Single-Cell Analyses Confirm the Critical Role of LRP6 for Wnt Signaling in APC-Deficient Cells

The majority of colorectal cancers (CRCs) are initiated by constitutive activation of Wnt signaling resulting from truncation of the tumor suppressor adenomatous polyposis coli (APC) (Clevers and Nusse, 2012). APC promotes β-catenin turnover via its function within the destruction complex. In this traditional model, the Wnt receptor complex is dispensable for constitutive Wnt pathway activation and consequent CRC development following APC loss. Our published findings (Saito-Diaz et al., 2018) reveal an additional evolutionarily conserved function of APC: prevention of ligand-independent Wnt receptor complex signalosome formation. This novel APC function ensures receptor activation occurs only after Wnt stimulation. Thus, in our recently proposed model (Thorne et al., 2018), APC has two functions: (1) to promote β-catenin degradation (its traditional role) and(2) to prevent inappropriate receptor activation when Wnts are absent.

In this issue of Developmental Cell, Chen and He used CRISPR/Cas9 genome editing to generate cell lines in which APC and LRP6 and/or LRP5 were knocked out and failed to observe changes in β-catenin levels or Wnt reporter activity. They considered the possibility that compensation or adaptation, which are previously described phenomena (El-Brolosy and Stainier, 2017; Rossi et al., 2015), occurred in these cell lines as a result of the genetic deletions. To address this, the authors performed siRNA knockdown of LRP6 in APC-deficient cells but failed to observe decreased activity of a Wnt target gene reporter. Paradoxically, they observed increased Wnt reporter activity following LRP6 knockdown that they attributed to an LRP6 siRNA off-target effect. They concluded that APC loss results in Wnt signaling independent of LRP6/5, which is inconsistent with our newly proposed model. The implication was that other components of the Wnt receptor complex are also not necessary for the aberrant activation of Wnt signaling caused by loss of APC.

Genetic and biochemical instability of cancer cell lines leads to significant cell-to-cell heterogeneity (Ben-David et al., 2018; Liu et al., 2019). One explanation for our disparate results is that CRISPR/Cas9-mediated LRP5/6 knockout as performed by Chen and He resulted in compensatory genetic mechanisms that maintain elevated Wnt signaling over time in knockout clones. Although siRNA knockdown was utilized to rule out this possibility, β-catenin levels were not reported; thus a comparison of their data with our previous and current results cannot be made. We are unable to explain why Chen and He observed paradoxical enhancement of Wnt reporter activity upon knockdown of LRP6, but this apparently artifactual effect may have masked the requirement for LRP6 in APC-deficient cells. In addition, greater than 90% knockdown of a given component must be achieved to disrupt function of other signaling pathways (Krumins and Gilman, 2006). Thus, we routinely perform double siRNA transfections (12 h apart) to achieve efficient knockdown. As LRP6 levels following knockdown were not shown by Chen and He, the efficiency of LRP6 depletion in the various cell lines tested is not clear.

To rule out clonal idiosyncrasies or genetic compensation, we performed single-cell analyses of endogenous β-catenin levels in isogenic human colonic epithelial cell (HCEC) lines with defined oncogenic perturbations (Zhang et al., 2016). These immortalized cells are cytogenetically normal, retain the capacity to undergo multilineage epithelial cell differentiation, and exhibit a functional Wnt pathway (Roig et al., 2010). Using an LRP6-interfering oligonucleotide that contains a targeting sequence different from our previous study and from that used by Chen and He, we silenced LRP6 and measured nuclear and whole-cell β-catenin levels (Figure S1). Nuclear and total β-catenin levels were significantly reduced (Figures S1B and S1C). In controls, we observed increased β-catenin levels and increased cell-to-cell heterogeneity of b-catenin upon elimination of APC solely or in the context of activated KRAS, loss of p53, and reintroduction of a truncated APC protein frequently observed in CRCs. In each context, LRP6 silencing lowered β-catenin levels and suppressed their cell-to-cell heterogeneity. We obtained similar results using LRP6-interfering oligonucleotides (Saito-Diaz et al., 2018). In light of this heterogeneity, we speculate that the clones isolated by Chen and He may represent the skewed population of cells that are growing rapidly due to elevated β-catenin levels. Our results provide further support for our recent model for APC function and suggest that elevated levels of β-catenin in APC and LRP5/6 knockout clonal populations observed by Chen and He may simply be a consequence of clonal compensatory mechanisms.