Do p53 Repeat Fields Synchronize Target Cells and Natural Killer Cells?

Natural Killer cells face a remarkable biological problem. They must continuously distinguish healthy cells from stressed, infected or malignant cells and decide which should be destroyed. Increasing evidence suggests that p53 participates directly in coordinating that decision. A recent review in Frontiers in Immunology brings together evidence that p53 affects not only the internal fate of a damaged target cell, but also the signals by which Natural Killer cells recognize, engage and ultimately destroy it.

This creates an interesting question for Codondex. If p53 coordinates several different components of the NK-target interaction at the same time, could the unusually dense repeat structures that Codondex detects within TP53 help identify genomic regions involved in regulating that coordination? We describe these structures as High-Density Nested Repeat Fields, or HDNRFs: regions in which repeated, overlapping and nested short sequences become unusually concentrated. HDNRFs are not simply conventional tandem repeats. They represent a broader sequence geography that becomes visible when DNA is examined simultaneously across multiple sequence lengths and positions.

The Codondex methodology approaches this problem without first needing to know the biological function of the sequence. It divides genetic sequence into overlapping components, preserves their relative positions and relationships, and identifies sequence combinations that recur or associate unusually strongly. The important distinction is that Codondex can identify unusual sequence architecture first and ask what that architecture means biologically afterwards. The progression is therefore from sequence, to pattern, to biological association and finally, if experimentally demonstrated, to mechanism.

We have already seen an intriguing example of this process within TP53. In an earlier Codondex analysis, the algorithm selected 28-nucleotide sequences constructed from a set of shorter key sequences and found a remarkable concentration of recurring sequence components within a very short region of TP53. That analysis converged on the intron 3 region containing the known PIN3 polymorphism, or Polymorphism Intron 3. PIN3 is a 16-base-pair duplication within a region of TP53 already known from conventional research to possess unusual structural and regulatory characteristics, including G-rich sequence capable of participating in G-quadruplex formation and regulation of TP53 RNA processing.

The importance of the PIN3 observation is not that Codondex discovered PIN3 itself. The polymorphism was already known. What was significant was the route by which Codondex arrived there. The algorithm was not instructed to search for PIN3, a G-quadruplex or a known p53 regulatory element. It arrived at the region because of an unusual concentration of relationships within the underlying sequence. Biology then independently provided a reason why that region might matter. That is an early example of the utility we are now trying to test more systematically with HDNRFs: whether unusual repeat geography can lead us to biologically important regulatory locations before we understand why those locations are important.

A second observation has made the p53 question considerably more interesting. In our tumour-section sequence data, the Codondex-selected sequences were compared with multiple immune measurements, including Chromium-release cytotoxicity under pNK, IL-2-stimulated pNK and sNK conditions, together with IFN-γ, cell-number and Elispot measurements. The experimental data therefore allow sequence characteristics identified independently by the algorithm to be compared against several different manifestations of NK-cell behaviour.

In repeated analyses of these data, pNK has continued to emerge as one of the most interesting and apparently most correlated biological candidates. This has occurred when the data of 48 tumour-sections were approached in different ways, including ranking the occurrence of duplicate algorithm-selected sequences and examining shorter repeated, overlapping and tandem-repeat relationships within the selected 28-base sequences. When those independent sequence rankings were compared with the experimental measurements, pNK repeatedly returned as a candidate association. This recurrence is more interesting because it is not simply tracking the assay with the largest numerical values.

Indeed, the unstimulated pNK experimental signal is often remarkably weak. In the Chromium pNK measurements many tumour sections are at or close to zero, although a smaller number of sections show substantially greater cytotoxicity. The corresponding pNK IFN-γ measurements are even more striking, with many samples recording zero and only a subset producing detectable values. In contrast, IL-2-stimulated pNK and sNK measurements frequently produce much larger absolute responses. Yet despite this weak unstimulated signal, pNK has repeatedly surfaced when the Codondex sequence characteristics are ranked against the experimental results.

