Optimization and development of a high-throughput TR-FRET screening assay for SLIT2/ROBO1 interaction

Abstract

The SLIT2/ROBO1 signaling axis plays a critical role in cell migration, angiogenesis, and immune regulation, contributing to tumor progression, metastasis, and therapy resistance. SLIT2 is highly expressed in various malignancies, where it promotes immune evasion by recruiting tumor-associated macrophages and disrupting vascular integrity, ultimately diminishing therapeutic efficacy. Beyond cancer, SLIT2/ROBO1 is implicated in neural development, fibrosis, and vascular remodeling, making it a potential but underexplored therapeutic target. However, no small-molecule inhibitors of SLIT2/ROBO1 interaction currently exist. Herein, we describe the development and optimization of a time-resolved fluorescence resonance energy transfer (TR-FRET) assay for high-throughput screening of small-molecule inhibitors targeting this pathway. Using recombinant SLIT2 and ROBO1, we established a robust assay that enables high-throughput screening (HTS) of chemical libraries of small molecules for SLIT2/ROBO1 inhibition. Screening a focused chemical library of protein-protein interaction (PPI) inhibitors identified SMIFH2 as a SLIT2/ROBO1 inhibitor, demonstrating its ability to disrupt the interaction in a dose-dependent manner. Our study introduces a novel screening platform for identifying small molecule inhibitors of SLIT2/ROBO1, laying the foundation for future drug discovery efforts aimed at targeting this signaling axis in cancer and other diseases.

Introduction

Slit guidance ligands (SLITs) are secreted proteins that regulate cell positioning during development by interacting with roundabout (ROBO) receptors [1]. In mammals, SLIT1, SLIT2, and SLIT3 bind the second leucine-rich repeat region (D2) to the Ig1 domain of ROBO1 and ROBO2 [2,3]. The SLIT2/ROBO axis plays essential roles in organogenesis, homeostasis, and cancer progression. Beyond cell migration, it influences proliferation, apoptosis, adhesion, and angiogenesis in normal and malignant cells. Additionally, SLIT2-ROBO signaling contributes to liver fibrosis via PI3K/Akt activation. By recruiting adaptor proteins to the ROBO cytoplasmic domain, SLIT2 modulates the cytoskeleton, affecting cell motility, adhesion, and growth [3–5].

In the nervous system, SLIT2/ROBO signaling regulates axonal repulsion [6,7], neuronal migration [8], and axon guidance [9]. In immunity, SLITs attract neutrophils but repel lymphocytes and dendritic cells [10–12]. In macrophages, they suppress macropinocytosis and cytotoxic polarization [13]. In endothelial cells, SLIT2-ROBO activation enhances angiogenesis in retinal and bone tissues by guiding tip cell migration and polarization [14–16]. SLIT2 plays a proangiogenic role in tumors [17], fostering tumor cell migration [18–20], metastasis [21], and therapy resistance [22], particularly in colorectal and pancreatic cancers and osteosarcoma. However, SLIT2/ROBO signaling also exhibits tumor-suppressive effects in lung and breast cancers [23–25]. In glioblastoma (GBM), conflicting evidence suggests both tumor--suppressive[26–28] and tumor--promoting roles [29,30], highlighting the pathway’s context-dependent functions.

Recent research has identified SLIT2/ROBO signaling as a novel immune evasion mechanism within the tumor microenvironment (TME) of GBM [31]. Elevated SLIT2 expression in both GBM patients and mouse models led to the accumulation of immunosuppressive tumor-associated macrophages (TAMs) and vascular dysmorphia [31]. Notably, silencing SLIT2 in glioma cells or systemically inhibiting SLIT2 using the SLIT2-trapping protein ROBO1Fc prevented TAM polarization toward a tumor-supportive phenotype and suppressed angiogenic gene expression [31]. These effects resulted in improved tumor vessel function and enhanced the efficacy of both chemotherapy and immunotherapy in GBM mouse models [31].

Notably, SLIT2 inhibition demonstrated a greater impact on angiogenesis and T cell response than previously explored therapeutic approaches aimed at modulating TAMs within the GBM TME [31]. The only known clinical trial (NCT03448692) involving a SLIT2/ROBO-targeting biotherapeutic protein (PF-06,730,512) was conducted for focal segmental glomerulosclerosis (FSGS) but was terminated in 2023 due to a lack of efficacy at safe doses [32,33]. Moreover, three ongoing clinical trials (NCT03940820, NCT03941457, and NCT03931720) are investigating chimeric antigen receptor-natural killer (CAR-NK) cell therapy targeting ROBO1 for solid tumors.

