is not associated with
   the phenotype of DRD4‐promoted migration and invasion in CRC cells.
   Instead, DRD4 interacts with transforming growth factor beta receptors
   (TGFBR1 and TGFBR2) to activate Smad2 phosphorylation and promote
   Smad2/Smad4 complex nucleus translocation. Then, SNAI1 and JAG1 are
   transcriptionally activated to induce epithelial‐mesenchymal transition
   and enhance the metastatic potential of CRC. Notably, the COOH‐terminal
   domain is identified as the key intracellular region for the
   pro‐metastatic roles of DRD4. Furthermore, treatment with a TGFBR1
   inhibitor combined with a BMP inhibitor effectively counteracts the
   pro‐metastatic effects induced by DRD4 both in vitro and in vivo. In
   conclusion, these findings uncover an unconventional role for DRD4
   beyond its classic function as a neurotransmitter receptor. The
   intracellular signaling of DRD4 interacting with TGFBR1 can be targeted
   pharmacologically for CRC therapy.

   Keywords: DRD4, colorectal cancer, TGF‐β, TGFBR1, TGFBR2
     __________________________________________________________________

   This study identifies an unconventional role of dopamine receptor D4
   (DRD4) in colorectal cancer (CRC), revealing its overexpression
   correlates with poor prognosis. DRD4 promotes metastasis by interacting
   with transforming growth factor beta receptors (TGFBR1 and TGFBR2) to
   activate Smad2 signaling and induce epithelial‐mesenchymal transition,
   independent of dopamine signaling. The COOH‐terminal domain is critical
   for this function. Blocking the DRD4‐TGFBR1 interaction effectively
   suppresses metastasis, suggesting a potential therapeutic strategy for
   CRC.

   graphic file with name ADVS-12-2413953-g001.jpg

1. Introduction

   Dopamine receptors including DRD1, DRD2, DRD3, DRD4, and DRD5 belong to
   the G protein‐coupled receptor (GPCR) family with a seven‐transmembrane
   structure. Based on their regulatory effects on adenylyl cyclase, they
   are further divided into two subfamilies.^[ [32]^1 ^] DRD1 and DRD5, as
   D1‐like receptors (D1R), activate the Gα[s/olf] family of G proteins
   upon ligand binding, stimulating an increase in intracellular cyclic
   AMP (cAMP) levels. DRD2, DRD3, and DRD4, as D2‐like receptors (D2R),
   couple with the Gα[i/o] family of G proteins, leading to a decrease in
   cAMP levels.^[ [33]^2 , [34]^3 ^] Neurotransmitter dopamine (DA)
   interacts with distinct dopamine receptors on cell membranes upon
   release from presynaptic terminals, mediating diverse physiological
   functions including reward, motivation, addiction, etc.^[ [35]^4 ,
   [36]^5 , [37]^6 , [38]^7 , [39]^8 , [40]^9 ^]

   Among those dopamine receptors, DRD4 presents the lowest expression in
   the brain; however, DRD4 expression is observed in the retina, kidneys,
   adrenal glands, sympathetic ganglia, gastrointestinal tract, blood
   vessels, and heart.^[ [41]^10 , [42]^11 ^] In these organs, the
   physiological functions of DRD4 primarily rely on its activation in two
   distinct stages. At the first stage, when in its inactive state, DRD4
   binds to Gα[i/o] protein. Upon ligand or agonist binding, it becomes
   active, releasing active Gα[i] protein that inhibits cAMP production.^[
   [43]^12 ^] This leads to decreased protein kinase A (PKA) activity and
   subsequent phosphorylation of cAMP‐response element binding protein
   (CREB), regulating downstream target gene transcription.^[ [44]^13 ,
   [45]^14 ^] At the second stage, beta‐Arrestin2 (β‐Arrestin2) acts as a
   signaling intermediary, involved in DA‐induced protein kinase B (AKT)
   inactivation independent of cAMP.^[ [46]^15 ^] Traditionally,
   β‐Arrestins are associated with GPCR signal termination and receptor
   internalization.^[ [47]^16 , [48]^17 ^] However, evidence from
   heterologous cell systems suggests that β‐Arrestins can also act as
   positive mediators of GPCR signaling independent of G proteins, via
   molecules like protein kinases.^[ [49]^18 , [50]^19 ^] In this context,
   DA binding to the DRD4 receptor triggers its activation, and the
   function of β‐Arrestin2 within this system involves the formation of a
   novel signaling complex that includes β‐Arrestin2, dephosphorylated
   AKT, and protein phosphatase 2A (PP2A), responding to DRD4 receptor
   activation, subsequently activating downstream glycogen synthase
   kinase‐3‐mediated signaling pathways.^[ [51]^20 ^]

   So far, it has been reported that DRD4 is abnormally expressed in
   various tumors. Besides nervous system tumors such as gliomas,
   glioblastoma, and pituitary tumors,^[ [52]^21 , [53]^22 , [54]^23 ^]
   DRD4 is also upregulated in breast cancer,^[ [55]^24 , [56]^25 ,
   [57]^26 ^] non‐small cell lung cancer,^[ [58]^27 ^] prostate cancer,^[
   [59]^28 , [60]^29 , [61]^30 ^] and colorectal cancer (CRC).^[ [62]^31 ,
   [63]^32 ^] Moreover, overexpression of DRD4 is associated with tumor
   progression and prognosis. For instance, both breast and CRC patients
   with elevated DRD4 expression exhibit poor prognosis.^[ [64]^24 ,
   [65]^31 , [66]^32 ^] In nervous system tumors and breast cancer, DRD4
   functions as an oncogene that promotes tumor cell proliferation and
   metastasis.^[ [67]^21 , [68]^24 , [69]^33 ^] However, the elevated
   expression of DRD4 exhibits tumor‐suppressive functions in certain
   cancers. For example, DA activation of DRD4 in pancreatic cancer leads
   to a reduction in cAMP, subsequently inhibiting the activation of the
   PKA/p38 signaling pathway, thereby suppressing tumor‐associated
   macrophage‐mediated pro‐tumorigenic inflammation.^[ [70]^34 ^]
   Additionally, in pediatric central nervous system tumors the histone
   methyltransferase the enhancer of zeste homolog 2 inhibits DRD4
   expression through enhancing the methylation of DRD4 promoter.^[
   [71]^35 ^] Nevertheless, the underlying molecular mechanisms of DRD4 in
   tumor progression remains unclear. In this study, we investigate the
   expression and function of DRD4 in CRC and elucidate the DRD4‐regulated
   signaling pathway in CRC metastasis.

2. Results

2.1. DRD4 Functions a Pro‐Metastatic Role in CRC

   To screen the neurotransmitter receptors that regulate CRC progression,
   we analyzed the relationship between the expression of 16 GPCR
   receptors and the prognosis of CRC patients using. The Cancer Genome
   Atlas (TCGA) database (Figure [72]1A; Figure [73]S1A, Supporting
   Information). Among them, not only was high expression of DRD4
   significantly correlated with poor prognosis (Figure [74]1A), but DRD4
   was also markedly upregulated in CRC tissues compared to both paired
   and unpaired normal tissues (Figure [75]1B). The correlation of DRD4
   with prognosis was further verified in the Gene Expression Omnibus
   (GEO) datasets [76]GSE14333 and [77]GSE38832 (Figure [78]S1B,
   Supporting Information). Moreover, DRD4 expression increased
   progressively from tumor‐node‐metastasis (TNM) stage I to stage IV in
   CRC (Figure [79]1C). Interestingly, DRD4 expression was also
   significantly elevated in colorectal polyps, adenomas, and
   adenocarcinomas, as analyzed through the GEO databases [80]GSE128435
   and [81]GSE41657 (Figure [82]1D). Additionally, [83]GSE131418 dataset
   indicated that DRD4 was notably upregulated in CRC distant metastases
   compared to primary tumors (Figure [84]1E). To validate these online
   data, we assessed DRD4 expression in normal colorectal tissues,
   colorectal adenomas, CRC primary tumors, and liver metastases by
   immunohistochemistry (IHC) assay. As depicted in Figure [85]1F, DRD4
   staining progressively intensified from normal tissues to adenomas,
   adenocarcinomas, and liver metastases. The statistical analysis results
   revealed that DRD4 expression was significantly higher in CRC tissues
   compared to 79 paired normal tissues and 30 paired adenoma tissues
   individually, and further increased in distant metastases
   (Figure [86]1G). In addition, we found that high expression of DRD4 was
   associated with poor prognosis in 69 cases with clinical prognostic
   information (Figure [87]1H). Multivariate Cox regression analysis
   indicated that DRD4 could be considered as an independent prognostic
   factor for survival in CRC patients (Table [88]S1, Supporting
   Information). These clinical data demonstrate that DRD4 may be
   associated with CRC progression.

