Tissue repair genes: the TiRe database and its implication for skin wound healing

Overview of experimental models used to establish wound healing-associated genes (WHAGs)

There is a great variety of methods available for the study of skin wound healing (WH), both with regard to types of genetic interventions and wounding assays [11]. In our dataset, the dominant interventions used in the mouse model are genetic (i.e. knockout or overexpression), whereas other interventions, such as protein administration, are more common in the other species (Table 1). As seen in Table 2, the most common wounding method by far is the dorsal full-thickness excision model.

Characterization of WHAGs

The TiRe data collection offers an opportunity to gain insight into the features of WHAGs. Most of these genes were identified in the mouse model. Notably, genes that were studied in rats, swine and humans were also studied in mice, and some in more than two species (Table 3). Despite the differences in intervention and wounding methods, targeting the common genes across the species mostly led to consistent results, i.e. to concordant effects. This suggests that WH across these species has much in common.

WHAGs are differentially conserved across vertebrates and invertebrates

To broaden this perspective, we further investigated the evolutionary conservation of WHAGs. For that purpose, we extracted the WHAG orthologs for all species available in the InParanoid database [13]. As seen in Figure 1, WHAGs are over-conserved among vertebrates, but are under-conserved in invertebrates (for specific details, see Suppl. Table 1).

This implies that (i) many of the skin WHAGs are a relatively recent acquisition in the course of evolution; and (ii) despite the significant differences in the anatomy and physiology of the skin between vertebrate species and the resulting differences in wound healing [14], the genetic basis of WH is conserved among vertebrates.

WHAGs are enriched in extracellular and immuno-inflammatory pathways

Complementary to the results above, our enrichment analysis on WHAGs sheds further light on this vertebrate-specific evolutionary conservation. As seen in Figure 2A, WHAGs predominantly encode for extracellular proteins and those involved in cell-cell/cell-ECM interactions. Furthermore, KEGG pathway enrichment analysis highlights a particular role for the focal adhesion pathway as well as for the ECM receptor interaction, regulation of actin cytoskeleton pathways, and various immune/inflammatory-related pathways (Figure 2B). Remarkably, the pathways involved in immune and inflammatory responses are even more over-represented when considering the genes that are conserved only in vertebrates (Figure 2C).

This was especially noted for the cytokine-cytokine receptor interaction and the associated JAK-STAT signaling pathways, hematopoietic cell lineage, complement and coagulation cascade pathways, and the adipocytokine signaling pathway which are enriched only in genes unique to vertebrates. Altogether, the results point to the importance of immuno-inflammatory reactions in wound healing, in vertebrates in particular. This is in line with numerous studies showing the importance of the inflammatory phase and with the unique immunity profile and functionality of vertebrates [15, 16].

WHAGs are highly interactive and form a protein-protein interaction network

Further supporting the notion that WH is a highly orchestrated and coordinated process [17] is the observation that WHAGs are greatly interconnected and more than two thirds of the WHAGs from the interactome (204/311) can be organized as a continuous protein-protein interaction (PPI) network (clustering coefficient of the entire set = 0.127). Moreover, as seen in Figure 3, many of these WHAGs serve as hubs connecting multiple signaling pathways.

Additionally, WHAGs have an unusually large number of first-order partners (average connectivity of 48.1 compared to 16.4 for the entire interactome) and together with their first-order interaction partners they would form a huge PPI network of 6,109 proteins, i.e., almost a third of the entire interactome. This incredible connectivity indicates that WHAGs are also in the “epicenter” of many other processes.

It has been previously shown that the most critical proteins in a given dataset have more connections within the network than is expected by chance [18-20]. Therefore, it is reasonable to assume that first-order partners of WHAGs could also be screened for their importance in WH using their connectivity (both total or with other WHAGs only) for prioritization. The power of such an approach has recently been proven useful for searching new longevity regulators of C. elegans among partners of longevity-associated genes [21]. As an example are 14 selected candidate genes that are not in our original WHAG list, but are highly enriched in connectivity to WHAGs and therefore have a high chance to be valuable for WH (Table 4). Of note, most of them participate in at least one of the WHAGs-enriched signaling pathways (Figure 2B). For example, SRC which interacts with 39 WHAGs, is at the crossroad of 5 WHAGs-enriched signaling pathways, and has been shown to be an important modulator of cell migration during WH after electric stimulation [22]. Another interesting example is EP300, which is connected to 40 WHAGs, and is involved in both adherens junctions, and the Jak-STAT signaling pathways. EP300 has been suggested to mediate the stimulatory effect of mechanical stress on WH [23]. Of course, further work is required to validate the significance of these candidates, yet this approach demonstrates well how the TiRe dataset can be used to find new wound healing targets.