Compared to therapeutic proteins and other biologic-based therapies (e.g., ROBO1Fc and anti-ROBO antibodies), small molecules offer a lower risk of immunogenicity and improved management of adverse events (AEs) due to their pharmacokinetic flexibility [34–37]. Notably, their oral bioavailability and shorter half-lives reduce the likelihood of prolonged on-target immune-related AEs, making them a more flexible therapeutic option [38,39]. Despite the therapeutic relevance of SLIT2/ROBO1 signaling, no small-molecule inhibitors currently exist for this pathway, highlighting a critical gap in drug development. The challenges associated with biologics, including limited tissue penetration, complex manufacturing, and potential immunogenicity, further underscore the urgent need for small-molecule modulators [34–39]. Such inhibitors could provide a scalable, cost-effective strategy to disrupt SLIT2-mediated tumor progression and immune evasion while enhancing therapeutic efficacy in GBM and other malignancies. Developing small-molecule inhibitors for SLIT2/ROBO1 could open new avenues for targeted therapies, overcoming the limitations of biologic-based approaches. Herein, we report the first high-throughput screening (HTS) assay for SLIT2/ROBO1 inhibition. The implementation of this assay in screening chemical libraries of small molecules will enable the discovery of drug candidates targeting the SLIT2/ROBO1 interaction for cancer therapy.

In the context of available protein-protein interaction (PPI) detection technologies, our TR-FRET assay offers a robust, homogeneous, and miniaturized format ideal for early-stage screening of PPI modulators. Alternative platforms such as lysate-based TR-FRET assays can preserve native protein conformations and post-translational modifications, potentially enhancing physiological relevance [40]. Live-cell BRET assays provide dynamic and spatial information on PPIs in intact cells and are particularly useful for capturing transient or context-dependent interactions [41]. Additionally, ultra-high-throughput adaptations of PPI assays in 1536-well formats enable broad compound library screening at scale [42]. While each of these approaches offers unique advantages, our current TR-FRET platform was selected for its reproducibility, amenability to miniaturization, and compatibility with pilot screening. Future work will explore adaptation to higher-throughput formats and integration with orthogonal assay systems to validate hits in more physiologically relevant contexts.

Methods

Our approach for developing a time-resolved fluorescence resonance energy transfer (TR-FRET) assay to evaluate SLIT2/ROBO1 interaction is depicted in Fig. 1. For this assay, we sourced human recombinant SLIT2 with a C-terminal His-tag from Sino Biological (Cat. No. 11967-H08H) and the extracellular domain (ECD) of ROBO1 fused to the Fc region of human IgG1 from Sino Biological (Cat. No. 30073-H02H). The fluorescent tags utilized included anti-His monoclonal antibody (mAb) d2-conjugate (Cat. No. 61HISDLF) and antihuman IgG polyclonal Ab (pAb) Tb-conjugate (Cat. No. 61HFCTAF), both obtained from Cisbio (a PerkinElmer company). The binding validation of ROBO1 to SLIT2 (Fig. S1) as well as the optimization of the TR-FRET assay (Figs. S2–S7 and Tables S1 and S2) are detailed in the Supporting Information. The results shown in Figs. S2–S7 are based on three replicates (n = 3) in single runs.

Fig. 1.

Our established workflow for the SLIT2/ROBO1 TR-FRET assay.

TR-FRET signal detection was performed using a Tecan Infinite M1000 Pro plate reader under the following conditions: donor excitation at 340 nm (bandwidth: 20 nm) with an emission at 620 nm (bandwidth: 10 nm), and acceptor excitation at 340 nm (bandwidth: 20 nm) with an emission at 665 nm (bandwidth: 10 nm). Each measurement included 100 flashes per well, with an integration time of 500 μs and a lag time of 60 μs. stock solutions of test compounds (prepared in 0.1 % DMSO and assay buffer) or vehicle control were added to medium-binding white assay plates (Greiner, Cat. No. 784,075) at a final concentration of 100 μM (n = 3). Each well received 2 μL of the test compound solution (or vehicle control), followed by 18 μL of the assay mixture. To prepare the assay mixture (total volume: 18 μL), SLIT2 (4 μL total volume) and ROBO1 (4 μL total volume) were each used at a final concentration of 5 nM, while antihuman IgG pAb Tb-conjugate (5 μL total volume) and anti-His mAb d2-conjugate (5 μL total volume) were included at 0.25 nM and 2.5 nM, respectively. Stock solutions of the fluorescent tags were prepared in PPI Tb detection buffer from Cisbio (Cat. No. 61DB10RDF). The assay plates were incubated at room temperature for 1 hour before measurement.