Figure 1.

   Figure 1
   [89]Open in a new tab

   High expression of DRD4 leads to malignant progression of CRC. A)
   Kaplan–Meier plots of patients with CRC with high (n = 187) and low (n
   = 187) DRD4 expression. B) Expression of DRD4 in CRC paired and
   unpaired samples from TGCA database. Paired samples: (n = 41); Unpaired
   samples: Normal (n = 41); Tumor (n = 466). C) Expression of DRD4 in TNM
   stages of patients. Normal (n = 51); Stage I (n = 57); Stage II (n =
   137); Stage III (n = 113); Stage IV (n = 52). D) GEO database:
   [90]GSE128435 dataset and [91]GSE41657 dataset represented the
   correlation between DRD4 and CRC progression. [92]GSE128435: Normal (n
   = 5); Polyp (n = 22); Tumor (n = 15); [93]GSE41657: Normal (n = 12);
   Adenoma (n = 51); Adenocarcinoma (n = 25). E). GEO database:
   [94]GSE131418 dataset represented the correlation between DRD4 and CRC
   distant metastasis. [95]GSE131418: Primary cancer (n = 332); Distant
   metastasis (n = 185). F) Representative images of standardized
   immunostaining for DRD4. G) Scores were assigned according to the
   results of (F): The positive area and intensity were evaluated
   double‐blind by two senior pathologists, and the evaluation results
   were assigned points according to the rules: positive index = positive
   area multiply by positive intensity. Assignment description: Positive
   area: <5%: zero points; 5–25%: one point; 26–50%: two points; 51–75%:
   three Points; >75%: four points; Positive intensity: negative (−),
   assigned zero points; Weak positive (+), assigned one point; Positive
   (++), assigned two points; Strong positive (+++), assigned three
   points; (+–++, 1.5 points; ++–+++, 2.5 points). Clinical samples:
   Normal (n = 79); Adenoma (n = 30); CRC (n = 79); Metastasis (n = 5). H)
   Kaplan–Meier plots of clinical CRC patients with high (n = 35) and low
   (n = 34) DRD4 expression. Data are presented as mean ± SD; statistical
   significance was assessed by an unpaired t‐test, paired t‐test, or
   one‐way ANOVA.

   Next, we investigated the expression of dopamine receptors including
   DRD1, DRD2, DRD3, DRD4, and DRD5 in eight CRC cell lines, which showed
   that DRD4 was upregulated in these CRC cells compared to other dopamine
   receptor family members (Figure [96]S2A, Supporting Information).
   Subsequently, we detected DRD4 expression in the normal colonic
   epithelial cell line (NCM460) and eight CRC cell lines using real‐time
   quantitative polymerase chain reaction (RT‐qPCR) and Western Blot (WB),
   which indicated that almost all CRC cell lines exhibited higher DRD4
   levels than normal cell line. Specifically, among these CRC cell lines,
   SW620, HT29, and HCT116 carried a relatively high DRD4 expression,
   whereas SW480, HCT8, and RKO with a relatively low DRD4 expression
   (Figure [97]2A; Figure [98]S2B, Supporting Information). Moreover,
   immunofluorescence assay (IF) showed DRD4 primarily localized on the
   cell membrane of RKO, HCT116, and SW620 cells (Figure [99]S2C,
   Supporting Information). To investigate the biological roles of DRD4 in
   CRC, we overexpressed DRD4 (DRD4‐OE) in RKO (Figure [100]2B) and HCT8
   cells (Figure [101]S2D, Supporting Information), confirming its correct
   localization to the cell membrane (Figure [102]2C). Overexpression of
   DRD4 did not affect cell proliferation (Figure [103]2D) but
   significantly promoted CRC cell migration and invasion (Figure [104]2E;
   Figure [105]S2E, Supporting Information). Additionally, mesenchymal
   markers such as Slug and Snail were significantly increased, while the
   epithelial markers ZO‐1 and Occludin were decreased by DRD4
   overexpression (Figure [106]2F). We further knocked down DRD4 (DRD4‐KD)
   in HCT116 and SW620 cells by Crispr/Cas9. Although complete deletion of
   DRD4 was not achieved due to multiple copies of the DRD4 gene in the
   genome, the expression of DRD4 was significantly reduced in both HCT116
   (Figure [107]2G) and SW620 cells (Figure [108]S2F, Supporting
   Information). Reduced DRD4 expression corresponded with decreased
   migratory and invasive capabilities of CRC cells (Figure [109]2H;
   Figure [110]S2G, Supporting Information), with no significant change
   observed in cell proliferation (Figure [111]2I; Figure [112]S2H,
   Supporting Information). Meanwhile, E‐cadherin and Occludin were
   increased and Snail was decreased in DRD4‐KD HCT116 cells
   (Figure [113]2J), while ZO‐1, E‐cadherin, and Occludin were increased
   and Vimentin was decreased in DRD4‐KD SW620 cells (Figure [114]S2I,
   Supporting Information).

Figure 2.

   Figure 2
   [115]Open in a new tab

   DRD4 promotes EMT process in vitro and metastasis in vivo. A) The mRNA
   levels and protein expression of DRD4 in NCM460, RKO, and HCT116 were
   examined by RT‐qPCR and WB. β‐Actin was run as an internal control. B)
   The mRNA levels and protein expression of endogenic and exogenous
   Flag‐tagged DRD4 were measured after stably transferred into
   pCDH‐DRD4‐N‐3×Flag plasmid by RT‐qPCR and WB. β‐Actin was run as an
   internal control in RKO cells. C) Immunofluorescence staining for
   endogenic (Alexa Fluor 546, red) and exogenous DRD4 with Flag‐tag
   (Alexa Fluor 488, green). Nuclei were stained with DAPI. Scale bars,
   40 µm. D) Effect of overexpression of DRD4 on cell viability of RKO
   cells in 96 h. CCK8 test wavelength was 450 nm. E) Transwell assay to
   investigate the migratory and invasive properties of DRD4 in
   overexpressed RKO cells. The histograms on the right show the
   quantification analysis results. F) Protein levels of EMT markers were
   examined by WB in overexpressed RKO cells. β‐Actin was run as an
   internal control. G) WB to detect the expression of DRD4 in DRD4‐KD
   HCT116 cells. β‐Actin was run as an internal control in HCT116 cells.
   H) Transwell assay to investigate the migratory and invasive properties
   of DRD4‐KD HCT116 cells. The histograms on the right show the
   quantification analysis results. I) CCK8 assay was performed to detect
   cell viability of the DRD4‐KD HCT116 cells within 96 h. CCK8 test
   wavelength was 450 nm. J) Protein levels of EMT markers were examined
   by WB in DRD4‐KD HCT116 cells. β‐Actin was run as an internal control.
   K) Splenohepatic metastasis of DRD4‐Mock and DRD4‐KD HCT116 cells in
   nude mice. Mock (n = 5); KD (n = 5). L) Statistical diagram of the
   number of metastatic tumors in the livers of nude mice. Mock (n = 5);
   KD (n = 5). Data are presented as the mean ± SD; statistical
   significance was assessed by one‐way ANOVA, an unpaired t‐test, or
   two‐way ANOVA.

   To further validate the pro‐metastatic role of DRD4 in vivo, we
   inoculated DRD4‐KD HCT116 cells into the spleen and subsequently
   monitored for liver metastasis formation 2 months later. The results
   showed that DRD4‐KD HCT116 cells formed fewer metastatic foci in the
   liver (Figure [116]2K,L; Figure [117]S2J, Supporting Information).

2.2. DRD4 Promotes CRC Metastasis in a Dopamine‐Independent Manner

   To explore whether DA is required for DRD4 to exert its pro‐metastatic
   role in CRC, we initially measured DA levels by ELISA, which revealed
   lower DA levels in both the supernatants and cell lysis of CRC cells
   compared to normal colonic epithelial cells (Figure [118]S3A,
   Supporting Information). Interestingly, neither DA nor the DRD4
   receptor agonist [119]PD168077 increased RKO cell migration capability,
   whether applied as pre‐treatment or co‐culture (Figure [120]3A).
   Although cell viability increased along with increasing DA
   concentrations, no differences were observed in either
   DRD4‐overexpressing RKO cells (Figure [121]S3B, Supporting Information)
   or DRD4‐KD HCT116 cells (Figure [122]S3C, Supporting Information). Cell
   proliferation assay also indicated no difference between
   DRD4‐overexpressing and control RKO cells with DA treatment (Figure
   [123]S3D, Supporting Information), and a finding similarly observed in
   DRD4‐KD HCT116 cells (Figure [124]S3E, Supporting Information).
   Notably, even under DA treatment, DRD4 overexpression could still
   promote CRC cell migration (Figure [125]3B), whereas DRD4 knockdown
   inhibited CRC cell migration (Figure [126]3C). These results
   demonstrate that the promotion of CRC metastasis by DRD4 may be
   independent of DA presence or absence.