Is accelerated wound healing “good” for longevity?

In an attempt to address this question, we have extended our previous analysis [10] by comparing the list of WHAGs with those reported as being involved in the control of lifespan [24]. The comparison yielded 17 genetic mouse models of extended lifespan (longevity phenotype), or reduced lifespan (premature aging phenotype), which were also tested for skin WH. The results are summarized in Table 5.

It is important to note that many studies used the rate of skin wound closure as a biomarker, assuming a priori that slower skin WH is indicative of an aging phenotype. Yet, our analysis shows that a slower or faster skin WH is indicative of an aging or longevity phenotype, respectively, only when assessed in advanced ages (Table 5), but not in the young. For example, Agtr1a knockout resulted in slower wound healing in young mice but also in an extended lifespan [25]. In contrast, Cav1 knockout, which accelerated wound closure, was accompanied by reduced longevity [26].

This means that pro- or anti-longevity effects of genetic interventions manifest in accelerated or delayed skin WH only in advanced ages, but not in young animals. Moreover, it seems that the association between the rate of WH and longevity is primarily attributed to an overall effect of the target gene on organismal aging rather than to its skin-specific action. This assumption is strongly exemplified by our study on the long-lived αMUPA mice, which preserve their skin WH capacity up to an old age (at least 25 months) [10, 27]. In this unique model [28], the uPa transgene is expressed in the ocular lens and the brain stem but not in the skin, thus excluding the gene-specific effects on WH. Overall, the results emphasize that the age factor should be taken into account when evaluating the links between skin WH, aging and longevity. To better understand these links, including older animals in the analysis is encouraged while using only young animals might yield confusing or misleading results. In particular, the opposite effect between the rate of skin WH in young age and the effect on life span could be explained by the links between WH and cancer, and the role of cancer in the determination of mouse longevity. Indeed, Schäfer and Werner [29] consider “cancer as an overhealing wound”. This could be especially relevant to mice as cancer is the main cause of death for a variety of murine strains [30, 31]. For example, Tert overexpression in the young leads to accelerated WH, a high incidence of cancer, and increased mortality [32]. Another example is the tumor suppressor gene Pten, known to negatively regulate the activity of the PI3K/mTOR pathway, which is involved in various cancers [33, 34]. Knockout of this gene resulted in accelerated WH in young age but a decreased lifespan [35], which is most likely associated with increased tumorigenesis.

Database content

Skin wound healing-associated genes (WHAGs) were determined based on genetic studies (knockout, knockdown, overexpression), or interventions that directly influence the level and/or activity of the protein product (antibody treatment, protein administration, etc.). A summary of the types of interventions is listed in Table 1. For all WHAGs included in the database, a given intervention has been observed to cause a marked change in the skin wound healing phenotype (such as accelerated or delayed wound closure, or alterations in the quality of repair). The list of WHAGs established thus far in model organisms and humans was compiled from scientific literature and manually curated. To date, the list contains 397 entries for four organisms: Mus musculus, Rattus norvegicus, Sus domesticus, and Homo sapiens (330, 40, 12 and 14 entries, respectively).

In addition to information about the WHAGs, the database also includes the genetic background of the animal model, the type of genetic/protein intervention, the wound model used, wound dimensions and location, and a brief description of the wound healing outcome, with a reference to the original research.

Interface

TiRe has a user-friendly website interface, with simple and intuitive navigation tools. Searching can be done either by gene symbol, its full name, or gene aliases. Alternatively, the data can be reached by species browsing. The website also allows for downloading the entire dataset from the download page, in order to carry out more extensive analyses offline. A build counter and a build release date are provided to keep track of different database versions.

Availability

The TiRe database is available at http://www.tiredb.org, with the data made available under the permissive Creative Commons license, allowing data to be used in other analyses. There are options to either download the entire database or its parts. Feedback is welcome.

Data analysis