Regeneration restores the diversity and the pattern of sensory organs in the legs of Parhyale

Get Complete Project Material File(s) Now! »

Regeneration restores the diversity and the pattern of sensory organs in the legs of Parhyale


Many animals have the ability to regenerate parts of their body that have been injured or amputated, including organs with complex architectures and multiple differentiated cell types. The regenerated organs resemble the injured organs that they replace, but it is usually not known whether they represent perfect or imprecise replicas of the original structures. Here we address this question in the regenerated limbs of the crustacean Parhyale hawaiensis, a genetically tractable system where it is possible to elucidate the cellular origins of regenerated tissues. We focus on the array of external sensory organs that decorate the distal part of Parhyale legs. These sensory organs represent complex markers that allow us to track the accuracy of regeneration in fine detail. We describe eight types of external sensory organs present on Parhyale legs, distributed in stereotypic positions or patterns on the surface of the leg. We find that regenerated legs carry the full diversity of external sensory organs, distributed precisely in their original locations. The numbers of regenerated sensory organs are indistinguishable from those found in size-matched un-amputated limbs. We conclude that regeneration faithfully restores the array of external sensory organs in the legs of Parhyale.
The ability to regenerate varies widely among animals. On one extreme of the spectrum are animals that are able to regenerate any part of their body perfectly from a small body fragment, such as some species of planarians and hydrozoans (Morgan, 1901). On the other extreme are animals that are incapable of regenerating (e.g. nematodes) or whose capacities to regenerate are limited to physiological turnover of some tissues (e.g. skin or blood renewal in mammals). Between these extremes are a wide variety of animals whose capacities to regenerate are limited to specific organs, particular stages in their lifecycle, or whose ability regenerate is imperfect. Well known examples of the latter category are lizards, whose regenerated tails replace bony vertebrae by cartilage (Goss, 1969), cockroaches, whose regenerated legs have differences in nerves, tracheae and musculature (Kaars et al., 1984), or fish, whose regenerated heart tissue beats asynchronously with respect to the rest of the heart (González-Rosa et al., 2014). In many cases regeneration is known to produce organs that have a normal appearance, but it is difficult to assess whether regeneration is perfect down to the fine structural and cellular details.
From an evolutionary point of view, the mechanisms of regeneration need only produce an organ or tissue that restores normal function. So long as function is restored, there is no added evolutionary benefit for regeneration to be perfect.
Here we focus on leg regeneration in arthropods, where external chemosensory and mechanosensory organs composed of small groups of cells are clearly visible on the external surface of the leg. These organs have been studied extensively in the fruit fly Drosophila, where each sensory organ typically consists of 4 cells – including hair, socket, neuron and sheath cells – which arise through stereotypic divisions from a single precursor cell (reviewed in Hartenstein, 2005). External sensory organs come into contact with the environment through highly specialised structures that protrude from the surface of the cuticle in the form of bristles or setae. These structures are produced by a single cell, the hair (trichogen) cell, in Drosophila, but could arise from several cells in other species (Garm and L. Watling, 2013; Guse, 1983). These structures can take a variety of shapes and sizes, depending on the type of sensory organ and its functions in each species. Within a given species, however, the morphology, cellular composition and spatial distribution of each type of sensory organ are usually stereotypic and well conserved (Hartenstein, 2005). Thus, sensory organs can provide markers for assessing the accuracy of regeneration with very high (almost cellular) precision.
Among arthropods, malacostracan crustaceans retain the ability to regenerate their limbs throughout their lifetime (Charmantier-Daures and Vernet, 2004; Maruzzo and Bortolin, 2013). We recently introduced the amphipod crustacean Parhyale hawaiensis as an experimental system for studying regeneration, taking advantage of genetic approaches (transgenesis, CRISPR-mediated gene editing) and genomic resources available in this species (Kao et al., 2016; Martin et al., 2016; Pavlopoulos and Averof, 2005). The ability to perform live imaging and to track cells during the course of regeneration (Alwes et al., 2016) makes Parhyale an attractive system in which we can study the mechanisms of regeneration at single-cell resolution. In this context, the external sensory organs of Parhyale limbs can serve as markers to investigate the fidelity of regeneration. Here, we describe the diversity and the pattern of sensory organs found in the distal part of Parhyale limbs. We assess the extent to which these structures are faithfully recovered and their functions restored following leg regeneration.