Figure 3.

   Figure 3
   [127]Open in a new tab

   The function of promoting cancer metastasis of DRD4 is independent of
   DA. A) Transwell assay to investigate the migratory properties of DA
   and DRD4 agonist [128]PD168077 in RKO cells with the treatment of
   pre‐culture (24 h) or co‐culture. The histograms on the right show the
   quantification analysis results. B) Transwell assay to investigate the
   migratory properties of 20 µm DA in DRD4‐overexpressing RKO cells. The
   histograms on the right show the quantification analysis results. C)
   Transwell assay to investigate the migratory properties of 20 µm DA in
   DRD4‐Mock and DRD4‐KD HCT116 cells. The histograms on the right show
   the quantification analysis results. D) Protein levels of endogenic and
   exogenous Flag‐tagged mutant DRD4 were examined by WB in exogenous DRD4
   mutant RKO cells. β‐Actin was run as an internal control. E) ELISA
   assay to detect the effect of 20 µm DA on the cAMP concentrations of
   exogenous DRD4 mutant RKO cells in 6 h. F) Protein levels of p‐CREB and
   PKA were examined by WB in exogenous DRD4 mutant RKO cells. β‐Actin was
   run as an internal control. G) Transwell assay to investigate the
   migratory and invasive properties of full‐length and DA‐binding sites
   mutant DRD4 in wild‐type and DRD4‐Mut7 RKO cells. The histograms on the
   right show the quantification analysis results. H) WB to test endogenic
   and exogenous Flag‐tagged mutant DRD4 in the DRD4‐KD HCT116 cells.
   β‐Actin was run as an internal control. I) Fluorescence intensity
   change of the effect of 20 µm DA on the full‐length and DA‐binding
   sites mutant DRD4 in the DRD4‐KD HCT116 cells with calcium probe
   incubation. J) Transwell assay to investigate the migratory and
   invasive properties of full‐length and DA‐binding sites mutant DRD4 in
   the DRD4‐KD HCT116 cells. The histograms on the right show the
   quantification analysis results. Data are presented as the mean ± SD;
   statistical significance was assessed by two‐way ANOVA.

   A previous study has predicted the binding sites of DRD4 to DA,^[
   [129]^36 ^] so we constructed two DRD4 mutants: one with seven
   mutations (D115A, L187A, Y192A, V193A, S196A, S197A, and S200A. Named
   as Mut7) and another with four mutations (D115A, V116A, C119A, and
   T120A. Named as Mut4) (Figure [130]3D; Figure [131]S3F, Supporting
   Information). As we know, activation of D1Rs typically increases cAMP
   levels, while DRD4 activation, a D2R, suppresses cAMP.^[ [132]^37 ^]
   However, the DRD4 mutants with seven mutations (DRD4‐Mut7) caused a
   significant elevation of cAMP, but the DRD4 mutants with four mutations
   (DRD4‐Mut4) still decreased cAMP level (Figure [133]3E). As the
   downstream signals, PKA and p‐CREB expression were also increased by
   DRD4‐Mut7, yet not by DRD4‐Mut4 (Figure [134]3F). These findings
   indicate that the seven amino acid sites are essential for the binding
   of DRD4 to DA, while mutations at these sites disrupt their
   interaction. Furthermore, the transwell assay showed that the DRD4‐Mut7
   still promoted CRC cell migration and invasion as same as wild‐type
   DRD4 (Figure [135]3G). The increased intracellular calcium influx could
   also serve as a marker of DRD4 activation.^[ [136]^37 , [137]^38 ^]
   When both wild‐type and Mut7 DRD4 were re‐expressed in DRD4‐KD HCT116
   cells (Figure [138]3H), intracellular calcium influx was recovered by
   expression of wild‐type DRD4, but not by Mut7 DRD4 (Figure [139]3I).
   However, not only wild‐type but also Mut7 DRD4 could rescue the
   migration and invasion capabilities of DRD4‐KD HCT116 cells
   (Figure [140]3J). Taken together, these findings suggest that the
   promotion of CRC cell migration and invasion by DRD4 is independent of
   DA signaling activation.

2.3. DRD4 Functions as Pro‐Metastatic Roles Depending on the COOH‐Terminal
Domain but not Through the Persistent Late‐Stage Activation State

   There are two crucial structural‐functional domains in DRD4: one
   located in the third intracellular loop with a 48‐bp variable number
   tandem repeat (VNTR), and the other at the palmitoylation site near the
   COOH‐terminal end. Notably, the third cytoplasmic loop of DRD4 encodes
   a VNTR tract comprising a 16‐amino acid segment, with 2–11 repeat
   variants.^[ [141]^39 ^] To investigate whether DRD4 promotes CRC
   metastasis through the variant VNTR, we produced DRD4‐overexpressing
   vectors with two or four VNTR repeats (Figure [142]4A). These results
   showed that DRD4 with either two or four VNTR repeats could rescue
   migration and invasion in DRD4‐KD HCT116 cells (Figure [143]4B),
   implying that the VNTR region is not the crucial structural domain for
   DRD4 in promoting the metastasis of CRC. Conversely, overexpression of
   DRD4 with a 16‐amino acids deletion at the COOH‐terminal end
   (DRD4‐ΔC16) in RKO cells failed to enhance migration and invasion
   capabilities (Figure [144]4C,D). Subsequent overexpression of DRD4‐ΔC16
   in DRD4‐KD HCT116 cells (Figure [145]4E) restored calcium influx levels
   similar to wild‐type DRD4 (Figure [146]4F), without affecting DA
   binding capacity. However, DRD4‐ΔC16 was not as capable of rescuing
   DRD4‐KD HCT116 cell migration and invasion as wild‐type DRD4
   (Figure [147]4G).

Figure 4.

   Figure 4
   [148]Open in a new tab

   The non‐DA‐dependent constitutive activation of DRD4 is dependent on
   the COOH terminal. A) WB to detect exogenous Flag‐tagged mutant DRD4
   with four (DRD4‐OE4.4) or two (DRD4‐OE4.2) VNTR regions in DRD4‐KD
   HCT116 cells. β‐Actin was run as an internal control. B) Transwell
   assay to investigate the migratory and invasive properties of exogenous
   Flag‐tagged mutant DRD4 with four (DRD4‐OE4.4) or two (DRD4‐OE4.2) VNTR
   regions in DRD4‐KD HCT116 cells. The histograms on the right show the
   quantification analysis results. C) WB to detect DRD4 with full‐length
   or DRD4‐ΔC16 in RKO cells. β‐Actin was run as an internal control. D)
   Transwell assay to investigate the migratory and invasive properties of
   DRD4 with full‐length or DRD4‐ΔC16 in RKO cells. The histograms on the
   right show the quantification analysis results. E) WB to detect
   exogenous Flag‐tagged mutant DRD4 with full‐length or DRD4‐ΔC16 in
   DRD4‐KD HCT116 cells. β‐Actin was run as an internal control. F)
   Fluorescence intensity change of the effect of 20 µm DA on the DRD4
   with full‐length or DRD4‐ΔC16 in DRD4‐KD HCT116 cells with calcium
   probe incubation. G) Transwell assay to investigate the migratory and
   invasive properties of DRD4 with full‐length or DRD4‐ΔC16 in DRD4‐KD
   HCT116 cells. The histograms on the right show the quantification
   analysis results. H) WB to detect the levels of p‐AKT with exogenous
   Flag‐tagged mutant DRD4 in DRD4‐KD HCT116 cells. Total AKT was run as
   an internal control. I) WB to detect the levels of β‐Arrestin2 and PP2A
   with exogenous Flag‐tagged mutant DRD4 in DRD4‐KD HCT116 cells. β‐Actin
   was run as an internal control. J) Representative images of metastatic
   lungs. EV (n = 5); OE (n = 5); Mut7 (n = 5); ΔC16 (n = 5). The white
   arrows represent the tumors. K) Statistical diagram of the number of
   metastatic tumors in the lungs of NCG mice injected via tail vein. EV
   (n = 5); OE (n = 5); Mut7 (n = 5); ΔC16 (n = 5). Data are presented as
   the mean ± SD; statistical significance was assessed by two‐way ANOVA
   or an unpaired t‐test.