That distinction could be important. If the relationship were simply that more repeat sequence generated more general immune activity, one might expect the strongest associations to appear consistently in the assays producing the largest absolute responses. That is not obviously what we are seeing. Instead, an intrinsically weak pNK phenotype appears repeatedly when the sequence data are interrogated using different measures of repeat structure. At present this remains an exploratory observation rather than proof of a biological relationship, but it suggests that the algorithm may be detecting something associated with the underlying susceptibility of the target cell to natural NK surveillance, before that response is amplified artificially by IL-2 stimulation.

The p53 literature offers a plausible biological framework for explaining why such a relationship might exist. p53 appears capable of translating the internal stress state of a target cell into an external immune-recognition state. Wild-type p53 can increase expression of ULBP1 and ULBP2, activating ligands recognized by NKG2D on NK cells. p53 can influence PVR/CD155, which participates in DNAM-1 signalling, and ICAM-1, which helps stabilize the interaction between an NK cell and its target. In some experimental systems p53 activation also alters inhibitory MHC-I signalling. The intracellular state of p53 can therefore affect whether the exterior of the target cell tells an NK cell, in effect, that something is wrong.

The same p53 response can also influence whether the target dies once it has been recognized. p53 participates in BAX-mediated mitochondrial apoptosis and intersects with pathways involving granzyme activity, Fas and TRAIL death receptors. The same regulatory system that contributes to making the abnormal cell recognizable can therefore also make it more susceptible to the execution mechanisms used by the NK cell. p53-dependent signalling can additionally regulate chemokines such as CCL2 and influence the inflammatory environment controlling recruitment of NK cells to the target.

This suggests that p53 is not controlling one isolated NK signal. It may be coordinating several requirements for successful immune surveillance. The abnormal cell must first become recognizable. An NK cell must encounter it. Activating signals must overcome inhibitory signals. Stable NK-target contact must be achieved. The target must then remain susceptible to the apoptotic machinery delivered by the NK cell. p53 appears capable of influencing several of these conditions within the same biological response.

We describe this concept as p53-driven NK-target synchronicity. The word synchronicity is useful because the ultimate NK cytotoxicity measurement may be the visible end point of several variables moving together. A weak pNK result therefore does not necessarily imply a biologically unimportant relationship. If unstimulated pNK represents the basal ability of the NK cell to identify and eliminate an abnormal target, even a relatively small signal may contain considerable information about whether the underlying recognition machinery is properly aligned.

Viewed in this context, the repeated Codondex association with pNK becomes particularly interesting. The question is no longer simply whether a p53 repeat sequence correlates with the magnitude of NK cytotoxicity. It is whether an HDNRF identifies a genomic state associated with the probability that the target cell and NK cell become correctly synchronized. That could explain why a relatively weak pNK phenotype repeatedly appears when sequence characteristics are ranked in different ways. The algorithm may not be identifying the magnitude of an immune response so much as a genomic condition influencing whether the target is naturally visible to the NK cell in the first place.

This also produces a much more precise experimental hypothesis. If a p53 HDNRF is genuinely associated with NK-target synchronicity, it should not correlate only with the final Chromium pNK measurement. The variables between the sequence and the cytotoxicity result should also move together. Tumour sections ranked according to the relevant HDNRF signal could therefore be measured for functional p53 activity, ULBP1 and ULBP2 expression, PVR/CD155, ICAM-1, BAX and apoptotic competence, together with NKG2D and DNAM-1 engagement, NK-cell CD107a degranulation, IFN-γ production and direct target-cell killing.

The important result would not be one statistically significant relationship. It would be a coherent progression in which the HDNRF state predicts p53 state, p53 state predicts the target-cell recognition phenotype, and that phenotype predicts the pNK response. If these variables rise and fall together across the same tumour sections, it would provide a mechanistic explanation for an association that has so far appeared computationally.

Temporal experiments could test the relationship even more directly. Controlled activation of wild-type p53 should first change p53-responsive transcription, followed by changes in NK-recognition ligands and apoptotic competence, and ultimately by a measurable change in NK engagement and killing. Suppressing or removing functional p53 should disrupt at least part of that progression. Restoring p53 should restore it. Such an experiment would allow the apparent synchronicity to be observed rather than inferred from static correlations.

