The AKT signalling axis: a brief orientation
The PI3K–AKT–mTOR pathway is one of the most studied signalling cascades in cell biology, and for good reason. It integrates signals from growth factors, nutrients, and cellular energy status, and translates them into decisions about cell survival, proliferation, protein synthesis, and metabolism. When a growth factor binds its receptor tyrosine kinase, the receptor activates phosphoinositide 3-kinase (PI3K), which phosphorylates phosphatidylinositol-4,5-bisphosphate (PIP2) to generate phosphatidylinositol-3,4,5-trisphosphate (PIP3) at the inner leaflet of the plasma membrane. PIP3 recruits AKT (also called protein kinase B) via its pleckstrin homology (PH) domain, bringing it to the membrane where it can be phosphorylated at two critical sites: Thr308 by PDK1, and Ser473 by mTORC2. Full catalytic activation of AKT requires phosphorylation at both sites, though Thr308 phosphorylation alone confers partial activity.
Activated AKT phosphorylates a large number of substrates. Among the most important are FOXO transcription factors (inactivating them, which suppresses apoptotic and cell cycle inhibitory gene expression), TSC2 (which relieves inhibition on mTORC1, promoting protein synthesis and cell growth), GSK3β (inactivating it, which among other things stabilises cyclin D1 and promotes cell cycle entry), BAD (preventing it from blocking BCL-2 anti-apoptotic proteins), and MDM2 (which it promotes to suppress p53). The cumulative effect of AKT activation is strong promotion of cell survival, proliferation, and anabolic metabolism.
Given this, it is not surprising that AKT is one of the most frequently hyperactivated kinases in cancer. Hyperactivation arises through several mechanisms: activating mutations in PI3K catalytic subunit (PIK3CA mutations are among the most common somatic mutations in human cancer overall), loss-of-function mutations in the tumour suppressor PTEN (which dephosphorylates PIP3 back to PIP2, opposing PI3K), amplification of AKT itself, and — critically for our purposes — loss of the phosphatases that terminate AKT signalling once activated. It is this last mechanism that brings us to PHLPP.
The discovery of PHLPP: Gao et al., 2005
For much of the early history of AKT research, the phosphatase that directly dephosphorylated AKT at Ser473 was unknown. PP2A had been implicated in AKT dephosphorylation in some studies, but a dedicated Ser473 phosphatase had not been identified. In 2005, Tianyan Gao and Alexandra Newton at the University of California San Diego published a landmark paper in Molecular Cell identifying PHLPP1 (originally called SCOP — suprachiasmatic nucleus circadian oscillatory protein) as a phosphatase that specifically dephosphorylates AKT at Ser473.
The discovery approach was elegant. Gao and Newton searched for proteins with PP2C-like phosphatase domains that also contained PH domains — reasoning that a physiological AKT Ser473 phosphatase would likely need membrane targeting capability to access AKT at the same plasma membrane location where it becomes activated. PHLPP1 fit this profile exactly. In cellular assays, overexpression of PHLPP1 reduced phospho-AKT-Ser473 levels without affecting Thr308, and knockdown of PHLPP1 elevated phospho-Ser473 and increased AKT target phosphorylation. Functionally, PHLPP1 overexpression reduced cell proliferation and induced apoptosis in multiple cancer cell lines.
A second family member, PHLPP2, was subsequently characterised. The two proteins share the overall domain architecture but differ in several important respects, including their substrate preferences and their regulation.
Domain structure: PP2C catalytic domain, RXXTL motif, and membrane targeting
Both PHLPP1 and PHLPP2 are large, multi-domain proteins. The defining features are: a PH domain at the N-terminus that provides membrane targeting and substrate recognition; a leucine-rich repeat (LRR) region that contributes to protein-protein interactions; a PP2C-type phosphatase catalytic domain (also called a PPM-type domain, meaning Mg2+/Mn2+-dependent protein phosphatase) that carries the catalytic activity; and a C-terminal PDZ-binding motif (the RXXTL sequence) that allows PHLPP proteins to be localised and scaffolded by PDZ domain-containing proteins at specific cellular compartments.
The PDZ-binding motif is particularly important from a regulatory perspective. PDZ scaffolding proteins such as NHERF (sodium-hydrogen exchange regulatory factor) and MAGI family proteins control where PHLPP is localised within the cell, and therefore which pools of AKT it can access. Disruption of PDZ interactions alters PHLPP substrate specificity and localisation. This means that PHLPP activity is not simply a matter of expression level — spatial organisation within the cell governs function.
The PP2C catalytic domain itself is characterised by its Mg2+/Mn2+-dependence (unlike PP2A, which uses a different catalytic mechanism) and its resistance to okadaic acid, a natural toxin that inhibits PP2A and PP1 phosphatases. This pharmacological distinction between PP2C-type and PP2A-type phosphatases has been useful in untangling which phosphatase is responsible for specific dephosphorylation events in cellular experiments.
