Dopamine D₂ receptor agonists abrogate neuroendocrine tumour angiogenesis to inhibit chemotherapy-refractory small cell lung cancer progression

Abstract

Small cell lung cancer (SCLC) is difficult to treat due to its aggressiveness, early metastasis, and rapid development of resistance to chemotherapeutic agents. Here, we show that treatment with a dopamine D₂ receptor (D₂R) agonist reduces tumour angiogenesis in multiple in vivo xenograft models of human SCLC, thereby reducing SCLC progression. An FDA-approved D₂R agonist, cabergoline, also sensitized chemotherapy-resistant SCLC tumours to cisplatin and etoposide in patient-derived xenograft models of acquired chemoresistance in mice. Ex vivo, D₂R agonist treatment decreased tumour angiogenesis through increased apoptosis of tumour-associated endothelial cells, creating a less favourable tumour microenvironment that limited cancer cell proliferation. In paired SCLC patient-derived specimens, D₂R was expressed by tumour-associated endothelial cells obtained before treatment, but D₂R was downregulated in SCLC tumours that had acquired chemoresistance. D₂R agonist treatment of chemotherapy-resistant specimens restored expression of D₂R. Activation of dopamine signalling is thus a new strategy for inhibiting angiogenesis in SCLC and potentially for combatting chemotherapy-refractory SCLC progression.

Introduction

Lung cancer is the leading cause of cancer deaths worldwide [1]. Small cell lung cancer (SCLC) accounts for 15–17% of lung cancer cases and ~200,000 deaths each year globally [2, 3]. Most newly diagnosed SCLC patients present with extensive-stage disease and initially respond to first-line chemotherapy but frequently develop drug resistance, resulting in a dismal five-year survival rate of only 7% [4, 5]. SCLC has recently been categorized into three subtypes with unique transcriptional characteristics and therapeutic vulnerabilities based upon differential expression of the transcription factors ASCL1 (SCLC-A; 51%), NEUROD1 (SCLC-N; 23%), and POU2F3 (SCLC-P; 7%) [6, 7], and a fourth subtype exhibits low expression of all three transcription factors with high inflammation and susceptibility to immune checkpoint inhibitors (SCLC-I; 17%) [6]. Molecularly targeted therapies have improved survival time of patients with non–small cell lung cancer (NSCLC) [5], but similar targeted treatments in SCLC have largely failed, and chemotherapy has remained standard of care for over three decades. Outcomes from chemotherapy are poor, with a median overall survival rate of 9.4 months [8, 9]. Recent progress has been made improving overall survival of SCLC patients with US Food and Drug Administration (FDA)-approved immunotherapies, including the PD-L1 inhibitors atezolizumab and durvalumab [8, 10, 11]. For example, the addition of atezolizumab to platinum-based frontline chemotherapy extends median overall survival by 2 months [11]. Unfortunately, however, only a small subset of patients with extensive-stage SCLC experience deep and durable responses to immune checkpoint blockade, and reliable prognostic biomarkers to identify such potential responders do not currently exist. In spite of robust initial responses to first-line platinum-etoposide with or without immunotherapy, nearly all SCLC patients eventually relapse [8]. Consequently, developing improved treatment approaches for extensive-stage SCLC is an urgent unmet need.

The tumour microenvironment has recently emerged as an active promoter of cancer progression. A dynamic and reciprocal relationship between cellular components of the tumour microenvironment and cancer cells is established early in tumorigenesis. For example, tumour-associated endothelial cells respond to cues within the tumour microenvironment to promote tumour angiogenesis; the newly formed vessels supply the tumour with oxygenated blood and provide a provisional matrix capable of supporting additional vascular sprouting and tumour growth. Vascular endothelial growth factor (VEGF) A is an essential regulator of tumour angiogenesis. VEGF was initially discovered as “vascular permeability factor”, a tumour-secreted protein that potently promotes microvascular permeability [12]. Only later was it discovered separately as an endothelial mitogen [13] essential for the development of blood vessels [14–16]. VEGF signals predominantly through VEGF receptor 2 (VEGFR2) to regulate endothelial cell function by activating downstream signalling, including phospholipase Cγ–mediated activation of the mitogen-activated protein kinase (MAPK)/extracellular-signal-regulated kinase-1/2 (ERK1/2) pathway and phosphatidylinositol 3′ kinase (PI3K)-induced stimulation of the AKT (protein kinase B) signalling cascade [17, 18]. VEGF plays a crucial role in the development and homoeostasis of the lung vasculature, and the lungs exhibit the highest level of VEGF gene expression among physiologically normal tissues [19]. In pathological conditions, the expression of VEGF and its receptors is frequently affected by the specific characteristics of the disease and its stage in progression [20]. VEGF levels are higher in SCLC patients than in healthy individuals [21, 22]; correspondingly, increased serum VEGF levels are the only independent prognostic factor other than tumour stage in untreated SCLC patients [23], as confirmed by Zhan and colleagues through a large review and meta-analysis of VEGF expression in lung cancer [24]. Increased secretion of VEGF by tumour cells and upregulation of VEGFR2 in SCLC promote tumour angiogenesis, which provides tumours with the blood supply necessary to grow and metastasize.

Inhibition of angiogenesis has improved progression-free survival in several human cancers, including NSCLC [25, 26]. Bevacizumab, a humanized monoclonal antibody that binds all forms of VEGF-A, was FDA-approved for the treatment of NSCLC in 2006 [27] but has limited efficacy against SCLC when combined with chemotherapy [28–35]. Many other angiogenesis inhibitors, including thalidomide [36, 37] and a variety of small molecule tyrosine kinase inhibitors of VEGF receptors, such as sunitinib [38–41], vandetanib [42], cediranib [43], sorafenib [44], and nintedanib [45], have failed, achieved modest responses, or caused an unacceptable degree of toxicity in SCLC. Anlotinib, an orally administered tyrosine kinase inhibitor of VEGFR and other growth factor receptors, was recently approved by the China FDA as a third-line therapy for Chinese patients with SCLC following clinical trials (ALTER 1202 study) [46, 47]; however, to date no anti-angiogenic agents have received regulatory approval in the US for treatment of SCLC. Importantly, because SCLC lacks predictive biomarkers for response to VEGF inhibition, whether a subset of SCLC patients do respond to anti-VEGF treatment is unknown. Here, we seek to overcome this issue by manipulating the dopamine signalling pathway to inhibit angiogenesis, progression, and drug resistance in SCLC.

Dopamine is an important neurotransmitter in the central nervous system that is produced by sympathetic nerves that end on blood vessels. Dopamine acts on its target cells to activate cyclic adenosine monophosphate (cAMP), an important intracellular second messenger that regulates many cellular functions, through its receptors, which belong to the G protein–coupled receptor superfamily and are broadly classified as D₁ and D₂ types [48]. The D₁ class includes D₁R and D₅R, which increase intracellular cAMP upon activation [49]. The D₂ types, including D₂R, D₃R, and D₄R, inhibit intracellular cAMP [49]. D₂R is expressed by a variety of cell types, including neurons, immune cells, endothelial cells, and endothelial progenitor cells [48] and colocalizes with VEGFR-2; D₂R/VEGFR-2 crosstalk mediates dephosphorylation of VEGFR2 [50, 51]. Dopamine and other D₂R agonists bind to D₂R expressed on the surface of endothelial cells to inhibit VEGF-mediated angiogenesis; they also completely block accumulation of tumour ascites and tumour growth in mice [52]. Conversely, increased angiogenesis, tumour growth, and VEGFR2 phosphorylation are observed in D₂R knockout mice [53]. D₂R agonists increase the efficacy of anti-cancer drugs in preclinical models of breast and colon cancer through their anti-angiogenic effect on tumour-associated endothelial cells [54]. Dopamine signalling effector molecules regulate lung cancer progression [55–58]. In a lung injury model, we showed that dopamine inhibits pulmonary oedema through the endothelial VEGF-VEGFR2 axis [59]. We also demonstrated that an FDA-approved D₂R agonist, cabergoline, reduced NSCLC growth in mice and reported that a subset of NSCLC patients have increased endothelial D₂R expression, suggesting potential precision oncology treatment strategies are possible [60]. Work by several laboratories clearly suggests that this strategy can be translatable for therapy in different diseases, including cancer [53, 61, 62]. Based on these collective prior observations, we sought to understand how dopamine signalling regulates therapy-refractory SCLC progression and to determine whether D₂R agonist treatment can inhibit growth of chemotherapy-resistant SCLC by reducing angiogenesis.