   As a D2‐type dopamine receptor, DRD4 activation by DA includes early
   and late states. p‐AKT dephosphorylation occurs in the late stage and
   depends on heterotrimer formation with PP2A and β‐Arrestin2.^[ [149]^20
   ^] As expected, the level of phosphorylated AKT significantly increased
   in CRC cells following DA treatment (Figure [150]S4A, Supporting
   Information). However, overexpression of DRD4 decreased phosphorylated
   AKT levels, while DRD4 knockdown increased phosphorylated AKT levels
   (Figure [151]S4B, Supporting Information). Interestingly, regardless of
   DA treatment, DRD4 overexpression could consistently lead to a decrease
   of phosphorylated AKT in CRC cells (Figure [152]S4C, Supporting
   Information). Furthermore, both wild‐type DRD4 and DRD4‐Mut7 decreased
   phosphorylated AKT levels, whereas DRD4‐ΔC16 did not (Figure [153]4H).
   Accordingly, β‐Arrestin2 and PP2A were up‐regulated by wild‐type DRD4
   and DRD4‐Mut7 but not by DRD4‐ΔC16 (Figure [154]4I). Given that many
   previous studies have confirmed AKT signal activation could promote CRC
   epithelial‐mesenchymal transition (EMT) and metastasis, DRD4 functioned
   as a pro‐metastatic role not through activating late states DA signal.
   Further, the ability of DRD4‐ΔC16 to inhibit migration and invasion
   capabilities was reconfirmed in vitro (Figure [155]S4D, Supporting
   Information), and in vivo tail vein injection model showed that
   DRD4‐ΔC16 reduced the lung (Figure [156]4J,K) but not kidney (Figure
   [157]S4E,F, Supporting Information) metastasis capacity of HCT8 cells
   (Figure [158]S4G, Supporting Information). Together, the COOH‐terminal
   domain of DRD4 simultaneously mediates CRC metastasis and late states
   DRD4 activation, while this latter activation is not linked to its
   migratory and invasive functions.

2.4. DRD4 Activates the TGF‐β Signaling Pathway

   To elucidate which pathway activated by DRD4 contributes to CRC
   metastasis, we conducted RNA‐seq in DRD4‐KD HCT116 cells. The results
   revealed enrichment of genes associated with TGF‐β‐related signaling
   pathways (Figure [159]S5A, Supporting Information). RT‐qPCR confirmed
   that snail family transcriptional repressor 1 (SNAI1) and Jagged1
   (JAG1), among others, downstream target genes of the TGF‐β pathway,
   were upregulated by DRD4 overexpression and downregulated by DRD4
   knockdown (Figure [160]5A; Figure [161]S5B,C, Supporting Information).
   Additionally, IHC results displayed a significant positive correlation
   between DRD4 and SNAI1, as well as DRD4 and JAG1 expression in CRC
   samples (Figure [162]5B,C). Furthermore, IF assays revealed
   co‐localization of p‐Smad2 and SNAI1, as the activation signaling of
   TGF‐β pathway, which were highly expressed with DRD4 in the clinical
   CRC samples (Figure [163]5D; Figure [164]S5D, Supporting Information),
   and p‐Smad2 was also expressed in the clinical liver metastasis samples
   with high DRD4 expression (Figure [165]5E). Otherwise, p‐Smad2,
   p‐Smad1/5/9, and Smad4 levels were increased with DRD4 overexpression
   and decreased with DRD4 knockdown (Figure [166]5F), which was verified
   in HCT8 cells (Figure [167]S5E, Supporting Information). In the liver
   metastases of mice with high DRD4 expression, the elevated expression
   of p‐Smad2, p‐Smad1/5/9, and SNAI1 was also observed (Figure
   [168]S5F,G, Supporting Information). Unexpectedly, p‐Smad3 levels
   decreased when DRD4 was overexpressed, conversely, increased upon DRD4
   knockdown (Figure [169]5F). Exogenous co‐immunoprecipitation (Co‐IP)
   results showed no interaction of DRD4 and Smad2 or Smad3
   (Figure [170]5G).

Figure 5.

   Figure 5
   [171]Open in a new tab

   Constitutive activation of DRD4 is associated with the TGF‐β signaling
   pathway. A) The mRNA levels of SNAI1 and JAG1 in DRD4‐overexpressing
   RKO and DRD4‐KD HCT116 cells were examined by RT‐qPCR. β‐Actin was run
   as an internal control. B) Representative images of standardized
   immunostaining for DRD4, SNAI1, and JAG1 in clinical CRC samples. C)
   Correlation analysis of TCGA database SNAI1 and DRD4, as well as JAG1
   and DRD4. D) Immunofluorescence staining for DRD4 (Alexa Fluor 546,
   red) and p‐Smad2 (Alexa Fluor 488, green). Nuclei were stained with
   DAPI. Scale bars, 100 µm; 20 µm. E) Immunofluorescence staining for
   DRD4 (Alexa Fluor 546, red) and p‐Smad2 (Alexa Fluor 488, green).
   Nuclei were stained with DAPI. Scale bars, 100 µm; 20 µm. F) WB to
   detect the levels of p‐Smad2, Smad2, p‐Smad3, Smad3, p‐Smad1/5/9,
   Smad1, and Smad4 in the DRD4‐overexpressing RKO and DRD4‐KD HCT116
   cells. Smad2, Smad3, Smad1, or β‐Actin was run as an internal control.
   G) Immunoblotting to detect the immunoprecipitation of exogenous
   Flag‐tagged DRD4 and HA‐tagged Smad2 or Smad3 by an anti‐Flag antibody
   in 293T cells. To prevent overexposure, images are displayed in
   different exposures and the same below. H) Immunoblotting to detect the
   immunoprecipitation of exogenous Flag‐tagged AKT with HA‐tagged Smad2
   or Smad3 by an anti‐Flag antibody in 293T cells, respectively. Data are
   presented as the mean ± SD; statistical significance was assessed by an
   unpaired t‐test.

   As we know, the phosphorylation of Smad3 has traditionally been
   considered essential in the classical TGF‐β and BMP pathways. However,
   we used AKT phosphorylation inhibitor, perifosine (Peri), to treat
   DRD4‐KD HCT116 cells and SW620, which showed that suppressing AKT
   phosphorylation increased p‐Smad2 and p‐Smad1/5/9 levels but decreased
   p‐Smad3 levels (Figure [172]S6A,C,E,G, Supporting Information).
   Conversely, the AKT phosphorylation activator SC79 reduced p‐Smad2 and
   p‐Smad1/5/9 expression while increasing p‐Smad3 levels (Figure
   [173]S6B,D,F,H, Supporting Information). Furthermore, AKT could
   interact with Smad3, but not Smad2 (Figure [174]5H). These data
   elucidate that DRD4 activates Smad2 and Smad1/5/9 phosphorylation, but
   inhibits Smad3 phosphorylation through repressing the AKT pathway.

2.5. DRD4 Interacts with TGF‐β Family Receptors

   To investigate how DRD4 activates the TGF‐β pathway, we conducted Co‐IP
   experiments in 293T cells, which showed an interaction between DRD4 and
   TGFB1 (Figure [175]6A). In DRD4 knockdown HCT116 cells, exogenous DRD4
   could also pull down TGFB1 (Figure [176]6B). These results suggest that
   DRD4 may be considered a TGF‐β receptor or may form an indirect complex
   with it. IF assays showed co‐localization of DRD4 with transforming
   growth factor beta receptor 1 (TGFBR1) and transforming growth factor
   beta receptor 2 (TGFBR2) in HCT116 cells (Figure [177]6C) and clinical
   CRC samples (Figure [178]S7A, Supporting Information). Proximity
   ligation assay (PLA) signal amplification experiments in HCT116 cells
   revealed a strong interaction between DRD4 and both TGFBR1 and TGFBR2,
   but a relatively weak interaction between TGFBR1 and TGFBR2
   (Figure [179]6D). Moreover, the exogenous reciprocal interaction of
   DRD4 with TGFBR1 and TGFBR2 was confirmed in 293T cells
   (Figure [180]6E,F). However, DRD4‐ΔC16 did not reduce the interaction
   between DRD4 and TGFBR1 or TGFBR2 (Figure [181]S7B, Supporting
   Information). Instead, it decreased the phosphorylation of TGFBR1 and
   TGFBR2 by reducing the transmission of phosphorylation (Figure
   [182]S7C, Supporting Information), thereby inhibiting the activation of
   the TGF‐β signaling pathway. Intriguingly, TGFBR1 was observed to bind
   to Smad2 but not Smad3 in HCT116 cells, and Smad4 was also found to
   bind only to Smad2 (Figure [183]6G), suggesting a ligand‐independent
   interaction. The aforementioned results indicate that activated TGFBR1
   phosphorylates downstream Smad2, facilitating the formation of the
   Smad2‐Smad4 complex. Consistent with this, DRD4 overexpression enhanced
   the interaction between smad2 and smad4 in the absence of ligands
   (Figure [184]6H). To explore whether DRD4 could bind to other members
   of the TGF‐β family, molecular docking predictions (Figure [185]S8A,
   Supporting Information) and Co‐IP assay further verified the
   interaction of DRD4 with 4 type‐I receptors (4/7) of TGF‐β family
   including ALK4, ALK5 (TGFBR1), ALK6, and ALK7 (Figure [186]6E,I; Figure
   [187]S8B, Supporting Information), as well as with all five type‐II
   receptors, including TGFBR2, ACVR2A, ACVR2B, AMHR2, and BMPR2
   (Figure [188]6E,J). Meanwhile, molecular docking predictions between
   DRD4‐ΔC16 and TGF‐β family receptors reconfirmed that the loss of COOH
   terminal did not affect molecular binding but reduced TGF‐β signaling
   through phosphorylation (Figure [189]S8C, Supporting Information).
   These results underscore the important role of DRD4 in regulating the
   TGF‐β pathway.