The decisive experiment, however, concerns the repeat field itself. If an HDNRF is merely a marker, it may reliably identify the biological state without causing it. If it is regulatory, directly altering the repeat field should alter the system. CRISPR-based sequence changes or targeted epigenetic manipulation of a candidate p53 HDNRF could therefore test whether changing its architecture alters p53 transcription, RNA processing or functional activity and whether those effects subsequently propagate through NK-target recognition and cytotoxicity.

PIN3 provides an important precedent for why this approach deserves investigation. Codondex sequence analysis independently converged on a densely repetitive TP53 region that conventional molecular biology had already identified as structurally and functionally unusual. The pNK results provide a second, independent clue. Different approaches to ranking the tumour-section sequence data have repeatedly pointed towards an NK phenotype that, paradoxically, is one of the weakest in absolute experimental magnitude. Together these observations raise the possibility that Codondex is detecting something more fundamental than a simple relationship between sequence abundance and biological response.

The hypothesis now becomes testable. If the relevant p53 HDNRFs vary together with p53 functional state, NK-activating ligand expression, NK-target engagement and unstimulated pNK cytotoxicity, then the previously observed pNK association acquires a mechanistic explanation. If altering the HDNRF changes that sequence of events, the implication becomes substantially stronger.

Codondex was designed to find relationships in genetic sequence before we necessarily know what those relationships mean. PIN3 showed how an algorithmically identified concentration of repeat relationships could converge on known p53 regulatory biology. The repeated pNK association suggests that these sequence patterns may also intersect with immune function. The p53–NK literature now gives us a biological framework capable of connecting those two observations.

The next research question is therefore no longer simply whether p53 contains unusual repeat fields. It is whether some of those fields participate in the extraordinary coordination by which a stressed or malignant cell becomes visible as a target, and a Natural Killer cell recognizes that signal and responds.

Chemo vs. Mecho

Data strongly suggests interaction between plasma membrane and submembrane at the endothelial surface controls the inflammatory response. 

A meta-analysis from six studies of global gene expression profiles of Blood Pressure (BP) and hypertension was performed in 7017 individuals. 34 genes were differentially expressed. Of these, 6 genes were linked including MYADM, which was the only gene, of 34 discovered across diastolic, systolic BP and hypertension. Knockdown of MYADM (19q13), a component of endothelial surface rafts induced an inflammatory phenotype altering barrier function through the increase of the adhesion receptor ICAM-1 (19p13). This is mediated by MYADM activation of ERM actin cytoskeleton proteins. 

Mechanical forces, without a definitive direction e.g., disturbed flow and relatively undirected stretch at branch points and other complex regions cause sustained molecular signaling of pro-inflammatory and proliferative pathways that include mechanical stretch tied to p53. 

ERM proteins also facilitate Sphingosine-1-phosphate (S1P) dependent egress for T-cells to migrate from lymphoid organs. Their directional migration, by blebbing is contained at the T-cell’s leading edge. This fundamentally different mode of migration is characterized by intracellular pressurization. Of the five S1P receptors S1P2 (19p13) is critical in the immune, nervous, metabolic, cardiovascular, musculoskeletal, and renal systems. Results suggest that the ratio between S1P1 and S1P2 (19p13) governs the migratory behavior of different T cell subsets. 

Human NK cells express S1P1 mRNA. Activation with IL-2 increases S1P1, promotes S1P4 (19p13) and S1P5 (19p13) but not S1P2 (19p13) expression. Unlike S1P1, S1P2 (19p13) signals through several different G-alpha subunits, Gi, G12/13, and Gq. S1P5 (19p13) is also expressed in human and mouse NK cells and was required for mobilization to inflamed organs. S1P5-deficient mice had aberrant NK cell homing during steady-state conditions. NK cell trafficking in vivo requires a dedicated sphingosine 1-phosphate receptor. 