The chemical library used in screening was procured from MedChem Express (Cat. No. HY-L109). This is the protein-protein interaction inhibitor library composed of 609 compounds. The TR-FRET signal was calculated as the ratio of fluorescence intensity at 665 nm to that at 620 nm, multiplied by 100. Each plate contained control wells (n = 3, three replicates in a single run) treated with 0.1 % DMSO and additional wells with all assay components except His-tagged SLIT2 (n = 3, three replicates in a single run) to serve as background controls. For hit identification, Compounds exhibiting at least 50 % inhibition of the TR-FRET signal were classified as hits. To account for potential assay interference, an additional analysis of donor fluorescence was performed. Any compounds that altered donor fluorescence in a manner consistent with FRET/donor ratio changes, indicative of fluorescence attenuation, were excluded from the final hit selection.

Results and discussion

Building on our previous efforts to develop TR-FRET assays for HTS of chemical libraries of small molecules [43,44], we designed a TR-FRET assay to evaluate the interaction between SLIT2 and ROBO1 (Fig. 2). TR-FRET relies on two fluorophores—referred to as a donor and an acceptor—that facilitate energy transfer when positioned in close proximity [42]. In this protein-protein interaction assay, one protein is conjugated (directly or indirectly) with the donor, while the other is linked to the acceptor [45]. When the two proteins bind, the donor fluorophore comes into proximity with the acceptor, enabling energy transfer upon donor excitation, which generates a measurable signal. Disrupting this interaction with inhibitory molecules results in a diminished TR-FRET signal.

Fig. 2.

Schematic representation of the TR-FRET assay designed for identifying small molecules that disrupt the SLIT2/ROBO1 interaction.

Using biolayer interferometry (BLI), the binding affinity of hFc-ROBO1 to His-tagged SLIT2 (from SinoBiological, as used in this assay) was measured at 42.8 nM (Fig. S1). To determine the optimal donor-acceptor pair for detecting SLIT2/ROBO1 binding, we tested a combination of various FRET donors and acceptors (Figs. S2 and S3). We have selected the concentrations of both SLIT2 and ROBO1 in our assay based on the outcome of 2D titration of SLIT2 and ROBO1 (Table S1). The assay performed optimally at final concentrations of 5 nM for both human Fc-tagged ROBO1 and His-tagged SLIT2. The strongest signal was obtained using antihuman IgG pAb Tb-conjugate as the donor and anti-His mAb d2-conjugate as the acceptor, achieving a 1:10 ratio. The assay was miniaturized for a 384-well plate format, and key parameters—including buffer composition, plate type, reagent addition sequence, and incubation duration—were systematically optimized. The selection of the PPI Tb detection buffer from Cisbio (Cat. No. 61DB10RDF) as the assay buffer was based on an evaluation of different buffers to optimize the S/B ratio (Table S2). Assessment of the impact of the incubation time on the TR-FRET signal revealed that maximal signal was achieved after 1 hour of incubation (Fig. S4). Moreover, DMSO tolerance screening revealed that the TR-FRET signal was stable and robust up to 4.0 % (v/v) DMSO (Fig. S5). Under these optimized conditions, the signal-to-background (S/B) ratio for the SLIT2/ROBO1 TR-FRET assay was determined to be 8.21. Notably, we demonstrated that untagged ROBO1 (Fig. S6) and SLIT2 (Fig. S7) can induce a dose-dependent decrease in the TR-FRET signal.

The titration of d2-conjugate-labeled SLIT2 with Tb-labeled ROBO1 demonstrated FRET efficiency, as evidenced by the hyperbolic relationship between the TR-FRET ratio and acceptor concentration (Fig. 3). Small molecules capable of disrupting the SLIT2/ROBO1 interaction would lead to a reduction in the TR-FRET signal. Notably, the introduction of an anti-SLIT2 mAb (Cat. No. AB7665 from abcam), resulted in a concentration-dependent decrease in the TR-FRET ratio (Fig. 4). Dose-response analysis of the anti-SLIT2 mAb in the TR-FRET-based assay determined a half-maximal inhibitory concentration (IC₅₀) of 5.72 ± 0.94 nM (Fig. 4). The assay’s suitability for HTS was confirmed by a mean Z’ factor of 0.61 ± 0.03, indicating a robust and high-quality screening platform. The Z’ factor quantifies assay performance by comparing data variability (standard deviation) to the dynamic range, defined by the difference in TR-FRET signal between positive and negative controls [46]. The Z’ factor for the SLIT2/ROBO1 TR-FRET assay was determined from three independent experiments.

Fig. 3.