Results

D₂R agonists inhibit SCLC growth in a human xenograft model

To begin testing the hypothesis that D₂R agonists inhibit SCLC growth by reducing tumour angiogenesis, we orthotopically injected luciferase-labelled human DMS-53 SCLC cells into the left thorax of SCID mice (Fig. 1a). Non-invasive luciferase imaging demonstrated that the SCLC tumours had established themselves within the lungs of mice by 7 days after injection of the DMS-53 cells. Starting on day 8, we injected vehicle or the D₂R agonist quinpirole (10 mg/kg) intraperitoneally every other day for 13 days (i.e., until day 20). Quinpirole is a selective D₂R agonist that has been shown to inhibit angiogenesis within the tumour microenvironment [52], including NSCLC [60]. Luciferase imaging on day 21 showed that DMS-53 lung tumour growth decreased in mice treated with D₂R agonist quinpirole relative to that in vehicle-treated control mice (Fig. 1b, c; Supplementary Fig. 1), supporting the hypothesis that D₂R agonists inhibit SCLC growth in an orthotopic mouse model of human SCLC.

Fig. 1. Activation of D₂R signalling by quinpirole reduces tumour growth in a human small cell lung tumour xenograft model.

a Experimental timeline: SCID mice were orthotopically injected with luciferase-labelled human DMS-53 SCLC cells, then imaged for bioluminescence 7 days after the injection but before the start of treatment. Mice then received intraperitoneal injections of PBS vehicle (control group) or 10 mg/kg quinpirole (treatment group) every other day for 13 days, after which they were imaged again for bioluminescence. b Distribution of tumour growth rate across the vehicle- and quinpirole-treated groups. Each circle represents an individual mouse. Data are shown as mean ± SEM. A value of P ≤ 0.05 (two-way unpaired t-test) was considered significant. c Representative luciferase imaging from day 7 (before treatment with D₂R agonist) and day 21 (after treatment with D₂R agonist).

D₂R agonist cabergoline reduces tumour growth in PDX models of SCLC

We next evaluated whether D₂R agonists reduce neuroendocrine tumour progression using a human SCLC patient-derived xenograft (PDX) model. Specifically, a previously described human SCLC PDX specimen, named MSK-LX40 [63], was subcutaneously injected into nonobese severe combined immunodeficient γ (NSG) mice. After establishment of tumours of at least 100–200 mm³ in size, mice were intraperitoneally administered either vehicle (PBS) or cabergoline (5 mg/kg) daily for 2 weeks (Fig. 2a). The FDA-approved dopaminergic medication cabergoline (Dostinex^®) is a potent D₂R agonist (K_i=0.7) that is used in the treatment of high prolactin levels [64], Parkinson’s disease [65], ovarian hyperstimulation syndrome [66], Cushing’s disease [67], and restless legs syndrome [68]. Cabergoline boasts an extended elimination half-life relative to other D₂R agonists, safely providing a long-lasting clinical effect [69]. Like quinpirole, cabergoline reduces tumour angiogenesis in NSCLC and other cancer types [52, 60].

Fig. 2. D₂R agonist treatment reduces tumour growth and promotes apoptosis of tumour-associated endothelial cells in a chemonaïve human SCLC PDX model.

a–d NSG mice were subcutaneously injected with 5 × 10⁶ human SCLC cells originally derived from a chemonaïve patient (MSK-LX40) [63]. Mice were treated with either vehicle (10% DMSO in PBS) or cabergoline (5 mg/kg) daily for 14 days once mean tumour volume reached 100–200 mm³. Tumour growth was recorded by measuring tumour volume with callipers three times a week (a). Experiments were concluded before average tumour volume exceeded 1500 mm³. At the endpoint of the experiments, mice were euthanized, and xenografted tumours were harvested and weighed using a digital balance (b). Tumour volume was calculated from calliper measurements after extirpation (c). Each circle represents an individual mouse (b, c). Images of harvested tumours were acquired (d). e Co-immunofluorescence staining was performed on formalin-fixed, paraffin-embedded (FFPE) tissues harvested from NSG mice (n = 3 mice per group) treated with either vehicle or cabergoline to detect both the TUNEL⁺ and CD31⁺ cells in the tumour microenvironment. f The colocalization of TUNEL and CD31 staining was quantified by counting the number of double-positive cells (i.e., TUNEL⁺ and CD31⁺ cells) per CD31⁺ blood vessel. Each circle represents one visual field (≥5 visual fields were counted for each tissue). g Immunofluorescence staining was performed using a monoclonal Ki-67 antibody on FFPE tumour tissue specimens harvested from chemonaïve human SCLC tumour–bearing mice treated with vehicle (left) or 5 mg/kg cabergoline (right). h The proliferation of tumour cells within tissues harvested from cabergoline- and vehicle-treated mice (n = 3 for each group) was quantified by counting the number of Ki-67-positive cells per visual field ( ≥ 5 visual fields were counted for each tissue). Data are shown as mean ± SEM. A value of P ≤ 0.05 (two-way unpaired t-test) was considered significant. i, j HUVEC treated with vehicle (DMSO) or 100 μM cabergoline for 24 h were transferred to a 1% O₂ hypoxia chamber for an additional 6 h (i) or treated with hypoxia-inducer 150 μM CoCl₂ for an additional 24 h (j). Lysates were collected and immunoblotted for HIF1α and reference protein, GAPDH.

Tumour growth was inhibited in SCLC PDX–bearing mice treated with cabergoline compared with growth in controls (Fig. 2a). At the endpoint, the mice were euthanized and their tumours harvested. Tumour weight and volume were lower in the cabergoline-treated group than in the controls (Fig. 2b–d). To test our hypothesis that D₂R agonist treatment reduces SCLC tumour growth by reducing tumour angiogenesis and creating a less favourable tumour microenvironment that limits cancer cell proliferation, we co-stained tumour specimens harvested from the mice for CD31 to detect tumour-associated endothelial cells and for TUNEL (terminal deoxynucleotidyl transferase–mediated deoxyuridine triphosphate nick end labelling) to quantify apoptosis. As expected, we observed more CD31 and TUNEL co-staining in cabergoline-treated tumours than in vehicle-treated tumours (Fig. 2e, f), suggesting that D₂R agonist treatment increases apoptosis of tumour-associated endothelial cells. Ex vivo immunostaining for Ki-67 revealed significantly less proliferation of cancer cells within the patient-derived SCLC xenografts of mice treated with cabergoline than of control mice treated with vehicle (Fig. 2g-h).

Inhibition of angiogenesis leads to transient normalization of tumour vessels [70], likely producing a temporary increase in oxygen, reduced hypoxia, and increased efficacy of conventional chemotherapies. To experimentally validate this theory, we exposed cultured human umbilical vein endothelial cells (HUVEC) to hypoxic conditions (i.e., 1% O₂ or CoCl₂) in the presence of vehicle control or D₂R agonist cabergoline. As expected, exposure to hypoxic conditions results in upregulation of hypoxia-inducible factor 1α (HIF1α). Cabergoline treatment during exposure to hypoxic conditions resulted in substantially reduced HIF1α protein expression, suggesting that D₂R agonists reduce hypoxia (Fig. 2i-j; uncropped blots: Supplementary Fig. 2).

Taken together, these findings suggest that the FDA-approved D₂R agonist cabergoline decreases SCLC progression by reducing tumour angiogenesis and hypoxia through apoptosis of tumour-associated endothelial cells, leading to decreased proliferation of cancer cells within the SCLC PDX.

To support the notion that D₂R agonist treatment reduces angiogenesis through apoptosis of endothelial cells, we reasoned that an apoptosis inhibitor should rescue human endothelial cells from D₂R agonist-induced apoptosis. To test this, we modelled angiogenesis using tube formation assays. HUVEC were treated with vehicle or cabergoline in the presence and absence of an apoptosis inhibitor. As expected, in the absence of apoptosis inhibitor, cabergoline treatment caused apoptosis of endothelial cells, resulting in dramatically reduced angiogenesis, as evidenced by inhibition of tube formation (Supplementary Fig. 3). However, the addition of apoptosis inhibitor completely reserved the effect of cabergoline, as no defects in tube formation occurred in HUVEC treated with both cabergoline and apoptosis inhibitor (Supplementary Fig. 3).