Virus-infected mast cells selectively recruit NK cells and positively modulate their functions through mechanisms dependent on soluble mediators, such as interferons. Skin mast cells protect mice against vaccinia virus by triggering mast cell receptor S1P2 (19p13) and releasing antimicrobial peptides. S1P2 (19p13),  a negative regulator of platelet derived growth factor (PDGF) induced migration and proliferation as well as SphK1 expression. 

S1P inhibits macropinocytosis (internalizing extracellular materials) and phosphorylation of Akt via S1P2 (19p13) stimulation resulting in diminished antigen capture.

S1P1, S1P2 (19p13) and S1P3 receptors have redundant or cooperative functions for the development of a stable and mature vascular system during embryonic development. S1P2 (19p13)  and S1P3 are involved in regulation of endothelial barrier function, fibrosis, and vasoconstriction. 

Adipogenic differentiation is inhibited by S1P2 (19p13) as mediated by C/EBPα and PPARγ, which induces PEPCK, a more recent gene of interest in cancer that acts at the junction between glycolysis and the Krebs cycle.

Mecho or chemo, chicken or egg, what first?

Systolic Blood Pressure and Innate Immunity vs. the Cancer Brain

Participants with a valid heart disease phenotype (atherosclerosis) were identified in a MESA blood pressure analysis conducted over 10 years. The valid group varied from 770 to 1113 patients from whom further blood analysis queried a primary and exploratory hypothesis of immune cell subsets. Four statistically significant innate cell subsets were discovered to be associated with Systolic blood pressure (SBP); Natural Killer (NK) cells, gamma delta T cells and classical monocytes.

Separately, an analysis of 7017 individuals from 6 international studies of gene expression signatures for SBP, diastolic blood pressure (DBP) and hypertension (HTN) found 7717 genes of which 34 were most differentialy expressed. Enrichment analysis for the systolic and diastolic gene group's associated to NK cell mediated cytotoxicity and 13 other pathways including antigen processing and inflammatory response, pointing strongly to innate and adaptive immunity. MYADM was the only gene identified for all groups SBP, DBP and HTN.

MYADM controls endothelial barrier function through ezrin, radixin, and moesin (ERM)-dependent regulation of ICAM-1 expression. ERM expression is required for ICAM-1 expression in response to MYADM suppression or TNF-α. ICAM-1 is a paradigmatic adhesion receptor that regulates leukocyte adhesion together with integrin LFA-1. This connection between endothelial membrane and cortical actin cytoskeleton appears to modulate the inflammatory response at the blood tissue barrier. 

Pressure overload activates the sympathetic nervous system (SNS) and up-regulates p53 expression in the cardiac endothelium and in bone marrow (BM) cells. Increased p53 expression promotes endothelial-leukocyte cell adhesion and initiates inflammation in cardiac tissue, which exacerbates systolic dysfunction. SNS activates, at least by significant increase of circulating norepinephrine (NE), which up-regulates p53 expressions, while forced expression of p53 increased ICAM-1 expression. 

On endothelial cells SNS is mediated via catecholamine-β2-adrenergic signaling, which up-regulates the production of reactive oxygen species (ROS), activates p53 and induces cellular senescence. Immune cells, including macrophages, monocytes, NK cells, B and T cells express the β2-adrenergic receptor and catecholamine. During pressure overload, NE cultured macrophages up-regulated p53 expression, whereas introduction of p53 increased Itgal (LFA-1) expression (which binds ICAM-1). Treatment with NE increased ROS, which was attenuated after inhibition of β2- adrenergic signaling in macrophages. Endothelial cell–macrophage interaction via NE-ROS-p53 signaling induces up-regulation of adhesion molecules, thus contributing to cardiac inflammation and systolic dysfunction.

During hypertension the vascular endothelium activates monocytes, in part through ROS by a loss of nitric oxide (NO) signaling, increased release of IL-6, hydrogen peroxide and a parallel increase in STAT activation in adjacent monocytes. NO inhibits formation of intermediate monocytes and STAT3 activation. Humans with hypertension have increased intermediate and non-classical monocytes and  intermediate monocytes demonstrate evidence of STAT3 activation. Mice with experimental hypertension exhibit increased aortic and renal infiltration of monocytes, dendritic cells, and macrophages with activated STAT3.

