The epidermis is a stratified epithelial tissue which forms the outermost barrier in all metazoan organisms and plays a critical role in homeostasis. The barrier function of the epidermis is intrinsically dependent upon its structure, with keratinocytes (‘skin cells’) arranged in a deep-to-superficial spatiotemporal differentiation gradient. Keratinocyte stem cells, located adjacent to the dermis, produce daughter cells which undergo cornification as they move towards the outermost surface, driven by proliferation within deeper layers. To ensure correct epidermal architecture and thus function, the rate at which fully differentiated KCs are lost from the superficial surface must be well-matched to the rate of proliferation within the basal layer.

Fig. 1.  Bi-levelled schematic of an epidermal cross-section with commonly identified layers. Keratinocyte differentiation (left to right) alters the cellular biochemistry and morphology (Table I).

Fig. 1. Bi-levelled schematic of an epidermal cross-section with commonly identified layers. Keratinocyte differentiation (left to right) alters the cellular biochemistry and morphology (Table I).

Multiple mechanisms regulate keratinocyte terminal differentiation, producing an intricate balance of life and death signals. Many environmental stimuli which influence keratinocyte fate are integrated within the mitogen-activated protein kinase (MAPK) intracellular signalling network. The extracellular-signal regulated (ERK) branch of MAPK utilises sequential phosphorylation through the Raf -> MEK-1/2 -> ERK-1/2 signalling cascade. The kinase activity of phospho-ERK-1/2 regulates a variety of substrates including transcription factors and cell cycle control proteins. Because ERK acts upon so many different proteins, signalling specificity is thought to require scaffolding molecules and sequestration of ERK within sub-cellular compartments, together with its substrates.

Fig. 2.  A co-labelled immunofluorescence image of phospho-Raf-1 (blue), phospho-MEK-1/2 (red) and phospho-ERK-1/2 (green) in human epidermis. Cytoplasmic and nuclear signal intensity for each target is plotted relative to the normalised distance (Fig. 1).

Fig. 2. A co-labelled immunofluorescence image of phospho-Raf-1 (blue), phospho-MEK-1/2 (red) and phospho-ERK-1/2 (green) in human epidermis. Cytoplasmic and nuclear signal intensity for each target is plotted relative to the normalised distance (Fig. 1).

in situ intracellular signalling

We are applying quantitative confocal microscopy with immunofluorescence labelling to investigate signalling proteins within human epidermis. Confocal microscopy is an ideal technique to examine the epidermis: it provides sufficient spatial resolution to identify major sub-cellular compartments; with a sufficiently large field of view to capture the entire gradient of keratinocyte differentiation.

Antibodies against fourteen proteins and four phospho-proteins within and around the ERK-MAPK pathway were used for immunofluorescence labelling and confocal microscopy. The cytoplasmic, nuclear and plasma-membrane localised signal was segmented across the depth of the epidermis for all targets which showed differential expression.

A tissue-based co-ordinate system which distinguishes discrete cell layers was applied to the epidermis, allowing us to describe the localised signal intensity for each target as a function of keratinocyte differentiation. Utilising the inherent relationship between a keratinocytes age and its position within the epidermis, we are applying a variety of time-dependent network inference techniques in an attempt to identify intracellular signalling cascades which are active within human epidermis.

Inference Results

Most network inference techniques output a collection of statistical relationships between observed variables – the signal intensity associated with a sub-cellular compartment for each target protein. Interpretation of the resulting networks with respect to known molecular and epidermal biology allows us to identify mechanisms which may influence the balance of life and death signals which regulate keratinocyte terminal differentiation.

Fig. 3.  A subset of the consensus network structure inferred using dynamic Bayesian Network inference methods. Edges (arrows) shown in represent indicate known molecular interactions which are being recapitulated within the inferred network. Grey edges represent putative biochemical interactions which may be occurring within human epidermis. Magenta, cyan and yellow nodes (diamonds) represent the membrane-, cytoplasmic and nuclear-localised signal for each target.

Fig. 3. A subset of the consensus network structure inferred using dynamic Bayesian Network inference methods. 

Edges (arrows) shown in represent indicate known molecular interactions which are being recapitulated within the inferred network. Grey edges represent putative biochemical interactions which may be occurring within human epidermis. Magenta, cyan and yellow nodes (diamonds) represent the membrane-, cytoplasmic and nuclear-localised signal for each target.

The inference models developed thus far identify amongst other things, relationships between variables which represent localisation-specific signal for the same protein, highlighting what are most likely translocation events. Edges between alternative proteins should be interpreted with care, but when combined with the appropriate biological knowledge, may provide hypotheses of control mechanisms which can be experimentally verified.

Leave a Reply

Your email address will not be published. Required fields are marked *