We next sought to determine whether D₂R agonists reduce tumour angiogenesis and SCLC progression through other endothelial cellular functions beyond apoptosis within the lung tumour microenvironment. To assess how D₂R agonist cabergoline affects endothelial cell proliferation, we treated human endothelial cells with cabergoline and measured proliferation in real time using Incucyte-based live cell imaging and after 72 h using a luminescence-based cell viability assay. We observed that cabergoline treatment significantly reduces human endothelial cell proliferation (Supplementary Fig. 4). To evaluate the impact of D₂R agonist treatment on endothelial cell migration, we performed scratch wound assays and found that cabergoline treatment substantially decreases the migratory capabilities of human endothelial cells (Supplementary Fig. 5). Collectively, these observations suggest that D₂R agonist cabergoline likely reduces tumour angiogenesis by decreasing proliferation and migration of tumour microenvironment resident endothelial cells, in addition to D₂R agonist-mediated stimulation of tumour-associated endothelial cell apoptosis.

D₂R agonist cabergoline shows anti-proliferative effects in PDX models of chemotherapy-refractory SCLC

Given that normalization of the tumour vasculature improves intratumoral accessibility of anti-cancer agents [70], we hypothesized that D₂R agonist treatment may reduce chemotherapy-refractory SCLC progression by decreasing tumour angiogenesis within the lung tumour microenvironment, enhancing the anti-cancer effects of the chemotherapy. To test this hypothesis, we relied on three SCLC PDXs (MSK-LX40R, JHU-LX108R, JHU-LX33R) that had been previously passaged through mice treated with the standard-of-care chemotherapy regimen, cisplatin + etoposide, such that the PDXs acquired chemoresistance [63]. We sought to evaluate the impact of D₂R agonist treatment on mice bearing chemotherapy-refractory SCLC PDX tumours. Therefore, we implanted chemotherapy-resistant SCLC PDX tumours into the right flank of NSG mice, monitored animals every week until the establishment of the palpable tumour (i.e., tumour volume 100–200 mm³), and then randomized the mice into two groups. Each group of mice received weekly regimens of 5 mg/kg cisplatin on day 1 and 8 mg/kg etoposide on days 1, 2, and 3. One of the two groups also received weekly regimens of 5 mg/kg cabergoline on days 1–5 of each week for 2–5 weeks (depending on how aggressive each PDX was). The combination treatment of cabergoline and chemotherapy reduced tumour growth across all three PDX models of chemotherapy-refractory SCLC (Fig. 3a–c). At the conclusion of treatment, the mice were euthanized and tumours excised, weighed, and measured. As expected, we observed a decrease in ex vivo tumour weight and volume in cabergoline-treated mice compared with those treated without cabergoline (Fig. 3d–l), suggesting that D₂R agonism sensitizes chemotherapy-refractory human SCLC to cisplatin and etoposide.

Fig. 3. Treatment with D₂R agonist and chemotherapy overcomes chemoresistance in human SCLC PDX models.

Five million human tumour cells harvested from chemoresistant SCLC PDXs (MSK-LX40R (a), JHU-LX108R (b), and JHU-LX33R (c)) were subcutaneously injected into the right flanks of NSG mice. When tumour volume reached 200 mm³ in size, mice were randomly divided into treatment groups such that each group had similar mean tumour volumes. Mice were treated by intraperitoneal injection of 5 mg/kg cisplatin on day 1 and 8 mg/kg etoposide on days 1, 2, and 3 with or without 5 mg/kg cabergoline daily for the indicated times. Tumour volume was measured 2–3 times per week until the final tumour reached either five times the initial tumour volume (a) or 2000 mm³ in size (b, c). d–f At the endpoint, the extirpated tumours were weighed on a digital scale. g–i The final volume of the tumours from euthanized mice was measured using a digital calliper. j–l Harvested tumours from each group were arranged randomly and photographed. Each circle represents a tumour harvested from an individual mouse. Data are shown as mean ± SEM. A value of P ≤ 0.05 obtained with two-way ANOVA followed by Sidak’s multiple test (a–c) or two-way unpaired t-test (d–i) was considered significant.

To evaluate whether D₂R agonist treatment promotes apoptosis of tumour-associated endothelial cells, we performed ex vivo co-immunofluorescence staining with TUNEL and anti-CD31 antibodies using SCLC tissue specimens derived from mice harbouring the chemotherapy-resistant MSK-LX40R PDX. More endothelial cells were TUNEL-positive in cabergoline- and cisplatin/etoposide-treated tissue specimens than in samples treated only with cisplatin/etoposide (Fig. 4a, b), suggesting that activation of D₂R signalling stimulates apoptosis in tumour-associated endothelial cells. Given that normalization of the tumour vasculature improves intratumoral accessibility of anti-cancer therapies [70], we hypothesized that D₂R agonist treatment may reduce chemotherapy-refractory SCLC progression. Indeed, histological analysis of the MSK-LX40R model revealed that mice treated with cabergoline and cisplatin/etoposide had lower Ki-67 expression than mice treated exclusively with cisplatin/etoposide (Fig. 4c, d).

Fig. 4. D₂R agonist treatment together with chemotherapy decreases tumour cell proliferation and promotes endothelial apoptosis in a chemoresistant SCLC PDX model.

a FFPE tumour tissues (n = 3 per group) derived from mice treated with either chemotherapy alone or both cabergoline and chemotherapy were used for a co-immunofluorescence study to detect double-positive TUNEL⁺ and CD31⁺ cells in the tumour microenvironment. b To measure D₂R agonist-induced apoptosis in tumour-associated blood vessels, colocalization of TUNEL and CD31 staining was quantified in FFPE tumour tissues (n = 3 per group) harvested from mice bearing MSK-LX40R SCLC PDXs. Each circle represents one visual field, and five visual fields were counted for each tissue. c Immunofluorescence using monoclonal Ki-67 antibody was performed on FFPE tumour tissues (n = 5 per group) obtained from a chemoresistant human SCLC PDX model (MSK-LX40R) treated with 5 mg/kg cisplatin on day 1 and 8 mg/kg etoposide on days 1, 2, and 3 with or without 5 mg/kg cabergoline daily. Scale bar, 50 µm. d. The number of Ki-67-positive cells from the experiment illustrated in (c) was quantified. Each circle represents one visual field, and five visual fields were counted for each tissue. Data are shown as mean ± SEM. A value of P ≤ 0.05 was considered significant, two-way unpaired t-test.

To further test our hypothesis, we first created 3D SCLC PDX organoids from JHU-LX33R (Fig. 5a). We next stably transduced HUVEC with either a D₂R-specific shRNA to silence D₂R or a control LacZ-specific shRNA and treated these HUVEC with a D₂R agonist (Supplementary Fig. 6, with full, uncropped blots). We collected the conditioned medium from the D₂R agonist–treated HUVEC cultures and placed it on the dissociated SCLC PDX organoids for 72 h (Fig. 5a). Conditioned medium derived from D₂R agonist–treated HUVEC stably transduced with the control LacZ shRNA elicited a higher apoptotic response of PDXs than vehicle-treated shControl HUVEC (Fig. 5b; Supplementary Fig. 7). As expected, conditioned medium derived from D₂R agonist–treated HUVEC stably transduced with D₂R shRNA did not cause a change in apoptosis relative to vehicle-treated shControl HUVEC (Fig. 5b; Supplementary Fig. 7), suggesting that D₂R agonist signalling through D₂R in endothelial cells is necessary to produce growth factors and cytokines present in the conditioned medium that stimulate apoptosis of the SCLC PDX cells. These observations correspond with our in vivo findings demonstrating that D₂R agonist treatment reduces chemotherapy-refractory SCLC progression.