A senescence-associated secretory phenotype (SASP) was induced in epithelial cells after DNA damage of sufficient magnitude. In premalignant epithelial cells SASPs induced an epithelial–mesenchyme transition and invasiveness, hallmarks of malignancy by a paracrine mechanism that largely depended interleukin (IL)-6 and IL-8. Strikingly, loss of p53 and gain of oncogenic RAS exacerbated the pro-malignant activities. This suggests a cell-non-autonomous mechanism by which p53 can restrain and oncogenic RAS can promote the development of age-related cancer by altering the tissue microenvironment. Oncogenic signaling pathways inhibit the p53 gene transcription rate through a mechanism involving Stat3, which binds to the p53 promoter in vitro and in vivo. Blocking Stat3 in cancer cells up-regulates expression of p53, leading to p53-mediated tumor cell apoptosis. 

Induced stretch or stretch from pressure overload may engage a non-autonomous, p53 centric micro-mechanical mechanism that escalates or deescalates innate responses against cells functioning outside the mechanical ranges that macrophages or NK cells permit. Thus, the neuro-immune extension through SNS signaling, may begin with circulating blood pressure or stretch promoted through inflammation. 

p53 vasoregulation and NK cell depletion in SARS-CoV2

p53 has earned first prize in the academic stakes. It is also the most mutated gene in cancer and elephant's have 20 copies, which probably explains their surprisingly low rate of cancer. Its associations to innate immunity, particularly Natural Killer (NK) cells through the mechanics of vasoconstriction-dilation have become a point of interest in COVID19 patients.

Remarkably COVID19 has inspired the global scientific community to focus a significant portion of its aggregate research toward the impact of  SARS-CoV2 (CoV2). For the first time in history global research is singularly focused because a large number of other protein's and gene's are affected by CoV2 binding Ace2. The Ace2 receptor is important in systems of vasoconstriction-dilation and has wide ranging impact.

CoV2 binding Ace2 reduces its availability to convert Angiotensin1 to Angiotensin 1-7 (Ang1-7) or Angiotensin 1-9 (Ang1-9), which primarily interact via MAS and Angiotensin2 Receptor (AT2R) respectively. These have been linked to signaling and stretch caused by vasoconstriction-dilation, mitochondrial dysfunction, mitochondrial fission as well as cardiac and vascular remodeling.

Ang1-7 and Ang1-9 interactions with MAS or AT2R cell surface receptors have been linked to signaling events that drive p53 binding DNA and transcription. Myocyte stretching activates p53 and p53-dependent genes, leading to the formation of Angiotensin II (Ang II) and apoptosis. AngII, stimulates phosphorylation of p53 (on serine 15) and CREB (on serine 133) and signaling converges on the p53-CRE enhancer to stimulate Bradykinin receptor 2 (BK2) gene transcription. BK2 is a key element in the p53 related kallikrein-kinin system (KKS) of vasodilation that counters the Renin-Angiotensin-Aldosterone-System (RAAS) of vasoconstriction. 

Aldosterone was shown to induce mitochondrial dysfunction and podocyte injury mediated by p53/Drp1-dependent mitochondrial fission. In neuronal cells p53 dependent declines in Drp1 and parkin contribute to altered mitochondrial morphology and cell death. Parkin, via Pink1 activity binds depolarized mitochondria to induce autophagy of mitochondria. Mutations in both Drp1 and Pink1 were fatal in Drosophila models. These events also implicate a direct functional link to chronic inflammation in ageing between p53 and expression levels of ICAM1 on endothelial and NK cells required to bind targets. The p53 mediated negative regulation of autophagy is Pink1 dependent and experiments have shown that mitochondrial antigens, recognized by NK cells presented on MHC's are Pink1 and parkin dependent. 

