The data presented here have demonstrated in two different models that depletion of Cα or Cβ induces a reduction of EGFR protein levels, and that this reduction is not due to decreased level of EGFR mRNA.
A recent report by  demonstrated that membrane expression of EGFR was inhibited by the PKA inhibitor H89 through increased internalization and endosome arrest of the EGFR in neuroblastoma N2a cells. Such an effect could be observed after 4 hours of H89 treatment. As H89 is known to have a broader inhibitory specificity [26–28], these observations did not necessarily prove involvement of PKA in maintaining EGFR levels. Here we demonstrate that EGFR was reduced in both mouse liver and HeLa cells after depletion of Cα and Cβ, respectively. The reduced levels of EGFR were not due to reduction in EGFR mRNA, as demonstrated in both HeLa cells depleted of either Cα or Cβ, as well as mouse livers isolated from Cα and Cβ KO mice. The results from mouse Cα and Cβ KO livers indicated a relative reduction in staining for EGFR along the plasma membrane, as well as a general reduction of immunoreactive protein. This is contrary to the previous observation that H89 and thus inhibition of PKA induces an increased internalization of EGFR, which suggests that reduced PKA activity leads to degradation of EGFR. Our results demonstrate in mouse liver as well as HeLa cells, that a reduction in total PKA activity either through elimination of Cα or Cβ leads to a reduced expression of EGFR protein. This was not observed in the brain of neither Cα nor Cβ KO mice. The explanation for this may be that brain has a much higher levels of both PKA Cα and Cβ than liver and HeLa cells [29, 15]. Thus, one C subunit may be expected to compensate for the loss of activity of the other isoform. Based on that we postulate that protein levels of EGFR depend on a certain threshold level of PKA and that this effect may not be isoform-specific. Such a hypothesis is in part supported by observations made by Huang and colleagues . In mice with one or two mutant C subunit alleles (Cα -/- and Cβ +/+ or Cα+/+ and Cβ -/- or Cα-/+ and Cβ -/+), either of them were born with apparent embryonic defects. However, when deleting three alleles (either Cα -/- and Cβ -/+, or Cα -/+ and Cβ -/-) they demonstrated a 100% penetrant spinal neural tube defect and spina bifida. This suggests that the C gene dose in the brain is sufficient to rescue a defect irrespective of the isoform, implying Cα and Cβ redundancy in the brain. Furthermore, in our experiments we observed that the regulatory effects on the EGFR levels were apparent even without prior stimulation of endogenous cAMP formation. This observation may be explained in several ways including that of chronic levels of endogenous cAMP, or free C subunits which activity is stimulated independently of cAMP. The latter hypothesis is supported by several studies demonstrating that C subunits when bound to proteins other than the PKA R subunit, the C subunits may be regulated in a cAMP independent fashion [31, 32].
The fact that Cβ ablated mice also displayed decreased levels of EGFR indicates that Cβ plays a role in mouse liver, despite previous data indicating low levels of Cβ in mouse liver . This was further substantiated by measurement of PKA activity in mouse liver, demonstrating reduced activity in both Cα and Cβ KO livers. As the sensitivity of Western blotting and mRNA measurements of Cα and Cβ will vary, our measurements of activity in KO mice is probably a better indication of the relative contribution of total C activity from the different isoforms. The reduction in C activity in Cα KO was 70%, while the reduction in C activity in Cβ KO mice was also 70%. This can best be explained by the presence of satiable PKA inhibitors, like endogenous PKI.
The fact that EGFR levels were reduced in the liver of Cβ KO mice is the first observation of a Cβ-dependent phenotype affecting non-nervous tissues. However, we were also able to demonstrate a small but significant reduction in size of the Cβ ablated mice compared to the wt mice, suggesting that the Cβ subunit may influence growth regulation as has been demonstrated for the Cα subunit. The Cα ablated mice showed a growth retardation phenotype consistent with disruption of the GH/IGF-I endocrine system, and while GH-hormone levels were normal, IGF-I mRNA and major urinary proteins (MUPs) levels were reduced, indicating a partial resistance to GH . It has been reported that Cα ablated mice are reduced in size by approximately 30% and approximately 90% die before puberty . Cβ ablated mice by contrast only show a very small reduction in size and they also differ from the Cα mice in that offspring normally survive to adulthood. It is tempting to suggest that Cβ and Cα have redundant effects. Moreover, this may also suggest that cellular signaling pathways involved in growth regulation are merely dependent on a certain level of PKA activity for proper function, rather than isoform-specific effects of Cα and Cβ. As Cα is the most abundant C subunit in non-nervous tissues, disruption of Cα would result in more severe effects in the whole animal than the ablation of the gene for the less abundant Cβ subunit.
High levels of EGFR are associated with many tumors with poor prognostic features. EGFR itself and its downstream signalling pathways are promising targets for anti-tumour drugs . PKA has also been proposed as a possible target for cancer therapy, and the therapeutic potential of the combined blockade of EGFR and PKA has been discussed . With respect to cancer therapy, the focus has been set on the regulatory isoforms of PKA. Site-selective cAMP analogue 8-Cl-cAMP and a series of modified antisense oligonucleotides targeting the PKA RIα subunit have been applied without conclusive effects [35–37]. However, the effects of the cAMP analog 8-Cl-cAMP may be mediated by metabolite rather than the cAMP analog itself, suggesting that PKA is not involved.