Isoform distinctions: substrate preferences of PHLPP1 versus PHLPP2
The three AKT isoforms — AKT1, AKT2, and AKT3 — are highly homologous but have distinct tissue distributions and biological roles. AKT1 is ubiquitously expressed and plays a dominant role in cell survival and proliferation. AKT2 is particularly important in insulin signalling and glucose homeostasis — it is the isoform most critical for GLUT4 vesicle translocation in insulin-responsive tissues (skeletal muscle, adipose). AKT3 has prominent roles in the brain and in certain cancer subtypes including melanoma and glioblastoma.
A key finding in PHLPP research is that PHLPP1 and PHLPP2 have non-redundant substrate preferences for the three AKT isoforms. PHLPP1 preferentially dephosphorylates AKT2 and AKT3 at Ser473. PHLPP2 preferentially dephosphorylates AKT1 and AKT3 at Ser473. This means that total AKT Ser473 phosphorylation — the readout used in most studies — reflects the combined actions of both phosphatases, but the specific effects on glucose metabolism (primarily AKT2-mediated) versus cell survival signalling (primarily AKT1) depend on which PHLPP isoform is active. The practical implication is that knockout or overexpression of the two isoforms produces distinct phenotypes that are not fully interchangeable, even though both proteins dephosphorylate Ser473.
Beyond AKT, PHLPP1 and PHLPP2 have additional substrates that expand their biological reach. PHLPP1 dephosphorylates and inactivates protein kinase C (PKC) isoforms, specifically at the hydrophobic motif analogous to AKT Ser473. PKC isoforms are themselves major regulators of cell survival, polarity, and differentiation, adding another layer to PHLPP’s tumour suppressive activity. PHLPP2 has been found to dephosphorylate and regulate Mst1/LATS kinases in the Hippo pathway, suggesting a role in organ size control and mechano-transduction that is entirely distinct from AKT regulation.
PHLPP as a tumour suppressor: deletion and silencing in cancer
If PHLPP dephosphorylates AKT at Ser473 and thereby limits cell survival and proliferation, loss of PHLPP would be expected to phenocopy constitutive AKT activation — and this is precisely what the cancer genomics literature shows. PHLPP1 is deleted or transcriptionally silenced in a substantial fraction of prostate cancers (where it cooperates with PTEN loss to drive full AKT hyperactivation), colorectal cancers, breast cancers, glioblastoma, and pancreatic cancer. PHLPP2 is similarly deleted in a subset of cancers, sometimes in patterns that differ from PHLPP1, consistent with their distinct substrate preferences.
The concept of redundant tumour suppression at the AKT-Ser473 level is important here. PTEN removes PIP3, preventing AKT from reaching the membrane in the first place. PHLPP removes the phosphorylation from Ser473 after AKT has been activated. These are not redundant in a simple sense — they operate at different points in the activation cycle — but they can partially compensate for each other. In cells where PTEN is lost, PHLPP provides a secondary braking mechanism. When both are lost, AKT activity becomes very difficult to restrain through physiological mechanisms. This PTEN-PHLPP redundancy has become clinically relevant: prostate cancers that have lost both PTEN and PHLPP1 are among the most aggressive, and they predict poor response to single-agent PI3K inhibitors.
Mechanistically, PHLPP transcription is itself regulated by upstream signals, creating feedback loops. One important loop involves mTORC1: when mTORC1 is active (driven by AKT phosphorylating TSC2), mTORC1 substrate S6K1 phosphorylates IRS-1 on inhibitory sites, providing negative feedback on PI3K activation. Separately, mTORC2 — the complex that phosphorylates AKT at Ser473 in the first place — is itself subject to feedback via mTORC1 activity. The regulatory interplay between mTORC1, mTORC2, AKT, and PHLPP forms a complex network rather than a simple linear pathway.
The mTORC2 connection: who phosphorylates Ser473 in the first place
PHLPP specifically reverses the Ser473 phosphorylation placed there by mTORC2 (mTOR complex 2, defined by its Rictor subunit rather than the Raptor subunit that defines mTORC1). This distinction is mechanistically important because mTORC2 and mTORC1 have different biological roles. mTORC1 is the nutrient and growth factor sensor that promotes protein synthesis and inhibits autophagy. mTORC2 is more specifically a regulator of cytoskeletal organisation and AKT full activation, and it is less sensitive to rapamycin in acute treatment settings (though chronic rapamycin does inhibit mTORC2 assembly in many cell types).
If mTORC2 places phosphate at Ser473 and PHLPP removes it, then the steady-state level of phospho-AKT-Ser473 reflects the balance between mTORC2 kinase activity and PHLPP phosphatase activity. Interventions that reduce mTORC2 activity would be expected to lower phospho-Ser473 independently of PHLPP, and this is relevant to understanding how spirulina might modulate this balance.
How spirulina connects to the PHLPP-AKT axis
Spirulina’s effects on the PI3K-AKT-mTOR pathway have been studied primarily in the context of its anti-proliferative effects in cancer cell lines and its metabolic effects in insulin resistance models. Several mechanisms of interaction are plausible and partially supported by data.