Results and discussion

The diversity of setae on Parhyale limbs

We have used scanning electron microscopy (SEM) to survey the surface of Parhyale limbs, focusing on the three most distal podomeres (the dactylus, propodus and carpus) of the T4 and T5 pereopods. These two limbs show almost identical patterns of setae (see below). The cuticle of the limbs has a polygonal pattern (Figure 2.2a), which reflects the architecture of underlying epithelial cells (Havemann et al., 2008), (Dillaman et al., 2013). On the limb surface we observed several types of setae, which we categorised as follows based on previous studies (Garm, 2004; Garm and L. Watling, 2013; Les Watling, 1989):
– Lamellate setae: setae consisting of a smooth shaft bearing a series of lamellae towards the tip (figure 2.1a,b). We identified two variants of lamellate setae: type-1 have a wider base and a shorter shaft with a terminal pore (figure 2.1a), type-2 tend to have a more slender base, a longer shaft and no pore (figure 2.1b).
– Plumose setae: setae with a long shaft, bearing two opposed rows of long setules, which give it a feathery appearance (figure 2.1c). There are no visible pores.
– Twin setae: composites of cuspidate and lamellate setae (figure 2.1d). The cuspidate-like main shaft bifurcates into a branch that resembles the tip of a typical lamellate seta with a terminal pore.
– Curved setae: simple setae with a long twisting shaft, bearing a pore approximately 2/3 along the length of the shaft (figure 2.1e).
– Hooked setae: simple setae with a long thin shaft tapering gradually towards the apex, a hooked shape, and a series of fine nicks prior to the tapering distal region. The shaft has a terminal pore (figure 2.1f).
– Cuspidate setae: relatively short and stout setae, bearing longitudinal ridges and no pore (figure 2.1g).
– Microsetae: very small setae bearing a terminal pore covered by a hood. One side of the shaft has a lamellated appearance. Microsetae are associated with characteristic dimples on the cuticular surface (figure 2.1h).
(a) Type-1 and (b) type-2 lamellate setae, (c) plumose seta, (d) twin seta, (e) curved seta, (f) hooked seta with some dirt on the shaft, (g) cuspidate seta, and (h) microseta with a specific design of the cuticle surrounding it. The scale bars are 5µm.
In addition to these sensory structures, we observe that the surface of limbs bears numerous pores, located at the junctions of epidermal cells (figure 2.2). These pores do not appear to be associated with axons (figure 2.7), we therefore conclude that they are unlikely to have a sensory function.
(a) The cuticle is organised in a polygonal pattern. The pores are covering all the surface except the ventral side where the polygons have a different architecture. (b) The arthrodial membrane covering the joint of two podomeres has no pores. (c) There are three different types of pores: small, medium, and large; labelled with arrowhead, double arrowhead and arrow respectively. (d-g) Large pores have different microstructures inside indicating different subtypes. The scale bars are 10µm for a-c, and 250nm for c-g.

