Background

The centrosome is the major microtubule-organizing center in cells and is crucial in defining mitotic spindle poles and forming cilia (Conduit et al., 2015). A centrosome consists of a pair of centrioles that serve as the duplicating elements of the centrosome and organize surrounding layers of pericentriolar material (PCM), a protein matrix with microtubule-nucleating activity (Bettencourt-Dias and Glover, 2007; Lawo et al., 2012). Centrosome numerical abnormalities, especially centrosome amplification, have been reported to be commonly present in a variety of human tumors and are correlated with advanced tumor grade, metastasis, recurrence, and poor survival (Bettencourt-Dias et al., 2011; Chan, 2011; Godinho and Pellman, 2014; Gonczy, 2015). Furthermore, centrosome numbers display both inter-tumor and intra-tumor heterogeneity in a broad range of tumors (Chan, 2011). Thus, a method to accurately quantify centrosome numbers on a single cell level in human tissue sample is useful to identify cancer cells with centrosome amplification, to study tumor heterogeneity, and to identify patients that might benefit from therapies that target cells with centrosome amplification.

Human centrioles are approximately 450-500 nm in length with a diameter of 200-250 nm (Gonczy, 2012; Greenan et al., 2018). Currently, centrosome quantification mainly relies on microscopy. Electron microscopy has been used as an accurate method to quantify centrosome numbers (Lingle et al., 1998). Super-resolution microscopy is another powerful way to study human centrosome structures (Sonnen et al., 2012). However, both electron microscopy and super-resolution microscopy require careful sample preparation and can be too laborious for quantifying centrosomes in a large number of cells and clinical tissue samples. Thus, most studies use high resolution microscopy like confocal and deconvolution microscopy. Further, the selection of centrosome markers is critical for accurate quantification. Most previous studies rely on PCM proteins like γ-Tubulin and pericentrin as centrosome markers, and centrosome numbers were quantified by counting PCM foci. However, PCM proteins are highly dynamic during centrosome assembly and can form assemblies or fragmentation without centrioles; they are not representative as bona fide centrosomes (Woodruff et al., 2014). Therefore, quantifying accentriolar PCM foci can lead to false positive results as centrosome amplification. On the other hand, cancer cells with centrosome amplification often undergo centrosome clustering during mitosis (Ganem et al., 2009; Silkworth et al., 2009). This leads multiple centrosomes in close proximity to each other and to appear as single PCM foci, which can result in false negative detection of centrosome numbers.

We have established a method using both centriolar protein CEP135 and PCM protein γ-Tubulin together as centrosome markers. We multiplexed centrosome markers with cell-cell junction markers to quantify centrosomes on a single cell level by high resolution microscopy (Wang et al., 2019a). We have tested different centriolar and cell-cell junction markers. Cell-cell junction proteins like E-Cadherin and α6 integrin are used to distinguish single cells from the neighbor cells and to identify the whole cell volume (Figure 1A). Foci with both centriolar proteins and PCM proteins are considered as centrosomes. Using this method, we have successfully performed centrosome quantification on a variety of normal and cancer tissues (Wang et al., 2019a; Wang et al., 2019b). Further, this method can be used to detect cells with centrosome amplification in both interphase and mitotic cells (Figure 1B). Here, we describe a detailed protocol to perform this method.


Figure 1. Centrosomes in human normal and cancer tissue. A. A human normal tonsil FFPE tissue sample stained with CEP135 (red) and E-Cadherin (green). Nuclei in blue. Scale bars: 10 μm. B. A human triple-negative breast cancer FFPE tissue sample stained with CEP135 (red), γ-Tubulin (green) and E-Cadherin (green). Nuclei in blue. Scale bars: 10 μm. Arrowheads indicate centrosomes.