Background

Fibroblasts are almost ubiquitous in human tissues, and the most common cell type in the stroma of a number of solid tumours, where they are referred to as cancer-associated fibroblasts (CAFs) (Ishii et al., 2016; Kalluri and Zeisberg, 2006; Rupp et al., 2014; Servais and Erez, 2013). CAFs are associated with multiple hallmarks of malignancy (Ishii et al., 2016; Tao et al., 2017) and correlate with poor prognosis in multiple solid tumours (Hanley et al., 2018).

Given these tumour-promoting effects, and their genetic stability relative to cancer cells (Ishii et al., 2016), it is unsurprising that fibroblasts are an attractive therapeutic target. However, clinical trials targeting CAFs have so far yielded disappointing results (Hofheinz et al., 2003; Narra et al., 2007). This may, in part, be due to variation within the fibroblast population: these cells are known to be heterogeneous in both normal and disease states (Desmoulière et al., 2004; Sugimoto et al., 2006; Anderberg and Pietras, 2009; Servais and Erez, 2013; Witowski et al., 2015; Kalluri, 2016; Mellone et al., 2017). However, this heterogeneity remains poorly-characterised and it is not yet clear how many subtypes are present within a given tissue or tumour type, or the nature of functional differences between groups (Herrera et al., 2013; Servais and Erez, 2013; Ishii et al., 2016).

Characterising heterogeneity of cell populations within human tissues often requires analysis at a single-cell level. Single-cell RNA sequencing is a valuable platform for characterising multicellular ecosystems. However, fibroblasts are embedded within extracellular matrix and are particularly difficult to isolate: this has led to under-representation in, for example, single-cell RNA sequencing datasets (Lambrechts et al., 2018). Unlike many murine models, there is no standardized disaggregation protocol for human solid tissues. A number of different immune cell populations have been successfully isolated and analyzed directly from tissues (Holt et al., 1986; Perrot et al., 2007; Grange et al., 2011; Quatromoni et al., 2015; Ganesan et al., 2017). However, epithelial and non-immune stromal cells are usually cultured in vitro prior to analysis (Lurton et al., 1999; Koumas et al., 2003; Comhair et al., 2012; Barkauskas et al., 2013; Mackay et al., 2013). Culture in vitro has been shown to change fibroblast phenotypes (Öhlund et al., 2017; Waise et al., 2019); thus, how and whether the functional differences described in vitro are maintained in vivo is not yet known (Lurton et al., 1999).

Here, we describe an optimized protocol for the isolation of fibroblasts from primary human tissues, allowing immediate analysis without the need for culture in vitro. In brief, primary samples undergo mechanical and enzymatic dissociation, followed by incubation with TrypLE and red cell lysis buffer (to disrupt intercellular adhesions and remove red blood cells, respectively). Use of this approach yields a single-cell suspension consisting of multiple cell types, with approximately a 4-fold greater proportion of fibroblasts compared to other disaggregation strategies (Waise et al., 2019). We describe use of this protocol for the ex vivo analysis of fibroblasts in both normal and disease states using single-cell RNA sequencing, and highlight alternative downstream applications. In addition, this approach may have applications in the analysis of other cell types (e.g., epithelial cells).