The distribution of setae on Parhyale limbs

We find that each type of seta has a well-defined distribution on the distal part of Parhyale T4 and T5 limbs: either in stereotypic positions, or in arrays of several setae arranged in specific patterns (figure 2.3). The only exception to this stereotypic arrangement are the microsetae, which are well spaced on the surface of the cuticle but without an apparent conserved pattern (figure 2.3h).
The most distal podomere, the dactylus, bears three distinctive large setae placed in specific locations: a hooked seta located at the base of the terminal claw, a large curved seta on the ventral side, and a plumose seta on the dorsal side (figure 2). This distribution was invariable in all the limbs we inspected (n=22).
In the next two podomeres, the propodus and the carpus, most of the large setae are organised in four distinct groups. The first group, which we name the ‘crown’, consists of a row of type-2 lamellate setae, located on the dorsal margin at the distal end of each podomere (figure 2.2f-f’). The crown is present in every individual, but consists of a variable number of setae; we found that on average the propodus has 3.3 setae (s.d. 1.3, n=23) and the carpus has 3.0 setae (s.d. 0.9, n=23) on the crown (figure 2.4a and figure 2.6a).
The SEM image of the first walking leg of Parhyale (T4) focusing on three distal-most podomeres: dactylus (turquiose), propodus (green), and carpus (purple). The dactylus bears a terminal claw (a) and three distinctive setae: hooked (b), curved (c), and plumose (d) setae. The propodus and carpus has the crown groups (f, f’) and the comb groups (e, e’) both consist of a tight row of type-2 lamellate seta. The last group on the propodus and carpus is the row group: an array of small clusters of twin and type-1 lamellate setae on the ventral side of the podomeres (g, g’). Apart from the setal groups, there are several microsetae covering the posterior surface of the limb (h). The dashed line is marking the amputation plane.
The second and third groups, which we name ‘combs’, are tight rows of type-2 lamellate setae located in the distal part of the propodus and carpus. One comb is found consistently on the posterior face of the propodus and the carpus (figure 2.3e-e’). A second comb is usually found on the anterior face of the propodus (figure S2.1), but occasionally this is reduced to a single seta. On the carpus, single type-2 lamellate setae are found in place of the anterior comb. We find that each comb consists of 1 to 6 setae, with an average of 3.6 setae (s.d. 1.1, n=22) at the posterior of the propodus, 2.6 setae (s.d. 1.0, n=22) at the anterior of the propodus, and 3.0 (s.d. 0.9, n=23) at the posterior of the carpus (figure 2.4b and figure 2.6b).
The fourth group, which we name the ‘ventral’ setae, consists of an array of setae that are distributed with a regular spacing along the ventral side of the propodus and carpus (figure 2.3 g-g’). The number of elements in these ventral arrays of setae correlates with the length of the podomere, suggesting that new elements are added to this pattern as the limbs grow during the lifetime of Parhyale (figure 2.5). Each element of the ventral array is made up exclusively of twin setae and type-1 lamellate setae, but there is considerable variation in the number of setae per element (figure 2.4c, figure 2.6c). In the propodus, most elements consist of a twin seta surrounded by type-1 lamellate setae; in the carpus they often consist exclusively of type-1 lamellate setae (see supp. Table S1). As we discuss later, type-1 lamellate setae may develop into twin setae, so this variation could reflect different stages in the maturation of these elements, as new ones are added during growth. We found that on average the propodus has 3.26 sets of ventral setae along its length (s.d. 0.86, n=22) and the carpus has 2.82 (s.d. 0.96, n=23) (figure 2.4b and figure 2.6b).
In addition to these groups, the propodus bears a single cuspidate seta ventrally, near the joint, on the most distal end of the podomere (n=23, supp. figure S2.1).
The distal parts of T4 and T5 legs carry almost identical patterns of setae (see suppl. Table S1), in subsequent analyses we have therefore combined the data from these limbs.
Figure 2.4: Quantification of setae on the propodus of uninjured and regenerated legs. Quantification of the number of seta in the (a) crown and (b) comb groups, and (c) number of arrays on the ventral side of the propodus in unamputated and regenerated legs. Every point represents the seta on a single propodus (blue for control, orange for regenerated legs). Green bars correspond to the mean values.

