Role of Wnt signaling in Malawi cichlid craniofacial development and diversification


What we know:

Ecological specialization and phenotypic variability

Cichlid fishes have undergone rapid evolution, filling a wide variety of niches and resulting in an array of diverse phenotypes pertaining to different ecological pressures (Kocher, 2004). Within Lake Malawi, there are hundreds of cichlid species that diverged from a common ancestor approximately 500,000 years ago. Genetically, these species are highly similar to one another; however, they are phenotypically diverse (Loh et al., 2008). A driving force behind the speciation of cichlids is the abundance of habitat and food types available within the African Rift Valley lakes.

Figure 1: A sample of the phenotypic variation found among African cichlids (Albertson & Kocher, 2006)

Malawi cichlids are found in two distinct habitats; species that dwell in rock habitats are termed mbuna and species found in sand habitats are called utaka. Some differences between fish in these groups are brain regionalization, body forms, pigment patterns, visual systems and behaviors (Sylvester et al., 2010; Carleton et al., 2008; Hulsey et al., 2007; Streelman & Danley, 2003). Species from these two groups also differ in craniofacial morphology, as they are adapted for specific feeding specializations. This craniofacial diversity may also be indirectly altered by other factors, such as social behavior and environment, but the main reason is probably diet. Many mbuna feed on the algae attached to the rock or comb through the algae for insects or loose auwfuchs. Utaka generally feed on insects in the sand or on organisms in the water column. In both habitat types, there are generalist feeders whom will feed from the algal beds or water column (Muschick et al., 2011).

Humans commonly keep cichlids as pets. Aquarium cichlids have a feeding behavior similar to a general pet fish; they feed on fish flakes within the water column. However, others will also feed off of the algae growing in the tank. See this video below to see algal foraging cichlids in action!

The natural diversity in cichlid craniofacial skeleton makes them a great model for studying the developmental mechanisms underlying facial morphology. Parsons et al. (2014) conducted a study in order to understand how extreme facial phenotypes develop in algal feeding cichlids and to determine if there is a cost to such specialization.


Meet the Cichlids

Labeotropheus fuelleborni (LF)

LF Male (

L. fuelleborni is a mbuna cichlid that feeds on filamentous algae by shearing it from the rock surface. To withstand the stress of this feeding method, they have adapted short, robust jaws with rounded craniofacial profiles (Figure 2).


Maylandia zebra (MZ)

MZ Male (

M. zebra is a generalist, mbuna species. It will feed on loose auwfuchs from the algal beds, as well as organisms from the water column. In relation to LF, it has a long jaw and shallow profile, but in relation to all cichlid species, it has an intermediate craniofacial morphology (Figure 2).


Tropheops sp. ‘red cheek’ (TRC)

TRC Male (

 T. sp. ‘red cheek’ is an mbuna cichlid that feeds on loose auwfuchs. It does this by combing through the algae attached to the rocks. This species has a craniofacial morphology that is in between LF and MZ (Figure 2)


a) LF, b) TRC, c) MZ

Figure 2: Craniofacial variation in mbuna cichlids: a) LF, b) TRC, c) MZ (Parsons et al., 2014)


Wingless (Wnt) signaling

Wnt signaling is essential and plays many roles throughout embryonic development. Secreted glycoproteins bind to Wnt receptors, leading to a signaling cascade inside the cell. In craniofacial development, Wnt acts through the canonical ß-catenin-dependent pathway. Two key pathway members in this Wnt signaling path and in craniofacial development are ß-catenin and lef1 (Willert & Nusse, 1998). Parsons et al. (2014) uses ß-catenin and lef1 to study Wnt signaling in craniofacial development and evolution.