Fig. 5. Activation of D₂R signalling in the SCLC tumour microenvironment promotes apoptosis of SCLC organoids and contributes to an enhanced CD8⁺ T cell response.

a, b Conditioned medium from human endothelial cells treated with a D₂R agonist promotes apoptosis of SCLC organoids. a The SCLC PDX JHU-LX33R was suspended in Matrigel and plated at a density of 1 × 10⁵ cells in 50 µL. Complete growth medium was added to the PDXs, which were grown in a humidified chamber at 37 °C supplied with 5% CO₂. After 14 days, the Matrigel domes were dissolved by adding dispase (1 U/mL) containing cold complete medium, followed by TrypLE, then the cells were replated at a density of 10⁴ cells per well of a 96-well plate in Matrigel and topped off with 100 µL conditioned medium harvested from HUVEC that had been stably transduced with plasmids encoding control LacZ-specific shRNA (shControl CM) or D₂R-specific shRNA (shD₂R CM) and subsequently treated with a D₂R agonist (quinpirole). b Conditioned medium collected from shControl HUVEC treated with the D₂R agonist quinpirole (50 µM) increased the caspase-3-mediated apoptotic response in the SCLC chemotherapy-resistant organoid model. c, d Treating human SCLC PDX organoids with D₂R agonist quinpirole contributes to an enhanced CD8⁺ T cell response in the SCLC immune microenvironment. c JHU-LX33R SCLC PDX organoids were treated with D₂R agonist quinpirole (50 µM) for 72 h, and conditioned medium from the organoids was collected and placed on human CD8⁺ T cells. d Human CD8⁺ T cells were cultured in growth medium containing 100 IU/mL IL-2 (IL-2: negative control) or a T cell activation cocktail of 100 IU/mL IL-2, CD3 antibody, and CD28 antibody (Activation: activation cocktail only; Vehicle CM: activation cocktail and vehicle-treated SCLC PDX conditioned medium; Quinpirole CM: activation cocktail and 50 µM quinpirole-treated SCLC PDX conditioned medium; Quinpirole: activation cocktail and 50 µM quinpirole). The human T cells were subjected to flow cytometry using fluorophore conjugated antibodies to detect granzyme B and results have been plotted in a bar graph. Micrograph (right of the bar graph) showing the gating strategy for identification of CD8⁺ T cells expressing Granzyme B. Two independent biological repeats and 4 technical repeats were performed. One-way ANOVA followed by Tukey’s multiple comparisons test was performed.

To explore the impact of D₂R agonists on the immune microenvironment, we sought to investigate whether treating SCLC PDX organoids with D₂R agonist affects human T cell activation. To that end, we collected conditioned medium from JHU-LX33R SCLC PDX organoids treated with D₂R agonist quinpirole or vehicle control and put the conditioned medium on cultured human CD8⁺ T cells. The serine protease granzyme B is a key component of cytotoxic granules released by T cells that facilitates T cell-mediated cancer cell killing [71]. We observed increased granzyme B production by human CD8⁺ T cells exposed to conditioned medium from JHU-LX33R SCLC PDX organoids treated with D₂R agonist quinpirole compared to CD8⁺ T cells grown in vehicle-treated PDX conditioned medium (Fig. 5c, d). Our finding suggests that D₂R agonists may help promote activation of CD8⁺ T cells in the lung tumour microenvironment. Furthermore, we observed that treatment with D₂R agonist quinpirole significantly reduced the expression of PD-L1 on the surface of cells from the JHU-LX33R SCLC PDX organoids relative to treatment with control vehicle (Supplementary Fig. 8). The interaction between immune checkpoint protein PD-L1 on cancer cells with PD-1 on T cells inhibits the anti-cancer cytotoxic activity of T cells [72]. Therefore, the D₂R agonist-mediated reduction in PD-L1 expression on the surface of SCLC cells within the organoids suggests that D₂R agonist may improve the CD8⁺ T cell responses against cancer cells. Taken together, D₂R agonists may contribute to an enhanced T cell response against SCLC cells within the immune microenvironment.

Chemotherapy-resistant SCLC-A specimens express less D₂R on the surface of tumour-associated endothelial cells than matched chemotherapy-naïve specimens

We previously demonstrated a positive correlation between endothelial D₂R expression and tumour stage through immunostaining of tumour specimens from NSCLC patients [60]. Therefore, we sought to assess D₂R protein expression by immunostaining in paired chemotherapy-naïve and chemotherapy-resistant specimens from a cohort of SCLC-A patients (Supplementary Table 1). Briefly, tumour specimens were collected from SCLC patients before chemotherapy (i.e., chemonaïve) and following the development of chemotherapy-refractory disease progression. Following immunostaining with a D₂R-specific antibody, a pulmonary pathologist (Y-CL) reviewed the D₂R immunohistochemistry staining and scored the percentage of D₂R-positive endothelial cells present in each stained lung tissue specimen. The number of D₂R-positive endothelial cells in tumour specimens obtained following chemotherapy resistance was lower than that in the paired samples biopsied prior to chemotherapy treatment (Fig. 6). Collectively, our results suggest that protein expression of D₂R decreases as SCLC-A tumours acquire resistance to chemotherapy.

Fig. 6. D₂R is expressed by tumour-associated endothelial cells derived from SCLC patients.

a, b Immunohistochemistry was performed using a monoclonal D₂R antibody on FFPE human SCLC-A tissue samples biopsied from each patient before the start of the chemotherapy (i.e., chemonaïve) and following progressive disease after chemotherapy (i.e., chemoresistant). Representative images from paired chemonaïve (a) and chemoresistant (b) SCLC tumour specimens used to assess D₂R expression in tumour-associated endothelial cells before and after the chemotherapy. c Each tissue was scored for the percentage of endothelial cells that were D₂R-positive, and each circle on the graph represents an individual tissue (n = 9). Data are shown as mean ± SEM. A value of P ≤ 0.05 (two-way unpaired t-test) was considered significant.

Endothelial expression of D₂R increases in chemotherapy-resistant SCLC PDX models in response to D₂R agonist treatment

Because D₂R expression on the surface of SCLC tumour-associated endothelial cells decreases as the tumours developed resistance to chemotherapy, we evaluated whether D₂R agonist treatment affected the expression patterns of D₂R in this context. Specifically, chemotherapy-sensitive MSK-LX40 or chemotherapy-resistant MSK-LX40R SCLC-A PDX specimens were subcutaneously implanted into the right flank of NSG mice. The mice were monitored until the establishment of palpable tumours (i.e., tumour volume 100–200 mm³). Mice were then randomized into four groups. Two groups received weekly regimens of 5 mg/kg cisplatin on day 1 and 8 mg/kg etoposide on days 1, 2, and 3. One of these two groups also received weekly regimens of 5 mg/kg cabergoline on days 1–5 of each week for three weeks. The third group received this cabergoline regimen in the absence of cisplatin and etoposide. The fourth group of mice were administered vehicle as a negative control. The combination treatment of cabergoline and cisplatin/etoposide reduced tumour growth relative to either treatment alone or vehicle control (Fig. 7a–c). An immunofluorescence-based colocalization study showed colocalization of D₂R and CD31, indicating that endothelial cells express D₂R in both MSK-LX40 and MSK-LX40R tumour tissues, although endothelial D₂R expression was lower in vehicle-treated chemotherapy-resistant MSK-LX40R tumour tissues than in vehicle-treated chemotherapy-sensitive MSK-LX40 specimens (see vehicle-treated group in Fig. 7d; Supplementary Fig. 9), similar to our observation in paired chemotherapy-naïve and chemotherapy-resistant specimens from the cohort of SCLC-A patients (Fig. 6). Our immunofluorescence results showed no statistically significant changes in endothelial D₂R expression upon cisplatin/etoposide treatment between chemotherapy-naïve and chemotherapy-resistant SCLC PDXs (see cisplatin/etoposide treated group in Fig. 7d; Supplementary Fig. 9). However, cabergoline treatment alone or together with cisplatin/etoposide was associated with a statistically significant increase in endothelial D₂R expression in chemotherapy-resistant MSK-LX40R tumour tissues compared to chemonaïve tissue samples (see cabergoline- and cisplatin/etoposide- & cabergoline-treated group in Fig. 7d; Supplementary Fig. 9). Our collective results suggest that while lung endothelial D₂R expression decreases as SCLC develops acquired resistance to chemotherapy, administration of the D₂R agonist cabergoline results in increased expression of D₂R, suggesting that low D₂R expression in chemotherapy-refractory SCLC will not necessarily render D₂R agonist treatment ineffective.

Fig. 7. Endothelial D₂R expression is increased by D₂R agonist treatment in a chemoresistant SCLC PDX model.

a–c NSG mice were subcutaneously implanted with 5 × 10⁶ cells obtained from chemoresistant (MSK-LX40R) human SCLC PDXs. Mice were randomly divided into four groups: 1) vehicle; 2) cisplatin/etoposide; 3) cabergoline; and 4) cabergoline & cisplatin/etoposide. When mean tumour volume reached 200 mm³, mice were intraperitoneally administered vehicle (10% DMSO in PBS); 5 mg/kg cisplatin on day 1 and 8 mg/kg etoposide on days 1, 2, and 3, with or without 5 mg/kg cabergoline; or 5 mg/kg cabergoline alone five times a week. Tumour volume was measured three times a week until the vehicle-treated final tumour volume reached 2000 mm³ in size (a). At the endpoint, tumours were harvested from euthanized mice and photographs were taken to visualize gross morphology (b). Weight of the extirpated tumours was recorded using a digital scale (c). d Co-immunofluorescence staining was performed on FFPE tumour tissues (n = 3 per group) using primary antibodies against CD31 and D₂R. The number of endothelial cells expressing D₂R was quantified by counting the number of double-positive cells (i.e., D₂R⁺ and CD31⁺) divided by the CD31⁺ blood vessels present in a single visual field (≥5 visual fields were counted for each tissue). Data are shown as mean ± SEM. A value of P ≤ 0.05 (two-way unpaired t-test) was considered significant.