Figure 6.

   Figure 6
   [190]Open in a new tab

   DRD4 interacts with TGF‐β family receptors. A) Immunoblotting to detect
   the immunoprecipitation of exogenous Flag‐tagged DRD4 and HA‐tagged
   TGFB1 by an anti‐Flag antibody in 293T cells. B) Immunoblotting to
   detect the immunoprecipitation of exogenous Flag‐tagged DRD4,
   endogenous TGFB1 by an anti‐Flag antibody in the DRD4‐KD HCT116 cells.
   C) Immunofluorescence staining for DRD4 (Alexa Fluor 546, red), TGFBR1
   (Alexa Fluor 647, purple) and TGFBR2 (Alexa Fluor 488, green) in HCT116
   cells. Nuclei were stained with DAPI. Scale bars, 10 µm. The white
   arrows represent the co‐location of DRD4 and TGFBR1. The yellow arrows
   represent the solitary presence of TGFBR2. The blue arrows represent
   the co‐positioning of DRD4, TGFBR1, and TGFBR2. D) PLA was performed to
   detect the interaction of DRD4, TGFBR1, and TGFBR2 in HCT116 cells.
   Alexa Fluor 546‐labeled positive sites (red). Nuclei were stained with
   DAPI. Scale bars, 10 µm. E) Immunoblotting to detect the
   immunoprecipitation of exogenous Flag‐tagged DRD4 with HA‐tagged TGFBR1
   or HA‐tagged TGFBR2 by an anti‐Flag antibody in 293T cells,
   respectively. F) Immunoblotting to detect the immunoprecipitation of
   exogenous Flag‐tagged TGFBR1 or TGFBR2 with HA‐tagged DRD4 by an
   anti‐Flag antibody in 293T cells, respectively. G) Immunoblotting to
   detect the immunoprecipitation of endogenous TGFBR1 or Smad4 with Smad2
   or Smad3 by an anti‐TGFBR1 or anti‐Smad4 antibody in HCT116 cells,
   respectively. H) Immunoblotting to detect the immunoprecipitation of
   endogenous Smad4 with Smad2 or Smad3 by an anti‐Smad4 antibody in
   DRD4‐overexpressing RKO cells, respectively. I) Immunoblotting to
   detect the immunoprecipitation of exogenous Flag‐tagged DRD4 with
   HA‐tagged ALK4, ALK6, or ALK7 by an anti‐Flag antibody in 293T cells,
   respectively. J) Immunoblotting to detect the immunoprecipitation of
   exogenous Flag‐tagged DRD4 and HA‐tagged ACVR2A, ACVR2B, AMHR2, or
   BMPR2 by an anti‐Flag antibody in 293T cells, respectively.

2.6. DRD4 Induces Ligand‐Independent Activation of TGF‐β Pathway to Promote
CRC Metastasis

   First, we treated RKO cells with varying concentrations of recombinant
   TGF‐β and BMP4 at different time points, which showed that TGF‐β
   dramatically increased p‐Smad2 and p‐Smad3 levels (Figure [191]S9A,B,
   Supporting Information), while BMP4 activated p‐Smad1/5/9 (Figure
   [192]S9C,D, Supporting Information). In DRD4 overexpressing RKO cells
   treated with TGF‐β and BMP4, DRD4 further enhanced Smad2 and Smad1/5/9
   phosphorylation levels and inhibited Smad3 phosphorylation (Figure
   [193]7A). Conversely, DRD4 knockdown decreased the Smad2 and Smad1/5/9
   phosphorylation levels activated by TGF‐β and BMP4 and increased the
   Smad3 phosphorylation (Figure [194]7B). Subsequently, wild‐type DRD4
   and DRD4‐ΔC16 were re‐expressed in DRD4‐KD HCT116 cells, and the
   results showed that wild‐type DRD4 but not DRD4‐ΔC16 activated p‐Smad2
   and p‐Smad1/5/9, inhibited p‐Smad3, and COOH‐terminal domain of DRD4
   was key to activating the TGF‐β pathway (Figure [195]7C).

Figure 7.

   Figure 7
   [196]Open in a new tab

   Inhibition of TGF‐β receptors inhibits DRD4‐induced metastasis. A) WB
   to detect the effect of concentration gradient of TGF‐β or BMP4 on
   p‐Smad2 and p‐Smad3 or p‐Smad1/5/9 in DRD4‐overexpressing RKO cells at
   24 h, respectively. Smad2, Smad3, or Smad1 was run as an internal
   control. B) WB to detect the effect of concentration gradient of TGF‐β
   or BMP4 on p‐Smad2 and p‐Smad3 or p‐Smad1/5/9 in DRD4‐KD HCT116 cells
   at 24 h, respectively. Smad2, Smad3, or Smad1 was run as an internal
   control. C) WB to detect the effect of wile‐type DRD4 and DRD4‐ΔC16 on
   p‐Smad2, p‐Smad3, and p‐Smad1/5/9 in DRD4 knockdown HCT116 cells,
   respectively. Smad2, Smad3, or Smad1 was run as an internal control. D)
   WB to detect the effect of DRD4 inhibitor L745870 and TGFBR1 kinase
   inhibitor Ga on p‐Smad2 in DRD4‐overexpressing RKO cells. Smad2 was run
   as an internal control. E) WB to detect the effect of DRD4 inhibitor
   L745870 and BMP receptors (ALK2, ALK3, and ALK6) inhibitor Dd on
   p‐Smad1/5/9 in DRD4‐overexpressing RKO cells. Smad1 was run as an
   internal control. F) Animal model diagram. G) Representative images of
   metastatic lungs. EV (n = 7); EV+G+D (n = 7); OE (n = 7); OE+G+D (n =
   7). The white arrows represent the tumors. H) Statistical diagram of
   the number of metastatic tumors in the lungs of NCG mice injected via
   tail vein. EV (n = 7); EV+G+D (n = 7); OE (n = 7); OE+G+D (n = 7). I)
   WB to detect the knockdown efficiency of TGFBR1 in DRD4‐overexpressing
   HCT8 cells. β‐Actin was run as an internal control. J) Representative
   images of metastatic lungs. EV‐DRD4+Scramble‐TGFBR1 (n = 5);
   EV‐DRD4+Sh‐TGFBR1 (n = 5); OE‐DRD4+Scramble‐TGFBR1 (n = 5);
   OE‐DRD4+Sh‐TGFBR1 (n = 5). The white arrows represent the tumors. K)
   Statistical diagram of the number of metastatic tumors in the lungs of
   NCG mice injected via tail vein. EV+Scramble (n = 5); EV+Sh (n = 5);
   OE+Scramble (n = 5); OE+Sh (n = 5). Data are presented as the mean ±
   SD; statistical significance was assessed by an unpaired t‐test.

