Oncogenic PKCι decides tumor-initiating cell fate

Loss of cellular polarity is a well-documented characteristic of transformed cells, and genes involved in establishing and maintaining cellular polarity can act as either tumor suppressors (i.e. Numb) or oncogenes (i.e., atypical protein kinase Cι (PKCι), aurora kinase) in human cancers.¹ The PKCι oncogene, PRKCI, is frequently overexpressed as a result of tumor-specific copy number gains (CNGs) at chromosome 3q26 or the presence of oncogenic KRAS in many human tumor types.²

We recently demonstrated that elevated PKCι expression is associated with loss of cell polarity and acquisition of a highly aggressive, stem-like, tumor-initiating cell (TIC) phenotype in 2 major forms of lung cancer, lung squamous cell carcinoma and (LSCC) and oncogenic KRAS-driven lung adenocarcinoma (KRAS-LADC).^3,4 In LSCC, PKCι activates cell autonomous Hedgehog signaling to drive TIC behavior,⁴ whereas in KRAS-LADC, PKCι controls TIC behavior through activation of a PKCι-ELF3-NOTCH3 signaling axis.³ PKCι maintains TIC function and drives tumorigenesis by shifting the balance between symmetric cell division (SCD) and asymmetric cell division (ACD) in TICs toward ACD (Fig. 1).³ Either genetic or pharmacologic blockade of PKCι-ELF3-NOTCH3 signaling leads to selective inhibition of ACD and a profound loss of tumor-initiating potential in KRAS-LADC TICs, indicating the importance of ACD for tumor initiation and maintenance.³

Figure 1.

Two modes of tumor-initiating cell (TIC) replication. Symmetric cell divisions (SCDs; left) generate 2 stem-like TICs. SCD drives tumor initiation, but also tumor latency/dormancy. SCDs may also initiate tumor relapse after chemotherapy. Asymmetric cell divisions (ACDs; right) generate one TIC and one transiently-amplifying cell (TAC). TICs co-segregate template DNA to the daughter TIC, while the daughter TAC receives newly-replicated DNA; thereby ensuring the genomic integrity of TICs while TACs can acquire and accumulate mutations. TACs occasionally de-differentiate allowing newly acquired, advantageous mutations to enter a genetically-evolving TIC lineage. PKCι-ELF3-NOTCH3 signaling drives ACDs in KRAS-LADC TICs (see text).

TICs can undergo either SCD, which gives rise to 2 identical TIC daughter cells (Fig. 1A), or ACD which gives rise to a daughter TIC and a transient-amplifying cell (TAC) (Fig. 1B). Interestingly, our recent findings indicate that PKCι-ELF3-NOTCH3 signaling not only drives ACD, but is also necessary for daughter TICs to exhibit their highly aggressive stem-like behavior.³

Due to the inherently low proliferative rate of TICs, a shift to increased SCD favors TIC expansion, but limited tumor growth. Thus, SCDs may be critical for tumor initiation from a newly mutated tissue stem cell, but given the low proliferative rate of TICs, exclusive SCD may ultimately lead to tumor latency or dormancy (Fig. 1A). In contrast, a shift to ACDs allows for rapid exponential tumor growth by generating highly proliferative TACs. Perhaps not surprisingly then, TICs isolated from many human tumor types exhibit a substantial proportion of ACDs; we observed approximately equal numbers of SCDs and ACDs in KRAS-LADC TICs.³ ACDs may play other important roles in tumor biology beyond driving rapid tumor growth. Recent studies demonstrate that TICs exercise co-segregation of template DNA during ACD (reviewed in ref. 8). This interesting mode of cell division ensures that stem-like daughter TICs inherit only founder (unreplicated) DNA strands, while the differentiated daughter TACs acquire the newly synthesized DNA strands. This mechanism serves to maintain the genomic integrity of the TIC lineage by ensuring that newly synthesized DNA strands, which are susceptible to replication error-induced mutations, are not incorporated into the TIC progeny. Thus, ACD with co-segregation of template DNA may allow for more rapid accumulation of genetic mutations in TACs, thereby driving tumor progression (Fig. 1B). However, since TACs exhibit a finite replicative lifespan, such mutations are subject to extinction through replicative senescence. This situation is desirable from the perspective of a tumor, since deleterious mutations will be extinguished without affecting the genetic integrity of the TIC population. However, recent evidence indicates that partially-differentiated TAC progeny retain sufficient plasticity to re-acquire stem-like properties and re-enter the TIC pool.⁵ Thus, it is plausible that TACs that acquire an advantageous mutation can de-differentiate into TICs, thereby introducing the advantageous mutation into the immortal TIC population. Indeed, TACs that acquire an advantageous mutation would be expected to rapidly outnumber TACs lacking the mutation, thereby increasing the likelihood of such cells de-differentiating into TICs. Thus, ACD with template DNA co-segregation and TAC plasticity may allow acquisition and eventual immortalization of favorable mutations, while minimizing the risk of acquiring deleterious mutations in the TIC population, which could lead to tumor extinction. PKCι maintains the ability of TICs to undergo ACD.³ Thus, pharmacologic inhibition of PKCι represents a therapeutic approach to directly target the TIC phenotype. We have identified the anti-rheumatoid drug auranofin (ANF) as a selective PKCι inhibitor.² ANF selectively inhibits ACD of KRAS-LADC CSCs,³ exhibits potent anti-tumor activity,² and has shown promise clinically as an anti-tumor agent.^6,7 Several key questions remain regarding this novel therapeutic approach. First, does ANF suppress tumor heterogeneity and enhance genomic stability in addition to inhibiting TIC growth? This interesting question remains unanswered. Second, since ANF selectively inhibits ACD of TICs while at least partially sparing SCD, will this intervention invite cancer relapse by promoting the clonal expansion of residual, treatment-resistant TICs by SCD? While this question ultimately must be answered clinically, our pre-clinical data indicate that ANF not only inhibits ACD of KRAS-LADC TICs, but also inhibits the tumor-initiating, clonal expansion potential of symmetrically-dividing TICs. At present, it is unclear whether the mechanisms by which PKCι inhibits both ACD and the tumorigenicity of symmetrically-dividing TICs are the same, or are mediated through distinct signaling pathways.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.