Discussion

In this study, we sought to overcome the ineffectiveness of angiogenesis inhibitors for SCLC treatment by manipulating the dopamine signalling pathway to inhibit angiogenesis, progression, and drug resistance in human SCLC. We demonstrated that D₂R agonists sensitize chemotherapy-resistant SCLC tumours to cisplatin and etoposide using several PDXs to model acquired chemotherapy-refractory SCLC progression in mice. Biochemical analyses of tissue from these mice suggest that treatment with a D₂R agonist reduces tumour angiogenesis through increased apoptosis of tumour-associated endothelial cells, leading to a less favourable microenvironment for tumours and impeding cancer cell proliferation. In paired specimens derived from individuals with SCLC-A subtype tumours before and after the progression of chemotherapy-refractory disease, we showed that tumour-associated endothelial cells express D₂R before exposure to chemotherapy. As SCLC tumours develop resistance to chemotherapy, D₂R levels decrease, but treatment with a D₂R agonist enhances D₂R expression in endothelial cells from chemotherapy-resistant specimens.

Tumour angiogenesis is induced by VEGF-A binding to VEGFR2 on the surface of tumour-associated endothelial cells to activate downstream signalling pathways. Therefore, VEGF-A is a therapeutic target for inhibition of angiogenesis to normalize the tumour vasculature [73, 74]. Bevacizumab, a humanized anti-VEGF-A monoclonal antibody, was approved by the FDA for the treatment of NSCLC in 2006 [27] and for the treatment of recurrent glioblastoma in 2009 [75]. Moreover, bevacizumab treatment has been linked to inhibition of vessel growth, regression of newly formed vessels, and normalization of the vasculature to facilitate the delivery of cytotoxic chemotherapy [73]. Bevacizumab in combination with cisplatin and etoposide chemotherapy has been shown to improve progression-free survival in a selected SCLC patient population, but no significant improvement of overall survival benefits was observed [76]. Similarly, the combination of bevacizumab and paclitaxel in chemotherapy-sensitive SCLC has failed to yield any noteworthy clinical outcomes [77], highlighting the fundamental differences between SCLC and NSCLC [78]. The failure of anti-angiogenic agents to improve overall survival in SCLC underscores the need for finding an effective targeted therapy to improve the outcome of chemotherapy and/or anti-angiogenic therapy in relapsed chemotherapy-refractory SCLC.

Our rationale for assessing the anti-cancer and anti-angiogenesis effects of D₂R agonists as a potential therapy to inhibit human SCLC progression stems from (1) our prior finding that D₂R agonists inhibit angiogenesis and reduce tumour progression in murine models of NSCLC [60], (2) the role of dopamine signalling in physiological neuronal function [79], and (3) the neuroendocrine origin of SCLC [80]. In support of our hypothesis that D₂R agonists inhibit SCLC growth by reducing tumour angiogenesis, a meta-analysis of 348,780 patients with Parkinson’s disease showed that patients who received dopaminergic therapy for the Parkinson’s had a 47% reduction in risk of developing lung cancer [81]. Furthermore, the FDA-approved D₂R agonist cabergoline reduces tumour size in prolactinoma, a non-cancerous adenoma of the pituitary gland which typically has high D₂R expression [82, 83].

Identifying biomarkers that predict responsiveness to therapy beyond classification of the tumour into one of the four current subtypes of SCLC will be critical to advancing treatment. Schlafen family member 11 (SLFN11) is a factor implicated in DNA damage repair deficiency. SCLC-A cell lines with high expression of SLFN11 are more resistant to cisplatin, whereas SLFN11 low expression was accompanied by increased sensitivity to cisplatin [63], suggesting a potential clinical biomarker. Similarly, our observation that dynamic changes in D₂R expression in tumour-associated endothelial cells depend on tumour responsiveness to chemotherapy and the activation status of D₂R signalling may lead to a clinically useful biomarker.

Dopamine and D₂R agonists have been shown to selectively inhibit VEGF-induced angiogenesis and vascular permeability by negatively regulating VEGFR2 phosphorylation [52], resulting in the inhibition of endothelial cell migration [51]. Specifically, dopamine treatment increases VEGF-induced phosphorylation of phosphatase-2 containing Src homology region 2 domain (SHP-2) and its phosphatase activity in HUVEC. Subsequent dephosphorylation of VEGFR2 at Y951, Y996, and Y1059 by active SHP-2 inhibits VEGF-dependent signalling events, including those that promote angiogenesis and endothelial cell migration [51]. For example, in both dopamine-depleted and D₂R-knockout mice, VEGF-induced phosphorylation of VEGFR2, MAPK, and focal adhesion kinase is substantially increased relative to the levels in control mice, indicating that dopamine signalling through D₂R regulates these signalling pathways required for endothelial cell barrier integrity, proliferation, and migration [84]. Correspondingly, dopamine treatment in endothelial progenitor cells prevents their participation in tumour neovascularization by inhibiting their mobilization from the bone marrow niche [61]. Prior work provides a strong rationale for the concept that D₂R agonist-mediated inhibition could be an effective therapeutic strategy [53, 61, 62]. Studies have demonstrated that disrupting peripheral dopaminergic nerves promotes tumour growth by triggering VEGF-dependent angiogenesis [53], whereas dopamine treatment reduces the migration of tumour-promoting endothelial progenitor cells from the bone marrow [61]. D₂R agonists have been shown to enhance the effectiveness of anti-cancer drugs in preclinical models of breast and colon cancer [54].

Using paired tumour specimens collected from SCLC patients before chemotherapy (i.e., chemotherapy-naïve) and following the development of chemotherapy-refractory disease progression, we have observed that chemotherapy-resistant SCLC-A specimens express less D₂R on the surface of tumour-associated endothelial cells than matched chemotherapy-naïve specimens. Investigating whether decreased endothelial D₂R expression in chemotherapy-resistant tumours is associated with VEGF/VEGFR2 pathway activation warrants future investigation. Given that activation of D₂R signalling has been shown to inhibit VEGF/VEGFR2 signalling [52, 60], we speculate that downregulation of endothelial D₂R expression in chemotherapy-resistant tumours could potentiate increased VEGF/VEGFR2 signalling. Importantly, we have demonstrated in mouse models that D₂R expression on the surface of SCLC PDX tumour-associated endothelial cells is restored upon treatment with D₂R agonist, cabergoline. The precise molecular mechanisms regulating cabergoline-mediated upregulation of endothelial D2R expression are unknown and require future evaluation. However, a previous study suggests that cabergoline promotes vascular barrier stability through activation of the Wnt/β-catenin by downregulation of its natural inhibitor, DKK3, and upregulation of tight junction protein, ZO-1, in endothelial cells of mice treated with LPS to induce permeability [85]. In this prior study, cabergoline also inhibited the production of pro-inflammatory cytokines which had been elevated by LPS in the vessels [85]. We speculate that cabergoline treatment reduces vascular permeability and inflammation within the SCLC PDX tumour microenvironment, leading to increased expression of D₂R on the surface of tumour microenvironment resident endothelial cells, potentially through feedback mechanisms resulting from activation of the Wnt/β-catenin signalling pathway and upregulation of ZO-1.

Efforts to study the biology of SCLC have been hampered by the fact that most cases are inoperable, and biopsies are rarely obtained at recurrence. We sought to overcome this hurdle by relying on several previously generated PDX models of human SCLC resistance to cisplatin and etoposide in mice [63]. Moreover, we have established an SCLC organoid model derived from the SCLC PDX tissue to study the effects of D₂R agonists on functional processes, such as apoptosis, that could reduce chemotherapy-refractory SCLC progression. We have also taken advantage of paired SCLC-A subtype tumour specimens obtained from individual patients prior to chemotherapy (i.e., chemotherapy naïve/sensitive) and following chemotherapy-refractory disease progression (i.e., chemotherapy resistant). We anticipate that these models and specimens will drive future efforts to study how activation of D₂R signalling promotes anti-angiogenic responses in the tumour microenvironment, particularly focusing on how cancer-associated fibroblasts, immune cells, stromal cells, and endothelial cells alter the tumour microenvironment to regulate tumour cell function.