Figure 3: Canonical ß-catenin dependent Wnt signaling pathway (Mikesch, 2007)


What Parsons et al. (2014) Found:


H1: “Wnt pathway members [are] expressed within the preorbital region of the teleost skull during craniofacial development”

H2: LF will have higher levels of Wnt signaling during craniofacial development than MZ

H3: Wnt signaling mediates the outgrowth of the preorbital region and relates to species differences

H4: LF craniofacial skeleton is less influenced by alternate feeding behaviors than that of TRC


Wnt contribution to phenotypic novelty

Parsons et al. used in situ hybridization in cichlids and a Wnt reporter line in zebrafish (Tg(7xTCF-Xla.Siam:GFP)) to characterize ß-catenin and lef1 signaling in the developing head. Wnt signaling pathway members were expressed in the developing preorbital region of the skull in all three species, in support of H1. However, in Figure 4 we can see differences in the domains of expression; LF (algal scraper) had expanded regions of Wnt singaling relative to MZ (generalist), shown by whole mount in situ hybridization images and in support of H2.

Figure 4: ß-catenin and lef1 expression in LF (e,f) and MZ (b,c) (Parsons et al., 2014)

Once they established the locality of Wnt signaling, they experimentally manipulated Wnt signaling in zebrafish, LF and MZ to test H3. Lithium chloride (LiCl) is a known agonist of Wnt signaling, meaning it increases Wnt activity. A known Wnt antagonist, IWR-I, can cause the opposite effect.

When zebrafish were treated with LiCl, they showed increased Wnt signaling, which was phenotypically manifested as faster bone development and ectopic bone formation down the body-line (Figure 5).

Figure 5: Wnt expression and maninpulation in the zebrafish transgenic line (Tg(7xTCF-Xla.Siam:GFP)) (Parsons et al, 2015)

The cichlid chemical treatments resulted in a striking phenotypic effect! MZ with increased Wnt signaling (LiCl treated) developed a facial structure similar to LF, while LF with decreased Wnt signaling (IWR-I treated) developed a facial structure to MZ (Figure 6). This further supports that idea that Wnt is an essential pathway for craniofacial development and that it has played a key role in natural facial variation in cichlids.

Figure 6: Craniofacial alterations due to manipulation of Wnt signaling (Parsons et al., 2014)


Wnt-mediated facial morphology and evolutionary potential

Data from in situ hybridizations and chemical treatments suggest that the novel LF craniofacial phenotype is influenced by Wnt signaling. Next, Parsons et al. addressed whether the Wnt-derived novel phenotype influences phenotypic plasticity. The study consisted of forcing fish to feed with a biting or sucking action, to test if this feeding behavior would induce changes in the jaw length. For this study they used LF, an algal feeder with the novel phenotype, and TRC, an algal feeder with less extreme specialization. They found that TRC craniomorphology changed in the direction of MZ (elongate jaw and shallow profile) while LF craniomorphology did not segregate with forced water column feeding (Figure 7).

Figure 7: Images of LF (a,b) and TRC (d,e) before and after forced water column-feeding treatments (Parsons et al., 2014)

This suggests that the increased specialization of LF comes at a cost; LF is more resistant to environmental changes, which could limit their evolvability in with changing ecological pressures.



H1: “Wnt pathway members [are] expressed within the preorbital region of the teleost skull during craniofacial development”

Yes, they found that Wnt is expressed in the preorbital region of multiple cichlid species, as well as zebrafish, during craniofacial development.

H2: LF will have higher levels of Wnt signaling during craniofacial development than MZ

Yes, using in situ hybridization they found that LF has higher endogenous levels of Wnt signaling during craniofacial development than MZ.

H3: Wnt signaling mediates the outgrowth of the preorbital region and relates to species differences

Yes, not only did they visualize higher Wnt levels in the species with a a shorter/rounder preorbital region, but they also were able to show this function of Wnt. In cichlids, they manipulated Wnt signaling and made LF look like MZ and vice versa. In addition, they showed that increasing Wnt signal in zebrafish leads to a reduction in preorbital outgrowth, mimicking the effect it has in LF.

H4: LF craniofacial skeleton is less influenced by alternate feeding behaviors than that of TRC

Yes, they were able to show that TRC (intermediate craniofacial morphology) shows greater phenotypic plasticity in response to forced water-column feeding, in relation to LF.


