Microscopy-Based Automated Live Cell Screening for Small Molecules That Affect Ciliation

Abstract

The primary monocilium, or cilium, is a single antenna-like organelle that protrudes from the surface of most mammalian cell types, and serves as a signaling hub. Mutations of cilia-associated genes result in severe genetic disorders termed ciliopathies. Among these, the most common is autosomal dominant polycystic kidney disease (ADPKD); less common genetic diseases include Bardet-Biedl syndrome, Joubert syndrome, nephronophthisis, and others. Important signaling cascades with receptor systems localized exclusively or in part at cilia include Sonic Hedgehog (SHH), platelet derived growth factor alpha (PDGFRα), WNTs, polycystins, and others. Changes in ciliation during development or in pathological conditions such as cancer impacts signaling by these proteins. Notably, ciliation status of cells is coupled closely to the cell cycle, with cilia protruding in quiescent (G0) or early G1 cells, declining in S/G2, and absent in M phase, and has been proposed to contribute to cell cycle regulation. Because of this complex biology, the elaborate machinery regulating ciliary assembly and disassembly receives input from many cellular proteins relevant to cell cycle control, development, and oncogenic transformation, making study of genetic factors and drugs influencing ciliation of high interest. One of the most effective tools to investigate the dynamics of the cilia under different conditions is the imaging of live cells. However, developing assays to observe the primary cilium in real time can be challenging, and requires a consideration of multiple details related to the cilia biology. With the dual goals of identifying small molecules that may have beneficial activity through action on human diseases, and of identifying ciliary activities of existing agents that are in common use or development, we here describe creation and evaluation of three autofluorescent cell lines derived from the immortalized retinal pigmented epithelium parental cell line hTERT-RPE1. These cell lines stably express the ciliary-targeted fluorescent proteins L13-Arl13bGFP, pEGFP-mSmo, and tdTomato-MCHR1-N-10. We then describe methods for use of these cell lines in high throughput screening of libraries of small molecule compounds to identify positive and negative regulators of ciliary disassembly.

Keywords: ADPKD; aurora kinase A; ciliary disassembly; drugs; heat shock protein 90; high content imaging; screening; targeted therapy.

FIGURE 1

Proteins regulating ciliary assembly and disassembly. Mechanisms of ciliary dynamics are complex and require the orchestrated action of numerous proteins. Some examples of proteins regulating ciliary protrusion (red) and resorption (green) are shown. The Rab GTPases (e.g., Rab8 and Rab11) are active at the early stages of cilia formation, promoting targeted vesicular trafficking required for the primary ciliogenesis (Knodler et al., 2010). Assembly of the primary cilium is highly dependent on the proper functioning of the IFT – a bidirectional movement throughout the axonemal microtubules which is essential for cilium elongation and maintenance. One of the key proteins controlling IFT-dependent ciliogenesis is IFT protein 88 homolog (IFT88); mutations inactivating IFT88 to an abnormally short cilia phenotype and development of polycystic kidney disease (Pazour et al., 2000). The movement of IFT particles along the microtubules is supported by a group of microtubule-based motors, which include kinesin family proteins (KIFs), providing anterograde transport (toward the cilia tip) and dynein family proteins (DYNs), providing retrograde transport (toward the cilia base). Cilia formation is also controlled by NIMA-related kinases (NRKs), namely, Nek1 and Nek4, which are often determined as mitotic kinases (Mahjoub et al., 2005; Coene et al., 2011). Additionally, ciliogenesis is regulated by small GTPases, including Arl13b, which localizes exclusively to the primary cilium, where it stabilizes the association between IFT-A and IFT-B complexes (Li et al., 2010).