Our studies primarily used immunocompromised mice to study the role of dopamine signalling in human SCLC progression, and the inability to model the contributions of the immune system in vivo is a limitation of our work. To better understand how D₂R agonist treatment affects immunoregulation within the tumour microenvironment, future research is needed using well-established genetically engineered mouse models of human SCLC. The Rb1 and Trp53 (RP) model involves using immunocompetent mice with conditional mutant alleles for Rb1 and Trp53 that are flanked by lox sites (Rb1^flox/flox;Trp53^flox/flox mice) that receive intratracheal administration of an adenoviral vector expressing Cre recombinase under the control of a CMV promoter (Ad-CMV-Cre), resulting in loss of function of Rb1 and Trp53 in the lungs for SCLC initiation [86]. A limitation of the RP model is that only 1–5 tumours develop over a long median latency period of 210 days [87]. Consequently, a newer model incorporating deletion of a third gene was developed, the Rb1;Trp53;Rbl2 (RPR2) model, which also uses delivery of Ad-CMV-Cre to the lungs to induce recombination and deletion of the floxed alleles [88]. The RPR2 mice develop approximately 10-20 times more tumours in half the time as RP mice. The ASCL1 transcription factor is highly expressed in the tumours of RP and RPR2 mice [87], whereas the Rb1/Trp53/Myc^T58A (RPM) model is a good option for modelling NEUROD1-high SCLC [89, 90]. Recently, CRISPR/Cas9-driven models of human SCLC have been introduced as alternatives to Cre-driven models. For example, LSL-Cas9 mice have been used to generate a model similar to RPR2 through the use of a single adeno-associated virus encoding sgRNAs against Rb1, Trp53, and Rbl2 delivered intratracheally and an EFS promoter driving Cre recombinase [91]. The CRISPR/Cas9 method for creating immunocompetent genetically engineered mouse models of human SCLC has emerged as a straightforward, cost-effective, and easily adaptable approach to better understand how D₂R agonist treatment affects immunoregulation within the tumour microenvironment as well as answer other questions focused on the role of immune system function in SCLC biology.

Recent pursuits to better understand the molecular basis of chemotherapy-refractory SCLC progression have primarily focused on the contributions of tumour cell–intrinsic factors. Here, we highlight the importance of cancer cell–extrinsic regulation of the tumour microenvironment by demonstrating that activation of D₂R signalling in tumour-associated endothelial cells by D₂R agonists inhibits tumour angiogenesis and reduces chemotherapy-refractory SCLC growth. In accordance with our findings, dopamine, upon binding to D₂R, reduces stress-mediated ovarian cancer growth by inhibiting tumour angiogenesis and stimulating tumour cell apoptosis [92]. Although it is well known that anti-angiogenesis therapies disrupt the tumour vasculature, the D₂R agonist-mediated anti-angiogenesis process can also transiently “normalize” the abnormal structure and function of tumour vasculature to make it more efficient for oxygen and drug delivery [70], which was clearly seen in our preclinical models (Figs. 1–4). Similar effects of dopamine on increased uptake of 5-fluorouracil have also been reported in human HT29 xenograft mouse models of colorectal cancer [93]. Our results are supported by the reported anti-tumour effects of endothelial D₂R activation by dopamine in various solid tumours [49, 52, 60, 84]. We speculate that D₂R agonists may increase transient normalization of tumour vessels, thereby producing a temporary increase in oxygen, alleviating hypoxia, and increasing the efficacy of conventional chemotherapies. Indeed, our findings demonstrate that treating human endothelial cells exposed to hypoxic conditions with D₂R agonist results in decreased HIF1α protein expression, suggesting that D₂R agonists reduce hypoxia (Fig. 2i-j). Furthermore, we speculate that D₂R agonist treatment may reduce immunosuppression within the tumour microenvironment, based upon our prior studies demonstrating that D₂R agonist treatment reduces tumour-infiltrating myeloid-derived suppressor cells in NSCLC [60]. Inhibition of VEGF signalling has been shown to stimulate CD4⁺ and CD8⁺ T cell activation and tumour infiltration, thereby reprogramming the tumour microenvironment [94]. Correspondingly, human CD8⁺ T cells exposed to conditioned medium from SCLC PDX organoids treated with D₂R agonist produced substantially greater amounts of cytotoxic granule granzyme B relative to CD8⁺ T cells grown in vehicle-treated PDX conditioned medium (Fig. 5c-d). In conjunction with immune checkpoint inhibitors, anti-angiogenic drugs can sensitize cancer cells to treatment. For instance, a combination of anti-VEGFR2 and anti-PD-L1 antibodies induced formation of high endothelial venules that facilitated enhanced infiltration and activity of cytotoxic lymphocytes and tumour cell death in RT2-PNET (Rip1-Tag2) pancreatic neuroendocrine cancer and PyMT (polyoma middle T oncoprotein) breast cancer [95]. While future research is necessary, D₂R agonist treatment likely helps reprogramme the immunosuppressive tumour microenvironment to enhance immune responses through reduction of tumour-infiltrating myeloid-derived suppressor cells and stimulation of T cells, making the tumour highly susceptible to enhanced anti-cancer responses through immune checkpoint inhibitors like anti-PD1/CTLA4.

Methods

Cell culture

The human SCLC cell line DMS-53 was purchased from The European Collection of Authenticated Cell Cultures (Salisbury, UK) via Sigma-Aldrich. This cell line was maintained in RPMI-1640 medium (Corning) supplemented with 10% foetal bovine serum (FBS; Millipore) and antibiotics. Human embryonic kidney (HEK) 293T cells purchased from American Type Culture Collection were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Corning) supplemented with 10% FBS (Millipore), 1% penicillin/streptomycin antibiotics (Corning), and 25 μg/mL plasmocin (Invivogen). Human umbilical vein endothelial cells (HUVEC; Lonza) were cultured in endothelial cell growth basal medium (EBM; CC-3121, Lonza) supplemented with the Endothelial Cell Growth Medium SingleQuots (CC-4143, Lonza). HUVEC of passages four to five were used and cultured in plates coated with bovine collagen type I protein (354231, Corning). All cell lines were grown in a humidified chamber at 37 °C supplied with 5% CO₂. To induce hypoxia, HUVEC were placed in a 1% O₂ hypoxia chamber for 6 h or treated with 150 μM CoCl₂ to cells for 24 h. Cell lines were authenticated by their source at the time of purchase and were subsequently routinely authenticated via morphologic inspection.

Organoid culture

SCLC PDXs were suspended in growth factor–reduced Matrigel (Corning) and plated at a density of 10⁵ cells into 50 µL droplets in each well of a 24-well plate. Once the droplets solidified, complete growth medium consisting of DMEM supplemented with 10% FBS, 1% penicillin/streptomycin antibiotics (Corning), and 25 µg/mL Plasmocin (Invivogen) was added to the PDXs. Cells were grown in a humidified chamber at 37 °C supplied with 5% CO₂ for 14 days. On the 14th day, medium from each well was removed, and the Matrigel domes were dissolved by adding dispase (1 U/mL) containing cold complete medium. Following incubation at 37 °C for 90 min with gentle agitation, the cells were centrifuged at 433 × g for 5 min. TrypLE (Gibco) was added to the pellet and incubated at 37 °C for 10 mins. Following centrifugation at 433 × g for 5 min, the cells were replated at a density of 10⁴ cells per well of a 96-well plate in Matrigel and topped off with 100 µL media.

Cleaved caspase-3 assay

Collagen at a concentration of 20 µg/mL was used to coat 10-cm plates. HUVEC were seeded at a density of 6 × 10⁵ cells and were grown to 70% confluency. Quinpirole and cabergoline were added at concentrations of 50 µM and 100 µM, respectively. After 72 h, conditioned media were collected from the cells, filtered, and added to PDXs along with the Caspase-3 green dye (Sartorius) and BioTracker nuclear red dye (Sartorius). The plates were placed in IncuCyte for up to 96 h with images taken every 24 h.