The pattern of sensory organs is fully restored in regenerated Parhyale limbs

Following the identification of these setae, we investigated to what extent these structures are faithfully regenerated following amputations at the carpus of T4 and T5 limbs (data in suppl. Table S2.1). In order to minimise genetic, environmental or age effects on the number and patterns of setae, we measured the setal patterns of control (unamputated) and regenerated limbs simultaneously in the same cohort of individuals: the observations on regenerated limbs were carried out on T4 and T5 limbs that were amputated and allowed to regenerate on one side of the animal, while the observations on unamputated limbs (described in the section above) we carried out on the contralateral limbs of the same individuals.
On the dactylus, we find that the stereotypic hooked, curved and plumose setae described earlier consistently regenerate in their original positions, with no exceptions (n=16). In the propodus and the carpus, we find that all the patterns of sensory organs that we described – the crown, the combs and the ventral setae – are also regenerated, but the number of setae in each pattern may vary depending on the size of the regenerated podomere.
Thus, the propodus of regenerated T4 and T5 limbs bears a crown (with 2.67 type-2 lamellate setae on average, s.d. 1.33, n=23), anterior comb (with 1.57 type-2 lamellate setae on average, s.d. 0.65, n=22), posterior comb (with 2.07 type-2 lamellate setae on average, s.d. 0.62, n=22), ventral setae (4.35 sets of twin and/or type-1 lamellate setae on average, s.d. 0.97, n=23), and the distally-located cuspidate seta (found in 17 out of 18 cases) (figure 2.4).
Regenerated limbs are smaller than their unamputated contralateral limbs in the first molt following amputation (Alwes et al., 2016) and gradually recover their full size during subsequent molts. We therefore reasoned that the smaller number of setae observed in a regenerated propodus might reflect the smaller size of the field in which they develop. To test this idea, we quantified the number of sets of setae found in the ventral array of the propodus in relation to the length of this podomere, on the images obtained by scanning electron microscopy. To add to that limited dataset, we imaged additional unamputated and regenerated T4 or T5 limbs on a laser scanning confocal microscope, exploiting the cuticle’s autofluorescence to observe the sensory organs in the ventral arrays (figure 2.5). In the combined dataset of SEM and confocal data (n=29), we find that the length of the propodus is ~18% smaller in the regenerated limbs compared with their unamputated contralateral limbs. The propodus of regenerated limbs harbour the similar numbers of setae units on their ventral array as size-matched unamputated limbs of 350-450 µm in length. We need to sample unamputated limbs with smaller propodus (<350 µm) and regenerated limbs with a larger propodus (>450 µm) to extend this comparison over a wider range of limb sizes.. These data suggest that the regenerated limbs bear similar numbers of setae as size-matched unamputated limbs. However, we need to image and quantify setae from more unamputated limbs, especially smaller ones, to clarify this correlation.
In our experiments the site of amputation is the distal part of the carpus (see figure 2). Thus, distal elements in the carpus (such as the crown and the combs) are removed by the cut while proximal elements are retained. We find that this partly regenerated podomere bears all the patterns of sensory organs found in T4 and T5 limbs, including a crown (with 2.41 type-2 lamellate setae on average, s.d. 1.06, n=17; figure 5a), anterior comb (with 1.21 type-2 lamellate setae on average, s.d. 0.43, n=18), posterior comb (with 3.28 type-2 lamellate setae on average, s.d. 0.99, n=17), and ventral setae (2.50 sets of twin and/or type-1 lamellate setae on average, s.d. 0.70, n=18) with a composition that largely resembles the composition found in unamputated limbs (figure 2.6).
Figure 2.5: Correlation between the number of ventral arrays and the length of the podomere for propodus.
Scatter plot of the size of the propodus versus the number of ventral arrays of unamputated control (blue circles) and regenerated (orange circles) legs.
In addition to the recovery of all types of large setae, we find that the microsetae and the pores which are distributed on the limb surface are also restored following limb regeneration. We did not quantify the density and distribution of microsetae and pores.