First, phycocyanin has consistently been shown to reduce mTOR signalling — specifically, to reduce phosphorylation of canonical mTORC1 substrates such as S6K1 (Thr389) and 4E-BP1 (Thr37/46). Several studies in hepatocellular carcinoma, colon cancer, and breast cancer cell lines have documented this. The mechanism appears to involve upstream AMPK activation: phycocyanin and phycocyanobilin activate AMPK, likely through mitochondrial effects on the AMP:ATP ratio, and AMPK phosphorylates and activates TSC2 while simultaneously phosphorylating Raptor (a component of mTORC1) to inhibit mTORC1 activity.
The connection to Ser473 and PHLPP comes through mTORC2. While AMPK-driven mTORC1 inhibition is well established, mTORC2 is more complex. Some evidence indicates that when mTORC1 is inhibited (for example by rapamycin or AMPK activation), there can be a paradoxical upregulation of mTORC2 activity in some cell types, because the negative feedback from mTORC1-S6K1 on IRS-1 is relieved, allowing more PI3K signalling and therefore more mTORC2-driven Ser473 phosphorylation. However, other data suggest that phycocyanin reduces both mTORC1 and mTORC2 activity, possibly through effects further upstream in the PI3K pathway. The precise effect on phospho-Ser473 in different cellular contexts is not fully resolved.
If spirulina’s components reduce mTORC2 activity or reduce the upstream PI3K signalling that drives mTORC2, this would lower the rate of Ser473 phosphorylation — reducing the burden on PHLPP to dephosphorylate it. In cells with functional PHLPP, the combined effect would be lower steady-state phospho-AKT-Ser473. In cancer cells with deleted or silenced PHLPP, the mTORC2-suppressive effect of spirulina components would be the only brake available, making it more important as a modifier in PHLPP-deficient contexts.
Metabolic implications: AKT2, GLUT4, and insulin sensitivity
The metabolic angle of this story focuses on AKT2 specifically. In skeletal muscle and adipose tissue, insulin signalling activates PI3K, which generates PIP3, which recruits AKT2 to the membrane. AKT2-Ser473 phosphorylation by mTORC2 provides full AKT2 activation, and active AKT2 phosphorylates AS160 (TBC1D4), a Rab-GTPase-activating protein that when unphosphorylated suppresses GLUT4 vesicle docking at the plasma membrane. AKT2 phosphorylation of AS160 at Thr642 releases this brake, allowing GLUT4 translocation and increased glucose uptake. This is the canonical mechanism of insulin-stimulated glucose disposal.
PHLPP1, with its preference for AKT2 dephosphorylation at Ser473, is therefore a regulator of insulin sensitivity. Overexpression of PHLPP1 in muscle reduces AKT2 Ser473 phosphorylation and impairs insulin-stimulated GLUT4 translocation. Conversely, reduction of PHLPP1 activity could enhance insulin sensitivity — which may be relevant to understanding one mechanism by which spirulina improves metabolic markers in insulin resistance models.
Several clinical and animal studies have shown spirulina supplementation improving insulin sensitivity, fasting glucose, and HbA1c in people with type 2 diabetes or metabolic syndrome. The mechanistic credit has generally been attributed to AMPK activation, which has many downstream effects including GLUT4 translocation through AMPK’s own phosphorylation of AS160 (at Ser588, distinct from the AKT2 site). But the PHLPP1-AKT2-GLUT4 axis could be contributing in parallel. AMPK activation reduces mTORC1 activity, and through mTORC1-S6K1 pathway relationships, may relieve certain negative regulatory inputs that normally suppress AKT2 signalling — a somewhat counterintuitive route to improved insulin sensitivity via the same AMPK activation that also suppresses mTOR.
What remains unknown and what to watch for
The honest state of the evidence is this: the PHLPP literature is robust in establishing PHLPP1 and PHLPP2 as genuine AKT Ser473 phosphatases with tumour suppressor function. The spirulina literature is robust in establishing that spirulina components, particularly phycocyanin and phycocyanobilin, modulate AKT pathway activity. The direct connection — specifically, whether spirulina components regulate PHLPP1 or PHLPP2 expression, localisation, or activity — has not been directly studied in available literature.
What would be worth watching for in future research: whether phycocyanin affects PHLPP1/2 mRNA or protein expression in metabolically relevant cell types (hepatocytes, myocytes); whether the metabolic improvements seen with spirulina in T2D models are accompanied by increased PHLPP1 activity and reduced AKT2-Ser473 phosphorylation specifically; and whether spirulina’s anti-proliferative effects in cancer cell lines are attenuated when PHLPP1/2 are knocked down, which would implicate PHLPP upregulation as part of the mechanism.
The PHLPP story is also a reminder that the AKT pathway is not simply a cascade to be activated or blocked — it is a tightly regulated cycle with multiple feedback mechanisms and dedicated termination machinery. PHLPP represents the underappreciated dephosphorylation half of this cycle. Understanding how dietary and nutritional interventions influence not just the kinases but also the phosphatases of this pathway is a more complete picture of metabolic regulation.