Severe COVID-19 patients have highly elevated Bradykinin and AngII, perhaps an indication of elevated p53 trends that have been discovered in these patients. Under normal circumstances, on endothelial cells Bradykinin would act as a potent vasodilator via its BK2 receptor. However, since Ang (1-7) potentiates Bradykinin action on BK2 receptors its near absence may reduce KKS vasodilation. On the other hand RAAS, also via p53 and elevated AngII primarily interacts with AT1R to promote vasoconstriction.  

NK cells through their Renin Angiotensin System may counter-regulate target cells in response. However, in COVID19 patients depletion of NK cells, invasion of Neutrophils and endothelial cell damage, in part through elevated p53 autophagy and apoptosis is the overwhelming nasty work of CoV2 against the backdrop of dysregulated blood pressure in tissue.

 

 

Impotent Natural Killers by Cancer Stem Cells and Ageing

Cancer stem cells have been found, through various mechanisms to alter the sentinel function and innate, immune surveillance of Natural Killer cells (NK). In senescent cells that have stopped cell division, including in cancer stem cell niches and NK induced vascular remodeling (as found in the developing placenta) NK's sentinel vigilance is also reduced.

Senescence-associated mitochondrial dysfunction, a significant trigger of multiple dimensions of the senescent phenotype is caused by disruption of normal mitochondrial autophagy (mitophagy). Mitophagy increases with aging and this age-dependent rise is abrogated by PINK1 or parkin deficiency. Deletion of a p53 response element on PINK1 promoter impacts p53-mediated PINK1 transcriptional repression. This p53-mediated negative regulation of autophagy has been found to be PINK1-dependent and constitutes a p53-PINK1 loop in nucleus and cytoplasm.

Further, mitophagy controls the activities of tumor suppressor p53 to regulate, at least hepatic cancer stem cells via Nanog. Prostate cancer cells escape NK attack by Nanog down-regulating ICAM1 (LFA1), to which NK would normally bind its target. In lung cancer NK have been found to limit the efficient clearance of senescent tumor cells from the mouse lung after p53 restoration. This indicated p53 may promote conditions for cellular survival and NK induced vascular remodeling or angiogenesis, necessary for the growth of tumors.

When under stress and inner mitochondrial membrane pressure gradient moves toward depolarization, Pink1 slots into the membrane, binds and phosphorylates p53 at Serine 392 (p53s392) and aids phagophore formation to enhance mitophagy. Mitophagy traps cytoplasmic p53s392, which reduces its transport to the nucleus where it would otherwise disrupt transcription of Nanog. (As illustrated below). 

Activated p53s392 nucleoside concentrations are effected by mitophagy

On the other hand, the sentinel function of NK may be subject to this PINK1 mediated mitochondrial switch. In prostate cancer cells Nanog promoted ICAM1 transcription required for NK binding target and cell killing. In prostate cancer cells Nanog over-expression restricts ICAM1, which promotes tumor formation. (As illustrated below). Investigating further, the direct functional link between p53 and ICAM-1 (CD54) in senescence and age-related disorders appears to be deeply integrated in mitophagy, senescence and immunity.

Nanog over-expression appears to be deterministic 

In stem cells where normal expression of Nanog transcribes ICAM1 and cancer stem cells where over-expression of Nanog restricts ICAM1, the variable PINK1-p53 switch may represent a "canary" that signals the state of  mitochondrial health to sentinel NK. However in some cancer cells where normal mitophagy is impaired and Nanog expression is restricted by p53s392, other p53 isoforms may directly promote the transcription of ICAM1.

In  two manipulation experiments using five different fibroblast cell lines that accelerated development of senescent associated secretory phenotypes a striking result was observed: oncogenic RAS expression, which causes genotoxic stress and senescence in normal cells, and functional loss of the p53 tumor suppressor protein. Both loss of p53 and gain of oncogenic RAS also exacerbated pro-malignant paracrine signaling activities. Experiments show that PINK1 and Parkin, which are regulated by p53 specifically regulate mitochondrial antigen presentation of both MHC classes.

So, the question is whether the p53-PINK1 mitochondrial switch acts as cell-health "canary" for sentinel NK, where its inherent variables and regulatory loop may be fertile ground for the challenges of developing cancers?