   Interestingly, the DRD4 inhibitor L745870 could not reduce Smad2 and
   Smad1/5/9 phosphorylation levels. In contrast, the TGFBR1 kinase
   inhibitor Galunisertib (Ga) significantly inhibited Smad2
   phosphorylation (Figure [197]7D; Figure [198]S9E,F, Supporting
   Information). Similar results were observed with the BMP receptors
   (ALK2, ALK3, and ALK6) inhibitor Dorsomorphin dihydrochloride (Dd).
   Smad1/5/9 phosphorylation levels were unaffected by L745870 but were
   repressed by Dd (Figure [199]7E). These findings indicate that
   extracellular inhibition of DRD4 does not impair its ability to
   activate TGF‐β signaling, whereas blocking intracellular TGFBR1
   prevents DRD4 from activating TGF‐β signals. Furthermore, the transwell
   assay revealed that Ga and Dd could suppress CRC migration and
   phosphorylation of Smad2 and Smad1/5/9 enhanced by DRD4 overexpression
   without affecting cell proliferation within a certain time frame
   (Figure [200]S9G–I, Supporting Information). Meanwhile, Treatment with
   Ga and Dd significantly reduced the elevation of SNAI1 and JAG1 caused
   by DRD4 in DRD4‐overexpressing RKO cells (Figure [201]S10A, Supporting
   Information) and DRD4 knockdown HCT116 cells (Figure [202]S10B,
   Supporting Information). Finally, Ga and Dd were combined to pre‐treat
   DRD4 overexpression RKO cells, which were then injected into severe
   immunodeficient NCG mice via the tail vein (Figure [203]7F). The
   results indicated that RKO cells promoted CRC metastasis to livers and
   kidneys, and combination treatment with Ga and Dd markedly reduced the
   metastatic potential of DRD4‐overexpressing RKO cells to kidneys, but
   not to livers (Figure [204]S10C–E, Supporting Information). IF assays
   confirmed the co‐expression of DRD4 and p‐smad2 in mouse liver and
   kidney metastases, which could be inhibited by Ga and Dd treatment
   (Figure [205]S11A, Supporting Information). Moreover, the tail vein
   injection of DRD4‐overexpressing HCT8 cells also caused a reduction for
   the lung metastases following treatment with Ga and Dd
   (Figure [206]7G,H), while no reduction was observed in the kidneys
   (Figure [207]S11B–D, Supporting Information). To further validate that
   TGFBR1 mediated DRD4‐induced metastasis, we knocked down TGFBR1 in
   DRD4‐overexpressing HCT8 cells (Figure [208]7I; Figure [209]S12A,
   Supporting Information) and found that the migratory and invasive
   capabilities were suppressed in vitro (Figure [210]S12B, Supporting
   Information). Meanwhile, TGFBR1 knockdown significantly reduced
   DRD4‐induced lung (Figure [211]7J,k) and kidney metastases in vivo
   (Figure [212]S12C–E, Supporting Information). These in vitro and in
   vivo data suggest that DRD4 promotes CRC metastasis through interacting
   with TGF‐β family receptors independent of extracellular TGF‐β ligand
   signaling pathway.

3. Discussion

   GPCR receptors play pivotal roles in tumor biology, influencing various
   biological functions such as cell proliferation, apoptosis resistance,
   angiogenesis, stemness, and metastasis.^[ [213]^40 ^] Consequently,
   targeting GPCR receptors has emerged as a promising strategy in cancer
   therapy. Our investigation not only find DRD4 is up‐regulated in CRC
   and negatively associated with prognosis, but also highlights the roles
   of DRD4 in CRC metastasis. DRD4 overexpression in CRC causes a
   constitutive activation of the β‐Arrestin2/PP2A/AKT pathway independent
   of DA, but this activation does not mediate the pro‐metastatic roles of
   DRD4. Meanwhile, DRD4 binds to multiple TGF‐β family receptors
   (including TGF‐β and BMP signaling receptors) on the cell membrane and
   increases Smad2 and Smad1/5/9 phosphorylation to form a complex with
   Smad4. Furthermore, nuclear translocation of Smad‐complex leads to the
   transcriptional upregulation of downstream target genes such as SNAI1
   and JAG1, further promoting EMT and metastasis (Figure [214]8A).

Figure 8.

   Figure 8
   [215]Open in a new tab

   Schematic representation of constitutive DRD4 activates TGF‐β receptors
   independent of ligands to promote CRC metastasis. A) DRD4 activates the
   β‐Arrestin2/PP2A/AKT pathway independently of DA by interacting with
   TGF‐β receptors, thus triggering phosphorylation of Smad2/1/5/9 and
   promoting EMT through upregulation of SNAI1 and JAG1, ultimately
   enhancing metastatic potential.

   As we know, DRD4 is conserved as a receptor for the neurotransmitter
   DA,^[ [216]^37 , [217]^38 ^] however, it remains unknow whether the
   pro‐metastatic role of DRD4 in CRC depends on classical DA pathway.
   Despite DA‐inducing intracellular calcium influx in CRC cells featuring
   heightened DRD4 expression, this phenomenon does not translate into
   observable changes in migration and invasion capabilities.
   Consequently, the phenotypic effects initiated by DRD4 cannot be solely
   attributed to the activation of the DA signaling pathway, despite DRD4
   retaining its capability to bind to and activate this pathway. Thus, we
   hypothesize that the functionality of DRD4 in CRC may be intricately
   linked to cross‐talk with other signaling pathways. Previous studies
   confirmed that somatostatin (or somatotropin release‐inhibiting factor,
   SRIF), operating via specific G‐protein‐coupled membrane somatostain
   receptors (SSTRs), formed heterodimers with other GPCRs, such as the
   µ‐opioid receptor^[ [218]^41 ^] and DRD2,^[ [219]^42 , [220]^43 ^]
   thereby enhancing functional activity,^[ [221]^44 ^] even in the
   absence of ligands. The interplay between SSTR and DRD2 on the plasma
   membrane was implicated in promoting the proliferation of human
   prostate and lung cancer cell lines.^[ [222]^45 ^] This phenomenon, in
   which ligand‐independent receptor‐receptor interaction leads to
   functional changes, was also found in our findings.

   The VNTR in the third intracellular loop region of DRD4 typically
   comprises four or seven tandem repeats, with fewer repeats potentially
   indicating functional deficiencies or disease predisposition.^[
   [223]^46 , [224]^47 ^] Although DRD4 VNTR polymorphism has been
   associated with attention‐deficit hyperactivity disorder,^[ [225]^47 ^]
   abnormal personality traits,^[ [226]^48 ^] susceptibility to
   addiction,^[ [227]^49 ^] and Alzheimer's disease,^[ [228]^50 ^] no
   previous studies have linked the VNTR region of DRD4 to cancer. Our
   findings further elucidate that the pro‐metastatic functions of DRD4 in
   CRC are unrelated to the VNTR region. Additionally, studies have
   identified a palmitoylation site at the COOH‐terminus of DRD4,
   influencing receptor surface expression, endocytosis, signal
   transduction, and receptor function.^[ [229]^51 , [230]^52 ^] Protein
   S‐palmitoylation, a reversible post‐translational modification, plays a
   critical role in modulating the localization, stability, and function
   of various proteins within cells. This modification is crucial for the
   function of oncogenes (such as neuroblastoma RAS viral (v‐ras)
   oncogene and epidermal growth factor receptor) and tumor suppressors
   (such as scribble planar cell polarity protein and melanocortin 1
   receptor).^[ [231]^53 ^] Our findings affirm that the COOH‐terminal
   domain of DRD4 is essential for its pro‐metastatic function in CRC,
   potentially through its association with the palmitoylation site.

   DA can trigger immediate activation of the PI3K‐AKT signaling pathway,
   resulting in increased AKT phosphorylation. However, in CRC cells with
   elevated DRD4 expression, AKT phosphorylation remains persistently
   suppressed, which should be considered as a second‐stage activation of
   DRD4. Interestingly, we observed that DRD4 remained in a prolonged
   second‐stage activation state even in the absence of DA, termed as
   constitutive activation. Constitutive activation of dopamine receptors
   DRD1,^[ [232]^54 ^] DRD2,^[ [233]^55 , [234]^56 ^] DRD5^[ [235]^52 ,
   [236]^57 ^] has been reported in relation to development and
   neurological disease, but the impact of the functional change on cancer
   progression and metastasis remains unclear. Therefore, the
   physiological implications of this “weak” and “continuous” activation
   state deserves further investigation.