READ  dynamical symmetry of the 2d bose gas

Putative sensory functions of different types of setae

Stainings using an antibody for acetylated-tubulin, which labels nerve axons, as well as transgenic animals expressing lyn-tdTomato (Alwes et al., 2016), which allows us to visualise some neurons, show that all the types of setae that I described are innervated, as expected of sensory organs (figure 2.7).
Based on what we know from other arthropods, we expect these external sensory organs to have mechanosensory and/or chemosensory functions. Although it is not easy to identify the precise sensory functions of each type of seta, we could speculate on their functions based on their morphology and some previously defined functions of homologous setae in other crustaceans. To begin with, the presence of a pore on the apex of a seta is indicative of a chemosensory function (Brandt, 1988; Garm and L. Watling, 2013), therefore, we can consider type-1 lamellate setae, curved setae, and microsetae as putative chemosensory organs.
Figure 2.6: Quantification of the setae on the carpus of uninjured and regenerated legs. Quantification of the number of seta in the (a) crown and (b) comb groups, and (c) number of arrays on the ventral side of the carpus in unamputated and regenerated legs. Every point represents the seta on a single carpus (blue for control, orange for regenerated legs). Green bars correspond to the mean values.
Conversely, setae that bear setules along their shaft, like plumose setae, have never been shown to be chemosensitive (Garm and L. Watling, 2013). They either have non-sensory functions, such as generating stroking power for swimming or used as combs for moving food into the mouth, or they can have mechanosensory functions. In Parhyale pereopods, it is unlikely that they would have a structural function since there is only a single plumose seta positioned on the distal part of the leg. However, they could serve as mechanosensory organs, since two rows of long setules on the shaft makes them very sensitive for waterborne vibrations and might also provide information on the direction of the signal (Fields et al., 2002; Vedel, 1985).
The comb and crown groups consist of rows of type-2 lamellate setae of different length. With such organisation, they could provide 3D sensation of mechanical stimuli collectively, as rodent whiskers do (Carvell and Simons, 1990).
Figure 2.7: All setae of both uninjured and regenerated limbs are innervated.
(a) An uninjured and (b) a regenerated T5 limb from the same individual carrying the PhHS-lyn::tdTomato-2A-H2B::EGFP transgene, which expresses lyn-tagged tdTomato (marking cell outlines, in red) and histone-EGFP (marking cell nuclei, in green). The tdTomato label allows neurons to be tracked from the setae on the surface of the epidermis to the main nerve of the limb. In the insets, clockwise from top left: microsetae, type-2 lamellate seta, plumose seta, hooked seta, curved seta, cuspidate seta, lamellate-1 and twin setae, and type-2 lamellate setae. The axons of the neurons cannot be seen in this focal plane.
The dactylus has no muscles and possibly also no chordotonal organs. The cuspidate seta beneath the dactylus-propodus junction might be important to perceive the bending of the dactylus.
Twin setae have two components, resembling the cuspidate and lamellate setae. The lamellate part of twin setae is likely to have a function in chemoreception, as suggested by the terminal pore on the apex, and the cuspidate part could have mechanosensory functions. Apart from putative mechanosensory functions, the sturdy cuspidate setae might also perform mechanical functions, such as clinging to the substrate or manipulating food (Garm, 2004). Considering that Parhyale spends most of its life clinging on a substrate, the robust twin setae on the ventral side of the legs could be used for attachment to the substrate.

Recovery of sensory function after regeneration

All the setae of regenerated legs are innervated, similar to unamputated legs (figure 2.7), which suggests that, besides recovering their detailed morphology, regenerated Parhyale limbs may have also recovered their sensory functions. To test this hypothesis we developed a simple assay for the mechanosensory functions of limb setae. We found that touching the ventral row or comb setae of T4 or T5 limbs with a solid object evokes stereotypical escape response (Video 2.1). The assay is robust: 95% (confidence interval (CI) 90-98%) of stimulations to the carpus comb setae resulted in a response in T4 and T5 limbs of immobilised animals (n=22, figure 2.8, table S2.2). After amputating one of these limbs, either T4 or T5, at the carpus-propodus junction, and the assay was repeated 6 days post amputation (dpa): the response rate of the amputated limbs dropped to 8% (CI 3-25%), whereas the unamputated T4 or T5 control limbs in the same animals remained highly responsive, with an average of 98% (CI 92-100) response rate. The unresponsiveness of the amputated limbs was expected since the sensory organs at the distal part of the limb stump are thought to de-differentiate and become disconnected from the setae (figure 2.8a, see chapter 3 for details). Therefore the setae located on the distal part of the stump are not expected to respond to the stimuli at the later stages of regeneration. Once limb regeneration was complete and the animals molted, the response rate of both control and regenerated limbs were the same, 88% (CI 74-95%), which is comparable to the response rate of before amputation. These results suggest that the mechanosensory function is fully restored during leg regeneration.