FIGURE 2

Rationale for choosing Arl13b, SMO, and melanin-concentrating hormone receptor 1 (MCHR1) as cilia-targeting moieties. Mechanism of MCHR1 action (A) and a schematic (B) for tdTomato-MCHR1-N-10 plasmid. MCHR1 is a protein which belongs to a class A of G-protein-coupled receptors (GPCRs) and localizes in neuronal cilia in the hypothalamus (Green and Mykytyn, 2010). The translocation of MCHR1 to the ciliary membrane by the BBSome complex (Nachury, 2018) allows it to play an important role in controlling energy homeostasis potentially through modulating cAMP signaling (Schou et al., 2015). Mechanism of SMO action (C) and a schematic (D) for pEGFP-mSmo plasmid. Hedgehog (HH) pathway is uniquely associated with the primary cilium. Two key receptors for this pathway – Smoothened (SMO) and Patched (PTCH1) – localize within the ciliary membrane depending on the pathway’s activation stage. In the absence of HH ligands, PTCH1 accumulates at the ciliary membrane preventing translocation of Smo to the cilia. When the ligand binds PTCH1, it leaves the cilia allowing SMO to translocate and activate the pathway. Mechanism of Arl13b action (E) and a schematic (F) for L13-Arl13bGFP plasmid. Arl13b is a small guanosine triphosphate (GTPase) which localizes exclusively to the primary cilium, where it controls ciliogenesis through stabilizing the coordination between IFT-A and IFT-B complexes (Li et al., 2010). Additionally, Arl13b regulates translocation of SMO receptor to the cilia (Larkins et al., 2011).

FIGURE 3

Confocal images of tTERT-RPE1 cells expressing L13-Arl13bGFP, pEGFP-mSMO, and tdTomato-MCHR1-N-10. Top, representative images of RPE1-Arl13b-EGFP (green) pEGFP-mSMO (green), and tdTomato-MCHR1-N-10 (red) in cells also visualized with antibody to acetylated a-tubulin (red with Arl13b and SMO; green with MCHR1) and DAPI (blue). Images compare ciliation in cells with or without serum treatment. Bottom, quantification of data for L13-Arl13bGFP, pEGFP-mSmo, and tdTomato-MCHR1-N-10 from 250 cells, from three representative experiments; results following treatment with an inhibitor (alisertib) and a stimulator (ganetespib) of ciliary disassembly are also shown. Data are presented as mean ± SEM, ^∗p ≤ 0.05, ^∗∗p ≤ 0.001, ^∗∗∗p ≤ 0.0001 as compared to control.

FIGURE 4

Automated screening workflow and representative images from the ImageXpress microscope. (A) Steps for plating, imaging, drug/serum addition, and analysis. (B) Epifluorescent images acquired by ImageXpress from first (no serum) and second imaging (after addition of serum control), for the hTERT-RPE1-Arl13b-EGFP (green) hTERT-RPE1-pEGFP-mSMO (green), and hTERT-RPE1-MCHR1-tdTomato (red) models. Note higher background in hTERT-RPE1-MCHR1-tdTomato cells.

FIGURE 5

Quantification of ciliation based on plating density. hTERT-RPE1-Arl13b-EGFP (A) and hTERT-RPE1-SMO-EGFP (B) cells were plated at concentrations shown under conditions described in text, and ciliation assessed after 48 h. Quantitation is based on use of “vesicle total count per cell” in the Transfluor module. Data are presented as mean ± SEM.

FIGURE 6

ImageXpress images of cilia, emphasizing thresholding issues. Representative images of the three models of ciliation evaluated in this method, and typical image segmentation by MetaExpress software modules Transfluor (TF) and Multiwavelength Scoring (MWS). (A) hTERT-RPE1-Arl13b-EGFP. (B) hTERT-RPE1-SMO-EGFP. (C) hTERT-RPE1-tdTomato-MCHR1. Left panels show Hoechst stained nuclei, middle panels show labeled cilia, and right panels are representative image segmentation as recognized by the software (shown as an overlay of the cilia image. For TF, green is nuclear segmentation, red is the cilia segmentation; for MWS, nuclei and cilia are overlaid in cyan). For each cell model, no serum and + serum panels represent the same image field. In A, settings for both TF and MWS modules can be developed that discriminate between background and cilia; for B, settings for TF but not MWS can be used. The tdTomato-MCHR1 model, shown in C, has significant intracellular signal, particularly following addition of serum (see arrows in the right panel), which render automated counting unreliable by either TF or MWS approach.

FIGURE 7

Quantification of ciliation using Transfluor and MultiWavelength Scoring modules. (A,B) Analysis of ciliation in hTERT-RPE1-Arl13b-EGFP and hTERT-RPE1-Smo-EGFP cell lines based on “vesicle count per cell” and “vesicle per cell” in the TF module (A,B for Arl13b, D,E for SMO), and positive W2 mean stain area in the MWS module (C for Arl13b, F for SMO). Data are presented as mean ± SEM, ^∗p ≤ 0.05, ^∗∗p ≤ 0.001, ^∗∗∗p ≤ 0.0001 as compared to control.