Flow cytometry

T cells: 96-well plates were coated with 10 μg/mL anti-CD3 antibody (BD Pharmigen, cat no.: 567118) and incubated at 37 °C for 3 h. Frozen human CD8⁺ T cells (Stem Cell Technology, cat no.: 70027) were thawed, washed with PBS, and suspended in TCM medium, containing RPMI, 10% FBS, 1% Pen-Strep, 10 mM HEPES, 0.05 mM beta-mercaptoethanol, IL-2 (Stem Cell Technology, cat no.: 78036.1; 100 IU/mL), and CD28 (BD Pharmigen, cat no.: B567117; 5 μg/mL) and seeded at a density of 1.5 × 10⁵ cells per well. Human SCLC PDXs were cultured in Matrigel as described above. Conditioned media from the SCLC PDXs was harvested after 72 h of treatment with either vehicle or 50 µM quinpirole and placed on the cultured human CD8⁺ T cells at a volume of 100 µL per well. After about 72 h, T cells were collected and transferred to a U-bottom plate and centrifuged at 1500 rpm for 5 min and washed one time with PBS before selection for live cells using Fixable Viability Dye eFluor™ 780 (Invitrogen, cat no.: 65-0865-18; 0.5 µl/mL in PBS) was performed at room temperature for 15 min. After live/dead staining, medium was added, the suspended cells were centrifuged, the supernatant was discarded, and cells were fixed using the Mouse FoxP3 buffer set (BD Biosciences, cat no.: 560409) at 4 °C for 15 minutes. PBS was added to the fixed cells and centrifuged, then permeabilized (Invitrogen, cat no.: 00833356) at 4 °C for 30 min. Following centrifugation, primary antibodies (anti-human CD8, eBioscience, cat no.: 56-0088-42; anti-human granzyme B-PE, BD Pharmigen, cat no.: 12889642; anti-human PD-1-FITC, BD Pharmigen, cat no.: 561035) were added to the cells at a 1:50 ratio each and incubated for 1 h at 4 °C. After centrifugation, cells were washed in PBS (150 µl) and transferred to FACS tubes.

SCLC PDX organoid-derived cells: SCLC PDX organoid cells were harvested from the Matrigel droplets using Dispase (Stem Cell Technology, Cat No.:7913; 1 U/mL) and agitated on a shaker at 37 °C for 10 min, followed by the addition of TrypLE (ThermoFisher, cat no.: 12604013). Once Matrigel was dissolved, cells were resuspended in FACS Buffer containing 2.5% FBS in PBS and passed through a cell strainer (Falcon, cat no.: 352235). Cells were washed, centrifuged, counted, and selection for live cells was performed using Fixable Viability Dye eFluor™ 780 for 15 min at room temperature. After washing, the cells were fixed for 15 min at 4 °C. Primary antibody (anti-human PD-L1-APC, BD Pharmigen, cat no.: 568316) was added at a 1:50 ratio and incubated at 4 °C for an hour, followed by washing and transfer to FACS tubes.

An LSRFortessa™ X-20 (BD Biosciences) benchtop flow cytometer was used to acquire data. FlowJo software was used for the analysis of the data. Fluorescence minus one (FMO) controls were used to set gates for each of the fluorescence channels.

Generation of stable cell lines

Lentiviral pGIPZ-shD₂R plasmid designed to silence D₂R expression and control pGIPZ-shScramble plasmid were purchased from the University of Minnesota Genomics Center. Transient transfection of shRNA plasmids along with their corresponding packaging plasmids was performed in 293T cells using Effectene transfection reagent (Qiagen) in accordance with the manufacturer’s protocol. Lentivirus was collected from the cell culture medium at 48 and 72 h after transfection. Collected medium was then passed through a 0.45-µm syringe filter (Millipore) to remove cell debris. One-third volume of Lenti-X concentrator (Takara) was added to the cell culture medium and incubated overnight at 4 °C. Precipitated lentivirus was collected from the cell culture medium after a brief centrifugation at 2000 × g for 1 h at 4 °C. HUVEC were transduced twice with concentrated lentivirus diluted in fresh EBM containing 10 µg/mL Polybrene (Millipore). HUVEC were grown in EBM containing 2 µg/mL puromycin (Sigma) for 72 h to select positive clones that were resistant to puromycin and expressed GFP. The knockdown of D₂R was confirmed by western blotting.

Human DMS-53 SCLC cells were transduced with retrovirus containing luciferase genes. Briefly, MSCV Luciferase PGK-hygromycin plasmid was obtained as a gift from Dr. Scott Lowe through Addgene (https://www.addgene.org/18782/). Retroviral luciferase plasmids and their corresponding packaging plasmids were transfected in 293T cells using Effectene transfection reagent (Qiagen) in accordance with the manufacturer’s protocol. Retro-X concentrator (Takara) was used to concentrate retrovirus using cell culture media collected at 48 h and 72 h after transfection. Concentrated retrovirus diluted in fresh RPMI 1640 medium containing 10 µg/mL Polybrene (Millipore) was added to the DMS-53 cells. Seventy-two hours after transduction, 5 µg/mL hygromycin (Sigma-Aldrich) was added to the cell culture medium, and DMS-53 cells were incubated for additional 72 h. Luciferase-labelled stable human DMS-53 cells were confirmed based on the expression of luciferase and sustained growth in the presence of hygromycin.

Antibodies and drugs

Primary antibodies were used in western blotting experiments to detect expression of D₂R (Abcam, cat no.: ab85367; 1:1000) and α-tubulin (Santa Cruz Biotechnology, cat no.: sc-5386; 1:500). Horseradish peroxidase (HRP)-conjugated anti-rabbit (cat no.: 7074; 1:5000) and anti-mouse (cat no.: 7076; 1:5000) secondary antibodies were purchased from Cell Signalling Technology. Primary antibodies against CD31 (Abcam, cat no.: ab28364; 1:100), D₂R (Santa Cruz Biotechnology, cat no.: sc-5303; 1:50), and Ki-67 (Cell Signalling Technology, cat no.: 12202; 1:200) were used for immunofluorescence staining. Fluorescence-conjugated secondary antibody staining was performed in the dark using corresponding Alexa Fluor 488–conjugated anti-rabbit antibody (Molecular Probes; Cat no.: A11008; 1:400) and Alexa Fluor 594–conjugated anti-mouse antibody (Molecular Probes; Cat no.: A11005; 1:400). A mouse monoclonal D₂R antibody (Santa Cruz Biotechnology, cat no.: sc-5303, 1:50) was used for immunohistochemistry. Cisplatin (232120), etoposide (E1383), quinpirole (Q102), and cabergoline (C0246) were purchased from Sigma-Aldrich.

Immunoblotting

Stable HUVEC were lysed in SDS containing 2% 1× Laemmli buffer (Bio-Rad) supplemented with 5% 2-mercaptoethanol (Sigma). Cells were immediately scraped off the cell culture plate and transferred to microcentrifuge tubes for boiling. Cell lysates were heated to 95–100 °C for 5 min and allowed to cool at room temperature for 10 min. Later, 20 µL of cell lysates were separated via 4–20%–gradient SDS-PAGE (Bio-Rad) and transferred to polyvinyl difluoride (PVDF) membranes (Millipore). Following the completion of the protein transfer process, membranes were blocked with 5% bovine serum albumin (Sigma-Aldrich) diluted in 1× Tris-buffered saline (Growcells) containing 0.1% Tween-20 detergent (Fisher Scientific). Membranes were then incubated with primary antibodies overnight at 4 °C and corresponding secondary antibodies for 2 h at room temperature. Antibody-reactive protein bands on the membranes were detected in the dark using HRP-reactive chemiluminescence substrate (Thermo Fisher Scientific).