   We performed Gene Ontology pathway enrichment analysis on the RNA‐Seq
   results and identified a close association between the high expression
   of DRD4 and the TGF‐β signaling pathway. Notably, our findings indicate
   that DRD4 can modulate the activation of the TGF‐β signaling
   independently of its ligands. Co‐IP experiments confirmed that DRD4
   could interact with 4 type I receptors and all 5 type II receptors of
   the TGF‐β family (formation of heterotetramers with combinations of 2
   type I and 2 type II receptor molecules is essential for TGF‐β
   signaling pathway activation), indicating multiple possibilities for
   combinatorial activation and comprehensive TGF‐β family activation in
   CRC. Moreover, in the absence of ligands, DRD4 could interact with
   TGF‐β receptors to co‐activate, as demonstrated by PLA and IF
   experiments that showed intense binding between DRD4 and TGFBR1, and
   mutual binding even in the absence of TGFBR2. These results suggest a
   direct interaction between DRD4 and TGFBR1, leading to mutual
   activation. Otherwise, we found that the DRD4‐ΔC16 mutant does not
   alter the binding affinity with TGFBR1 and TGFBR2 but reduces the
   phosphorylation levels of both receptors. These results indicate that
   DRD4 modulates TGF‐β signaling, at least in part, by regulating the
   phosphorylation of TGFBR1 and TGFBR2. Thus, our findings suggest that
   DRD4 may regulate TGF‐β signaling through direct interaction and
   modulation of TGFBR1 and TGFBR2 phosphorylation. Meanwhile, in the
   absence of exogenous TGF‐β and BMP4, constitutively activated DRD4
   interacted with TGF‐β receptors, resulting in increased levels of
   phosphorylated Smad2 and Smad159, as well as elevated expression of
   target genes SNAI1 and JAG1. Given that the DA‐independent constitutive
   activation state of the DRD4 receptor still allowed it to bind to and
   activate ligands, we investigated TGF‐β receptors and found that the
   addition of recombinant TGF‐β and BMP4 still significantly activated
   the TGF‐β signaling pathway. These results demonstrate that both DRD4
   and TGF‐β receptors maintain their respective abilities to bind to
   ligands and activate downstream signaling pathways in CRC.

   Furthermore, our study reveals the ability of DRD4 to bind to TGFB1,
   but the precise mechanism of this interaction remains elusive. This
   interaction may be caused by faulty ligand recognition or indirect
   binding through TGF‐β receptors, but its occurrence underscores the
   significant correlation between DRD4 and the TGF‐β family. Besides, the
   activation of the TGF‐β signaling pathway induced by DRD4 exhibits
   deviations from the classical activation pathway. For example, we
   observed a significant inhibition of Smad3 phosphorylation by DRD4,
   which was different from classical TGF‐β signaling pathway activation.
   We demonstrated that Smad3 phosphorylation was inhibited by
   DRD4‐induced dephosphorylation of AKT. Conversely, recombinant TGF‐β
   stimulation increased Smad3 phosphorylation, highlighting disparities
   in the signaling pathways between DRD4‐induced and ligand‐induced
   activation.

4. Conclusion

   In conclusion, our findings elucidate that DRD4's interaction with
   TGFBR1/TGFBR2 constitutively activates the TGF‐β signaling pathway,
   thereby driving CRC metastasis. Consequently, DRD4 emerges as a
   potential biomarker for CRC progression and a promising pharmacological
   target for CRC therapy.

5. Experimental Section

Clinical Materials and Ethics Statement

   One hundred fourteen clinical CRC samples paraffin sections were
   collected from the Run Run Shaw Hospital, Zhejiang University, between
   2018 and 2019. This study included 79 samples of CRC, 30 samples of
   adenomas, and 5 samples of metastases, and prognostic information was
   obtained from 69 cases. Informed consent was obtained from all the
   included patients. The study was approved by the Ethics Board of
   Biomedicine, Zhejiang University (2017027), China. The animal
   experiments were performed according to the Guide for the Care and Use
   of Laboratory Animals (The Ministry of Science and Technology of China,
   2006). The use of animals in this study was approved by the Ethics
   Committee of Zhejiang University (reference number: ZJU20240236).

Main Reagents and Cell Culture

   Cell lines obtained from American Type Culture Collection include
   NCM460, RKO, HCT116, SW620, SW480, HT29, HCT8, DLD1, LoVo, and HEK293T
   were cultured with DMEM basic (1×) or RPMI Medium 1640 basic (1×). DMEM
   and 1640 were supplemented with 10% FBS and 1% penicillin‐streptomycin.
   Cells were cultured at 37 °C with 5% CO[2] condition.

Plasmid Construction and Transfection

   This study employed pcDH, pcDNA3.1, and pLKO.1 plasmids for the
   construction of overexpression and knockdown protein vectors. The
   overall strategy involved amplification of the target fragments with
   sticky ends from the cDNA of 293T cells, using a high‐fidelity enzyme
   (Phanta Max Master Mix Dye Plus). These fragments were then subjected
   to homologous recombination with double‐enzyme‐cut plasmids to form
   intact plasmids. Transfection of CRC cell lines utilized the DNA
   transfection reagent Neofect (Genomtech, China), while transfection of
   293T cells involved LipoD 293 reagent (SignaGen, USA), and siRNA
   transfection utilized GenMute siRNA Transfection Reagent (SignaGen,
   USA). Lentiviral packaging plasmids pMD2G and pXPAX2 were involved in
   lentivirus packaging, and the viruses were collected and filtered
   through 0.45 µm filters 48 h post‐transfection. Subsequently, depending
   on the requirement, different cell lines were infected, and puromycin
   (1 mg mL^−1, 1–20 µL) or G418 (100 mg mL^−1, 1–20 µL) was used for 4–7
   days of selection to obtain the desired cell lines.

Crispr‐Cas9 Technology

   Two lentiviral sgRNA clones targeting the human DRD4 gene were procured
   from GeneCopoeia (Rockville, MD, USA). Lentiviruses were generated in
   HEK293 cells using the pCRISPR‐LvSG06, pVSVG, and pSPAX2 vectors.
   Subsequently, these lentiviruses were employed to transduce HCT116 and
   SW620 cells. Following ≈7 days of selection with puromycin, individual
   viable cells were isolated and cultured for ≈2 weeks. Finally, the
   knockdown efficiency of monoclonal cells was identified through DNA
   sequencing and immunoblotting. Primer sequences utilized are provided
   in Table [237]S2 (Supporting Information).

Cell Proliferation Assay

   Cells were seeded in a 96‐well plate at a density of 2000–5000 cells 24
   h prior. The following day, the cells were treated with drugs. At
   specific time points, 100 µL of fresh culture medium was replaced, and
   after incubating for 1–2 h following the addition of 10 µL of CCK8
   reagent (Vazyme, China) to each well, the absorbance was measured at
   450 nm detection wavelength to determine the OD values.

Cell Migration and Invasion Assay In Vitro

   Twenty‐four‐well transwell plates with 8‐µm pore size (Corning, USA)
   were employed for investigating cell migration and invasion. The filter
   membrane on the back of each chamber was uniformly coated with
   20 µg mL^−1 fibronectin (FN) from Sigma (Germany). After air‐drying,
   50 µL of serum‐free culture medium, diluted 50 times, containing
   matrigel (Corning, USA) was added to the chambers, followed by
   incubation at 37 °C for 30 min. In the lower chambers of the 24‐well
   plate, 800 µL of complete culture medium with 10% serum was added.
   Cells (5×10^4–2×10^5) were diluted in 200 µL serum‐free culture medium
   and added to the chambers for 12–48 h of incubation. At specific time
   points, the medium in the chambers was aspirated, and the chambers were
   washed three times with PBS. Subsequently, they were fixed in 4%
   paraformaldehyde (LABLEAD, China) for 20 min. After three washes with
   PBS, the fixed chambers were stained in 0.001% crystal violet for 2–5
   min, followed by rinsing with tap water. Excess cells inside the
   chambers were wiped off using a cotton swab. Finally, random fields
   were captured using an inverted microscope (Nikon ECLIPSE TS100,
   Japan), and ImageJ was utilized for cell quantification.

Real‐Time Quantitative PCR (RT‐qPCR)

   Cells were subjected to digestion using the Total RNA Isolation Reagent
   Trizol (Pufei, USA) for mRNA extraction. Subsequently, mRNA (1 µg) was
   reverse transcribed into cDNA using the reverse transcriptase HiScript
   II Q SuperMix for qPCR (+gDNA wiper, Vazyme, China). Specific cDNA
   synthesis and amplification, as well as polymerase chain reaction
   (PCR), were conducted following the respective product manuals. To
   quantify gene expression levels, β‐Actin was used as an internal
   reference, and the relative quantification of mRNA was calculated using
   the 2^‐ΔCt method. The primer pairs used are provided in Table [238]S2
   (Supporting Information).

Western Blot

   In brief, protein samples were separated on 8%, 10%, or 12% SDS‐page
   gels and subsequently wet‐transferred to nitrocellulose membranes (NC,
   Bio‐Rad, Hercules, USA). After blocking with 5% high‐protein milk for 1
   h, the membranes were incubated with antibodies overnight at 4 °C and,
   on the following day, incubated with fluorescent secondary antibodies
   for 1 h in the dark. Exposures were performed using the Odyssey CLx
   Infrared Imaging System (LI‐COR, USA).