Variation in the TR-FRET ratio in the SLIT2/ROBO1 as a function of the acceptor-to-donor ratio. Error bars indicate the standard deviation (n = 3).

Fig. 4.

Dose-response curve depicting anti-SLIT2 mAb (AB7665) binding in the SLIT2/ROBO1 assay. Error bars indicate the standard deviation (n = 3).

To improve screening efficiency and maximize the likelihood of identifying small-molecule inhibitors of SLIT2/ROBO1 interaction, we utilized a focused PPI inhibitor library from MedChem Express (Cat. No. HY-L109). This library comprises 609 structurally diverse and medicinally active compounds designed to disrupt key PPIs implicated in various diseases. The selection includes inhibitors targeting well-characterized interactions such as MDM2-p53, Keap1-Nrf2, PD-1/PD-L1, and Myc-Max, ensuring a broad representation of scaffolds with established drug-like properties. Notably, many compounds in this library have demonstrated bioactivity and favorable safety profiles in preclinical studies. The inclusion of cell-permeable compounds further enhances the translational potential of identified hits, facilitating future medicinal chemistry optimization and biological validation in cellular and in vivo models. We performed the screening of this focused chemical library at a final concentration of 100 μM. As shown in Fig. 5A, this screening identified 8 potential hits that inhibited the TR-FRET signal by more than 50 %. The screening of the MedChem Express library exhibited excellent reproducibility, as illustrated in Fig. 5B

Fig. 5.

(A) Scatterplot showing single-dose inhibition data screened at 100 μM of the MedChem Express protein-protein interaction inhibitor library using our SLIT2/ROBO1 TR-FRET assay. (B) Correlation plot of the screened chemical library duplicate runs in the SLIT2/ROBO1 TR-FRET assay.

Screening the 609-compound MedChem Express library led to the identification of 8 preliminary hits that demonstrated at least 50 % inhibition of the SLIT2/ROBO1 TR-FRET signal (Fig. 6A). To validate these candidates, we conducted dose-response studies, ultimately identifying SMIFH2 (Fig. 6B) as the only compound exhibiting a concentration-dependent inhibition of the SLIT2/ROBO1 interaction. Dose dependent screening revealed that SMIFH2 possess an IC₅₀ value of 83.5 ± 7.14 μM (Fig. 6C), highlighting its potential for future hit-to-lead optimization studies. SMIFH2 is a small-molecule inhibitor of formin homology 2 (FH2) domains, originally identified for its ability to disrupt actin polymerization by targeting formin-mediated cytoskeletal dynamics [47]. In order to confirm the purity and chemical identity of SMIFH2, the ¹HNMR and LC-MS of SMIFH2 are shown in Figs. S8 and S9. To further validate SMIFH2 as an inhibitor of the SLIT2/ROBO1 interaction, we performed a competitive BLI assay, in which human ROBO1 (hFc Tag) was loaded on proA Biosensor and exposed to human SLIT2 preincubated with increasing concentrations of SMIFH2. A dose-dependent reduction in SLIT2 binding signal was observed (Fig. S10), confirming that SMIFH2 disrupts the SLIT2/ROBO1 interaction. This label-free approach complements the TR-FRET results and provides an orthogonal validation for SMIFH2 as a SLIT2/ROBO1 inhibitor.

Fig. 6.

(A) The screening funnel for testing the MedChem Express library using our SLIT2/ROBO1 TR-FRET assay. (B) Chemical structure of SMIFH2. (C) Dose-dependent screening of SMIFH2 in the SLIT2/ROBO1 TR-FRET assay. Error bars indicate the standard deviation (n = 3).

In summary, this study introduces a robust TR-FRET assay for HTS of small-molecule inhibitors targeting SLIT2/ROBO1 interaction. The development of such an assay is a crucial step toward addressing the current lack of small molecules for modulating this pathway. By enabling the identification of first-in-class small-molecule inhibitors, this platform lays the foundation for further medicinal chemistry optimization and preclinical evaluation. Future studies will explore the therapeutic potential of SMIFH2 and its derivatives, assessing their efficacy in disrupting SLIT2-mediated immune suppression and angiogenesis in relevant cancer models. Additionally, SLIT2/ROBO1 inhibitors could be investigated as combination therapies to enhance the effectiveness of existing FDA-approved immunotherapies, particularly in GBM and other aggressive malignancies where SLIT2-driven immune evasion contributes to poor treatment outcomes. Given the assay’s robustness, reproducibility, and miniaturized format, it is well-suited for automation and large-scale HTS, which will be pursued in future studies to identify potent modulators of the SLIT2/ROBO1 interaction.

Supplementary Material

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.slasd.2025.100240.