Table of contents :

Unipotent and multipotent progenitors of cnidarian regeneration
Pluripotent progenitors mediate planarian whole-body regeneration
De-differentiation and lineage-restricted progenitors of regeneration in vertebrates
Arthropod leg regeneration relies on lineage-restricted progenitors
Arthropod peripheral sensory organs
Research objectives
2. Regeneration restores the diversity and the pattern of sensory organs in the legs of Parhyale
Results and discussion
The diversity of setae on Parhyale limbs
The distribution of setae on Parhyale limbs
The pattern of sensory organs is fully restored in regenerated Parhyale limbs
Putative sensory functions of different types of setae
Recovery of sensory function after regeneration
Transformation of type-2 lamellate setae into twin setae
Materials and Methods
Parhyale culture
Scanning Electron Microscopy (SEM)
Laser Scanning Confocal Microscopy
Testing mechanosensory function
Author contributions
Supplementary Information
3. Continuous live imaging of Parhyale limb regeneration
Results and discussion
Extending and optimizing continuous live imaging of regenerating limbs
A complete overview of limb regeneration
4. CRISPR-mediated knock-in approach to generate cell type -specific marker lines of Parhyale peripheral sensory organs
The emergence of precise genome editing
CRISPR revolution
Design and synthesis of CRISPR-mediated genome editing reagents
DSB repair pathways: how to insert transgenes into genomes
CRISPR-mediated genome editing in Parhyale
Results and Discussion
Antp knock-in as a positive control
Designing sgRNAs for new targets
Targeting Pax3/7-2
Selecting putative sensory organ markers
Targeting the neuronal marker futsch
Testing sgRNA efficiencies on cut, sens, and elav
Overall lessons from CRISPR experiments
5. Generating and screening CRE-reporters to label the sensory organs of Parhyale legs
Identification of enhancers in the genome
Assay for transposase-accessible chromatin
using sequencing (ATACseq)
Methods for the functional characterization of enhancers
Identifying CREs in Parhyale hawaiensis
Results and discussion
ATACseq on Parhyale embryos and limbs
Selecting putative sensory organ markers and testing
putative CRE reporters
The DC5 reporter and enhancer traps
Supplementary Information
6. Combining live imaging with antibody staining to track the progenitors of Parhyale limb regeneration
The methodology of staining tissues with antibodies
Antibody staining in Parhyale
Results and Discussion
Developing an antibody staining protocol for regenerating legs
Registration of nuclei between live imaging and IHF stainings
Putative markers for recognizing sensory organ cell types using antibodies
Protein expression and immunization
Testing the antibodies raised against Parhyale proteins by IHF
7. Conclusion and perspectives
High fidelity regeneration of Parhyale peripheral sensory organs
Live recording the entire course of Parhyale limb regeneration
Transgenesis in Parhyale
Combining antibody staining with live imaging to identify progenitors in
regenerating limbs
Studying cell and tissue dynamics during Parhyale limb regeneration
A – Materials and Methods
Parhyale husbandry
Animal preparation for live imaging
Antibody staining
Microinjections for CRISPR
Image acquisition and analysis
Detailed Protocols
Fixation of adult Parhyale hawaiensis for SEM
RNA in situ hybridization in Parhyale hawaiensis embryos in vitro sgRNA synthesis
Microinjection to Parhyale embryos
Genomic DNA extraction from Parhyale embryos
T7 endonuclease assay
Antibody staining on Parhyale embryos
ATACseq on Parhyale embryos
Total RNA extraction from Parhyale tissues
B – Tracking cell lineages in 3D by incremental deep learning
Main text
Competing interests
C – References


Related Posts