Immunofluorescence

The tumours of mice bearing chemonaïve and chemoresistant human SCLC PDXs were harvested and fixed in neutral buffered 10% formalin (Sigma) at room temperature for 24 h before processing, embedding in paraffin, and sectioned. Tissues sections were deparaffinized, rehydrated, heat-retrieved with 1× Rodent Decloaker buffer (Biocare Medical, RD913L), and then incubated in Rodent Block M (Biocare Medical, RBM961G) for 1 h at room temperature to eliminate non-specific mouse IgG staining. Tissue sections were then incubated with primary antibodies overnight at 4 °C followed by fluorophore-conjugated secondary antibodies for 1 h at room temperature. For the TUNEL staining experiments, tissue sections were rinsed with 1× PBS (Sigma) and subjected to fluorescence-based TUNEL staining following the manufacturer’s protocols (Promega, G3250). Briefly, the tissue specimens were incubated with 1× equilibration buffer for 5 min and then with the reaction buffer containing recombinant terminal deoxynucleotidyl transferase (rTdT) enzyme and nucleotide mix for 1 h at 37 °C. The reaction was terminated with 2× saline sodium citrate buffer (SSC). Cell nuclei were counterstained with 4’, 6- diamidino-2-phenylindole, dihydrochloride (DAPI) in ProLong^® Gold Antifade Reagent (Cell Signalling Technology). Images were captured using a Zeiss Apotome.2 microscope (20× objective, 0.75 NA) and processed with ZEN microscope software (Zeiss). Double-positive cells (i.e., either TUNEL⁺-CD31⁺ or D₂R⁺-CD31⁺) per visual field were counted with the cell counter tool available in ImageJ software. The bar graphs were generated by dividing the number of double-positive cells by the number of CD31⁺ blood vessels. The percentage of Ki-67⁺ cells was determined by following the equation: percentage of Ki-67⁺ cells = (Count of Ki-67⁺ cells / Count of DAPI⁺ cells) × 100.

Immunohistochemistry

We obtained chemonaïve (i.e., before the start of chemotherapy) and matched chemoresistant (i.e., recurrence of tumour after chemotherapy) human SCLC whole-tissue specimens from nine SCLC patients at Mayo Clinic in Rochester, MN, in accordance with institutional review board–approved protocols. Formalin-fixed, paraffin-embedded whole tissues were serially sectioned, mounted on glass slides, and immunostained using primary and HRP-conjugated secondary antibodies. The staining was performed by using a Bond Autostainer (Leica), and the sections were incubated in hematoxylin (IHC World) to detect nuclei. A pulmonary pathologist (Y-CL) scored each lung tumour specimen of each group by observing the prevalence of D₂R staining in the tumour-associated endothelial cells using a light microscope.

In vivo orthotopic lung cancer model

Eight- to ten-week-old pathogen-free SCID/NCr mice (Strain Code: 561) purchased from Charles River Laboratories were bred and maintained in accordance with protocols approved by the University of Minnesota Institutional Animal Care and Use Committee (IACUC). One million luciferase-labelled human DMS-53 SCLC cells suspended in 80 µL PBS (Corning) and high-concentration Matrigel (Corning; Cat. no.: 354248) were orthotopically injected into the left thoracic cavity of 8- to 12-week-old male and female mice anesthetized with pharmaceutical-grade ketamine (90–120 mg/kg) and xylazine (5–10 mg/kg). Bioluminescence imaging of mice anesthetized with isoflurane was performed on the indicated days using the IVIS^® Lumina^™ S5 high-throughput 2D optical imaging system (Perkin Elmer) to monitor lung tumour growth non-invasively.

Human SCLC PDX subcutaneous mouse model

Ten- to twelve-week-old male and female NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ mice (NSG, Jackson Laboratory, Strain Code: 5557) were bred, maintained in a temperature-controlled room with alternating 12-h light/dark cycles under specific pathogen-free conditions, and fed a standard diet in accordance with protocols approved by the University of Minnesota IACUC. Male and female mice anesthetized with pharmaceutical-grade ketamine (90–120 mg/kg) and xylazine (5–10 mg/kg) via intraperitoneal injection under laminar flow hood in a specific pathogen-free room were subcutaneously injected with 5 × 10⁶ human SCLC cells collected from either chemonaïve or chemoresistant human SCLC PDX tissue samples. Briefly, frozen vials of tissues from one chemonaïve and three chemoresistant SCLC PDX-bearing mice were transported to our laboratory overnight on dry ice from our collaborator at Memorial Sloan Kettering Cancer Center. Due to the small amount of material available, the entire tumour sample was first resuspended in 100 µL high-concentration Matrigel on ice and later injected subcutaneously in the flanks of anesthetized NSG mice. Tumour growth was monitored weekly. When tumour volume reached 200 mm³ in size, mice were randomly divided into treatment groups such that each group had similar mean tumour volumes. Tumour volume was measured every week using the formula (length × width²)/2. When the P₀ tumours reached 2000 mm³ in volume, the mouse was sacrificed, tumour tissue samples were finely minced with sterile razor blades under aseptic conditions, vigorously triturated in Accutase™ cell detachment solution (BD Bioscience, 561527), passed through a 70-µm filter (Corning), centrifuged at 433 × g for 5 min, and either cryopreserved in 85% RPMI-1640 (Corning)/10% FBS (Millipore)/5% DMSO (MP Biomedicals) for future use or suspended in 100 μL PBS and high-concentration Matrigel on ice (5 × 10⁶ cells per mouse) for immediate use. Cells were then subcutaneously injected into the right flank of NSG mice, and the mice were monitored for tumour growth. After establishment of palpable tumours (≥100 mm³), the mice were randomly divided into groups and administered drugs at the indicated doses. At the endpoint, the mice were euthanized by CO₂ asphyxiation followed by cervical dislocation. Extirpated tumours were photographed, weighed, and preserved in neutral buffered 10% formalin (Sigma) for immunofluorescence analysis.

Clinical workflow and patient selection

Patients who met the following criteria were enroled in this study: (1) pathologically confirmed advanced SCLC; (2) defined subtypes based on the expression of transcription factors ASCL1, NEUROD1, and POU2F3; (3) treatment with platinum-based chemotherapy in the first-line setting; and (4) available biopsied tumour samples before and after the chemotherapy. A 20% increase of tumour burden after the completion of chemotherapy was considered as disease progression in accordance with RECIST 1.1. Pathological diagnosis and staging were carried out according to the staging system of the 2021 International Association for the Study of Lung Cancer (9th edition). Written informed consent was obtained from all of the patients prior to inclusion in this study. The Mayo Clinic Institutional Review Board Committee approved this study.

Statistics

To compare differences between two groups, two-way unpaired t-tests were performed and values of P ≤ 0.05 were considered significant. A two-way analysis of variance (ANOVA) followed by Sidak’s test was used to determine statistically significant differences between multiple groups (greater than two). Data expressed as mean ± SEM are representative of at least three independent experiments. For most animal experiments, the number of animals per group was calculated based on a one-way ANOVA analysis to allow 90% power when the mean in the test group is 1.25 standard deviations higher or lower than the mean in the controls.

Author contributions

SKA conducted most of the in vitro cell line-based experiments, including immunofluorescence studies, generation of stable cell lines, immunoblotting, etc. AP performed the organoid culture and cleaved caspase 3 assays. SKA, AP, LW, and LHH managed the mouse colony and performed tumour studies in mice. SKA, AP, and LW conducted murine in vivo imaging, tumour measurements, and necropsy. PM and CMR generated and provided the patient-derived xenograft models of chemotherapy-refractory small cell lung cancer. MCA and Y-CL coordinated the acquisition of small cell lung cancer specimens from patients at Mayo Clinic in accordance with IRB-approved protocols. BAT captured images of the small cell lung cancer specimens that had been immunostained to detect D₂R protein. Y-CL led the immunostaining, imaging, pathological review, and analysis of patient-derived lung tumour specimens. AP and SSMF performed the flow cytometry-based immunology experiments. SKA, AP, LW, SSMF, VV, CMR, Y-CL, and LHH provided technical and scientific support. SKA, AP, and LHH performed experimental troubleshooting, reviewed relevant scientific literature, critically analyzed data, prepared figures, and wrote the manuscript. LHH conceived the aims, led the project, and acquired funding to complete the reported research.

Funding

This work was supported by a Research Scholar Grant RSG-21-034-01-TBG from the American Cancer Society and The Hormel Foundation to LHH. The generation of the patient-derived xenograft models of chemotherapy-refractory small cell lung cancer was supported by R35 CA263816 and U24 CA213274 to CMR.

Data availability

All data generated or analyzed during this study are included in this published article (and its supplementary information files).

Competing interests

The authors declare no competing interests.

Ethics approval and consent to participate

All studies described in this manuscript were performed in accordance with the relevant guidelines and regulations. For studies involving human subjects, written informed consent was obtained from each of the individual human subjects prior to inclusion in this study, and the Mayo Clinic Institutional Review Board Committee approved this study. All studies involving animals were approved by the University of Minnesota Institutional Animal Care and Use Committee and performed in accordance with these approved guidelines (Protocol: 2410-42469A).

Supplementary information

The online version contains supplementary material available at 10.1038/s41419-025-07693-y.