Elisa Assay

   Cell culture supernatants or cell lysates, collected and diluted in PBS
   gradients, were prepared. For cAMP ELISA assay, 1×10^6 cells per 200 µL
   PBS with PMSF added was recommended. Standard wells, test sample wells,
   and blank wells were set up separately. For each well of the ELISA
   plate, 50 µL sample dilution was added, followed by the addition of
   working solution A. The plate was then incubated at 37 °C for 1 h and
   washed. Subsequently, working solution B was added, and the plate was
   incubated at 37 °C for another 1 h before washing. Chromogenic
   substrates were added to each well, and after terminating the reaction,
   absorbance was measured at a wavelength of 450 nm.

Co‐Immunoprecipitation (Co‐IP)

   FLAG beads (Sigma, Germany) were employed to capture proteins
   translated from exogenously transfected plasmids with a Flag tag in
   293T cells. After elution, the eluted proteins were incubated with an
   HA‐tag primary antibody (CST, USA) to examine protein interactions. The
   Biolinkedin Classic IP/Co‐IP Kit (Biolinkedin, China) was used to
   capture endogenously expressed proteins in CRC RKO and HCT116 cells.
   Similarly, after elution, the eluted proteins were incubated with the
   corresponding binding protein antibodies to examine protein
   interactions.

Immunofluorescent Staining

   Cells were pre‐seeded in wells of glass‐bottom dishes, and the
   following day, they were fixed in 37 °C paraformaldehyde for 30 min.
   Permeabilization was achieved using 0.1% Triton‐X100 at 37 °C for 10
   min, followed by blocking with 10% FBS for 30 min. The cells were then
   incubated with the primary antibody at 4 °C overnight. On the second
   day, after washing away the primary antibody, the cells were incubated
   with the secondary antibody (donkey anti‐mouse‐IgG‐Alexa Fluor 488,
   donkey anti‐rabbit‐IgG‐Alexa Fluor 546, or donkey anti‐goat‐IgG‐Alexa
   Fluor 647) from Invitrogen, USA, for 1 h. DAPI (Sigma, Germany) was
   used for nuclear staining and incubated for 20 min. Fluorescent signals
   were captured using a confocal microscope (FV2000, Olympus, Japan).

Duolink PLA Fluorescence

   Cells were pre‐seeded in wells of glass‐bottom dishes, and the
   following day, they were fixed in 37 °C paraformaldehyde for 30 min.
   Permeabilization was achieved using 0.1% Triton‐X100 at 37 °C for 10
   min, followed by blocking with Duolink blocking solution at 37 °C for
   60 min. After incubating with the primary antibody overnight at 4 °C
   and washing the next day, cells were incubated with PLA MINUS and PLUS
   probes for 1 h at 37 °C. Subsequent steps included washing, addition of
   the ligation enzyme with a 30‐min incubation at 37 °C, addition of the
   polymerase enzyme with a 100‐min incubation at 37 °C, and washing. The
   cells were then sealed with Duolink In Situ Mounting Medium containing
   DAPI and imaged using a laser confocal microscope (FV2000, Olympus,
   Japan).

Calcium Probe

   The day before, cells were seeded in immunofluorescence dishes. The
   following day, cells were incubated at 37 °C for 30 min with a solution
   containing 2 µm Fluo‐4 AM to load the fluorescence probe. After
   appropriate washing, an additional 30‐min incubation was carried out to
   ensure thorough esterase cleavage of Fluo‐4 AM within the cells,
   resulting in the formation of Fluo‐4, which then remained
   intracellular. Subsequently, after the addition of DA, fluorescence
   detection of Fluo‐4 was continuously performed for 15–30 min using the
   STORM Ultra High Resolution/Nikon A1 Confocal/Total Internal Reflection
   Microscope (N‐STORM/A1R, Nikon, Japan) to monitor changes in
   intracellular calcium ion concentrations.

Immunohistochemical Analyses

   The paraffin‐embedded tissue sections stored at −20 °C were brought to
   room temperature. Subsequently, they were placed in a 60 °C oven for 1
   h to completely dissolve the paraffin on the slices. The paraffin
   sections were sequentially immersed in three fresh xylene solutions,
   each for 3 min. Following this, the sections were immersed in a
   gradient series of ethanol solutions with concentrations of 100%, 100%,
   95%, and 75%, for 1 min each. They were then washed three times for 1
   min each with PBS solution. The sections were immersed in a 3%
   H[2]O[2]‐methanol solution for 15 min, followed by a PBS wash. The
   paraffin sections were immersed in a citrate buffer, brought to a boil
   at 200 °C, and kept under constant pressure and temperature for 3 min.
   After cooling on ice to room temperature, the sections were washed with
   PBS solution. Tissues were outlined with a pencil, sealed with 10%
   fetal bovine serum, and incubated at room temperature for 30 min. After
   PBS wash, the primary antibody was applied and left overnight at 4 °C.
   On the next day, after washing with PBS, the secondary antibody
   (ultrasensitive enzyme) was applied for 1 h at room temperature.
   Following PBS wash, 100 µL of a diluted DAB solution was applied, and
   the reaction was timed for 3–5 min. The color change in the tissue was
   observed, and the slides were placed in tap water to stop the color
   development. The slides were rinsed under running water until the color
   became clear. Hematoxylin staining was performed for 30 s, and then the
   slides were rinsed under running water until the color became clear.
   Subsequently, the slides were sequentially immersed in a gradient
   series of ethanol solutions at concentrations of 75%, 95%, 100%, and
   100%, each for 5 min. The sections were removed and air‐dried at room
   temperature on a slide rack. A drop of neutral resin was applied to the
   slides, avoiding bubble formation.

Animal Metastasis Assays

   All animal experiments were conducted in accordance with the protocol
   approved by the Zhejiang University Institutional Animal Care and Use
   Committee (Ethics Committee Number: 20742). Male NOD/ShiLtJGpt
   Prkdcem26Cd52Il2rgem26Cd22/Gpt mice aged 4 weeks, denoted as NCG
   (GemPharmatech, China), were intravenously injected with 1×10^6 RKO or
   HCT8 cells. Metastatic tissues were isolated for histological analysis
   at 5–8 weeks. Spleen‐liver metastasis experiments were conducted using
   5‐week‐old male nude mice. Under anesthesia, 1×10^6 HCT116‐Mock and
   HCT116‐KD cells were orthotopically injected into the spleen. After 2
   months, spleen and liver tissues were dissected for histological
   analysis.

RNA‐Seq

   Total RNA was extracted from HCT116‐mock, HCT116‐KD‐DRD4, SW620‐Mock,
   and SW620‐KD cells, and RNA sequencing and analysis were performed by
   RiboBio (Guangzhou, China). The construction of cDNA libraries was
   carried out using the Illumina TruSeq RNA Sample Preparation Kit
   (Illumina, San Diego, CA, USA). Individual RNA sequence libraries were
   pooled based on their respective sample‐specific 6 bp adapters, and
   sequencing was performed on the Illumina HiSeq 3000 sequencer with
   150bp/paired‐end reads. Differential gene expression analysis was
   conducted using the Cuffdiff program within the Cufflinks software
   package. Genes with padj ≤0.05 and fold change 1 or fold change ≤‐1
   were defined as differentially expressed genes, serving as candidate
   genes for further analysis and qPCR validation. The accession number
   for sequencing data was PRJNA1109554.

Statistical Analyses

   GraphPad PRISM and SPSS Statistics were used for statistical analysis.
   Normally distributed data were subjected to analysis using Student's
   t‐test, while non‐normally distributed data were analyzed using the
   non‐parametric Mann–Whitney U‐test. Comparison of two or more groups of
   data was analyzed using one‐way ANOVA or two‐way ANOVA. The data are
   presented as mean ± SD. Significance values are directly provided in
   the figures; “ns” indicates not significant.

Conflict of Interest

   The authors declare no conflict of interest.

Author Contributions

   Y.Z. performed the experiments, processed the data, and drafted the
   manuscript. J.T. collected clinical samples, performed IHC scoring, and
   gathered prognosis information for patients. M.W. analyzed the
   relationship between neurotransmitter receptors and prognosis. H.Z.
   provided experimental guidance and revised the manuscript. M.L.
   contributed to the experimental design.

Ethical Statement

   Informed consent was obtained from all the included patients. The study
   was approved by the Ethics Board of Biomedicine, Zhejiang University
   (2017027), China. The animal experiments were performed according to
   the Guide for the Care and Use of Laboratory Animals (The Ministry of
   Science and Technology of China, 2006). The use of animals in this
   study was approved by the Ethics Committee of Zhejiang University
   (reference number: ZJU20240236).

Supporting information

   Supporting Information
   [239]ADVS-12-2413953-s001.docx^ (13.7MB, docx)

Acknowledgements