Overall, Parsons et al. found that Wnt signaling is active early in craniofacial development and that variation in Wnt activity influences species-specific differences, contributing to the novel, algal scraper phenotype. Later in ontogeny, environmental variables further shape the facial structure. However, adaptation of the extreme LF phenotype decreases the ability of the individual to phenotypically respond to environmental changes. Figure 8 presents a concise model for the patterning and shaping of cichlid craniofacial structure from larval to adult stages.

Figure 8: A proposed model of Wnt signaling and environmental effects on craniofacial development and remodeling (Parsons et al., 2014)



This paper was nicely laid out with an elegant design. I like how Parsons et al (2015) started with characterizing the expression of Wnt, then added a functional component by manipulating the Wnt signal. This was they were able to understand where Wnt is expressed and the effect it has on the organism during development. In addition to this, they took it a step further and determined the level of phenotypic plasticity, in addition to evolution, to understand how evolution can be constrained at novel phenotypes. I think that it would have been interesting if they included MZ, the generalist species, in the plasticity assay. In this study, it was the most general species, making it seem that it would have the largest degree of plasticity. MZ was used as the least derived species in this study, showing the longest, shallowest preorbital region. When you look at all cichlid species in Lake Malawi, MZ have an intermediate profile. Rock dwellers (such as LF and TRC) have short rounded craniofacial profiles while sand dwellers have elongate, shallow profiles. It would be interesting to expand this experiment to include the five types: Novel sand, intermediate sand,  intermediate (MZ), intermediate rock (TRC), and novel rock (LF). It would be interesting to see if sand dwellers also follow this trend in Wnt signaling, with decreased Wnt signaling leading to elongate/shallow profiles. Also, it would be interesting to determine if MZ might actually have the widest degree of plasticity, meaning the intermediate profile in the lake would have the highest degree of adaptability.



Carlton, K.L., T.C. Spady, J.T. Streelman, M.R. Kidd, W.N. McFarland & E.R. Loew. 2008. Visual sensitivities tuned by heterochronic shifts in opsin gene expression. BMC Biology 6: 22.

Hulsey, CD., M.C. Mims & J.T. Streelman. 2007. Do constructional constraints influence cichlid craniofacial diversification? Proceedings of the Royal Society B 274: 1867-1875. 

Loh Y.H.E., L.S. Katz, M.C. Mims, T.D. Kocher, S.V. Yi & J.T. Streelman. 2008. Comparative analysis reveals signatures of differentiation amid genomic polymorphism in Lake Malawi cichlids. Genome Biology 9:R113.

Kocher, T.D. 2004. Adaptive evolution and explosive speciation: the cichlid fish model. Nature Reviews Genetics 5: 288-298.

Kocher, T.D. & R.C. Albertson. 2006. Genetic and developmental basis of cichlid trophic diversity. Heredity 97: 211-221. 

Mikesch, J.H., B. Steffen, W.E. Berdel, H. Serve & C. Muller-Tidow. 2007. The emerging role of Wnt signaling in the pathogenesis of acute myeloid leukemia. Leukemia 21: 1638-1647. 

Muschick M., M. Barluenga, W. Salzburger & A. Meyers. 2011. Adaptive phenotypic plasticity in the Midas cichlid fish pharyngeal jaw and its relevance in adaptive radiation. BMC Evolutionary Biology 11: 116.

Parsons, K.J., A.T. Taylor, K.E. Powder & R.C. Albertson. 2014. Wnt signaling underlies the evolution of new phenotypes and craniofacial variability in Lake Malawi cichilds. Nature Communications 5: 3629. 

Streelman, J.T. and P.D. Danley. 2003. The stages of vertebrate evolutionary radiation. TRENDS in Ecology and Evolution 18: 126-131.

Sylvester J.B., C.A. Rich, Y.H.E. Loh, M.J. Staaden, G.J. Fraser & J.T. Streelman. 2010. Brain diversity evolves via differences in patterning. PNAS 107: 9718-9723. 

Willert, K. & R. Nusse. 1998. ß-catenin: A key mediator of Wnt signaling. Current Opinion in Genetics and Development 8: 95-102.

Leave a Reply

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