Functional Recovery Associated with Dendrite Regeneration in PVD Neuron of Caenorhabditis elegans

Harsh touch response is affected by axotomy of PVD neuron but it remains unaffected by dendrotomy

PVD neuron displays an orthogonal array of dendritic branches that cover the major part of the body (Fig. 1A; Inberg et al., 2019). The higher-order branches are arranged in a menorah-like fashion (Fig. 1A; Oren-Suissa et al., 2010). Since both harsh touch sensation and proprioception are largely controlled by PVD neurons, it is not clear whether one or both modalities will be affected by dendrotomy or axotomy. Therefore, to understand whether dendrite regeneration would lead to functional restoration, it is important to have a quantitative behavioral deficit in a specific behavior caused due to dendrotomy. We first studied the effect of axonal and dendritic injury on the harsh touch response behavior. When the worms were prodded near the vulva with a platinum wire, they instantly exhibited an escape response (Fig. 1B, red mark, Movie 1) as seen before (Li et al., 2011). The distance traveled by the worm in each frame was shorter before the delivery of the harsh touch stimulus but it increased thereafter (Fig. 1C, Movie 1). Nearly 90% of the wild-type worms responded to the harsh touch (Fig. 1D) and the response remained comparable in the L4 and Day 1 stages, with a 20% decrease at the Day 2 stage (Fig. 1D). The percentage of animals responding to harsh touch was significantly reduced in mec-3 mutant (Fig. 1D). The mec-3 gene codes for a transcription factor, which is required for the expression of genes important for the development and function of PVD (Smith et al., 2013). However, this sensory behavior was unaffected in the gentle touch defective mutant of mec-4 (Fig. 1D), which codes for the DEG/ENaC mechanoreceptor channel for gentle touch (Chalfie and Sulston, 1981). This indicated that our assay condition is specifically sensitive to harsh touch.

We ablated PVD neurons on one or both sides (PVDL and PVDR, Fig. 1E,F) at the L4 stage to assess the role of these neurons in harsh touch response. The percentage of worms responding to harsh touch was significantly reduced when both the PVD neurons (PVDL and PVDR) were ablated as compared to the control (mock injury) (Fig. 1F). However, in cases of ablation on one side, only the harsh touch given to ablated side resulted in a significant reduction in response (Fig. 1F). Similarly, we found that following axotomy near the cell body (red cross, Fig. 1E,G), the percentage of worms responding to harsh touch decreased when both the PVD neurons were axotomized and on the axotomized side when one axon was cut (Fig. 1G). Surprisingly, when we dendrotomized both the major and minor primary dendrites on both sides (PVDL and PVDR, Fig. 1E,H), the harsh touch response remained unaffected (Fig. 1H). This indicated that dendrotomy does not affect the harsh touch response behavior associated with PVD neurons.

We also developed a more quantitative assessment of harsh touch response by analyzing video recordings of the worm's response to a harsh touch stimulus. We used an Image-J plugin software (Nussbaum-Krammer et al., 2015) that tracks the worm's speed (Extended Data Fig. 1-1A). The speed of the locomotion increased significantly following the delivery of harsh touch (Extended Data Fig. 1-1B). The Day 1 and Day 2 adults also showed a similar increase in speed upon harsh touch (Extended Data Fig. 1-1B). We defined the harsh touch response index (HTRI) as the ratio of the worm's speed after to before the harsh touch delivery (Extended Data Fig. 1-1A). As expected, the HTRI value in the mec-3 mutant is significantly lower than in the wild type (Extended Data Fig. 1-1C). On the other hand, the HTRI was unaffected in mec-4 mutant worms compared to wild-type worms (Extended Data Fig. 1-1C). Additionally, HTRI levels were significantly lowered after PVD ablation or axotomy (Extended Data Fig. 1-1D,E). This decline was specific to the injured side (Extended Data Fig. 1-1D,E). Dendrotomy of major plus minor dendrites of both of the PVD neurons did not affect the HTRI value (Extended Data Fig. 1-1F). To answer the question that whether the dendritic morphology may play any role in harsh touch response or not, we checked the harsh touch response for the mutants that have a developmental defect in the formation of the dendritic arbor of PVD neuron such as hpo-30(0) and tiam-1(0) (Zou et al., 2018; Tang et al., 2019). The harsh touch response was unaffected in these mutants (Fig. 1D) saying that the dendritic morphology may not be playing any significant role in harsh touch response. This indicated that harsh touch sensory modality may not be entirely dependent on the dendrites of PVD neurons.

The dendrotomy on PVD affects proprioception behavior

As the harsh touch response behavior was not affected by dendrotomy, we speculated that the proprioception behavior might be perturbed by dendrotomy. The dendrites release the neuropeptide NLP-12 to the neuromuscular junction that modulates the sinusoidal wave pattern formed during the movement of the worms on bacterial lawn (Tao et al., 2019). The amplitude and wavelength of the sinusoidal waveform made by a freely moving worm on the bacterial lawn (Fig. 2A) have been described as the hallmarks of proprioceptive behavior in worms (Albeg et al., 2011; Tao et al., 2019). These parameters are often altered in mutants missing higher-order dendritic branches in PVD neurons (Albeg et al., 2011; Tao et al., 2019). We used these parameters to test if this proprioception behavior was changed after dendrotomy. We took images of the sinusoidal traces made by the moving worms on the OP50 bacterial lawn on NGM plates (Fig. 2A). The amplitude, wavelength, and bending angle were computed from the sinusoidal trajectory (Fig. 2A). The mec-3, hpo-30, and tiam-1 mutants, which have reduced number of higher-order branches in PVD (Zou et al., 2018; Sundararajan et al., 2019; Tang et al., 2019), exhibited a decrease in all postural measurements (Extended Data Fig. 2-1A–C). The mutant for mec-10 which codes for DEG/ENaC channels also showed reduced values in these parameters (Extended Data Fig. 2-1A–C) as shown before (Tao et al., 2019). Ablation of the PVD cell body caused significant reduction in both the wavelength and amplitude of the waveform (Extended Data Fig. 2-1D–F). We found that the amplitude and wavelength values at Day 1 adult stage were significantly higher than those at larval stage L4 (Extended Data Fig. 2-1G–I), most likely due to an increase in the length of the worm body from 950 microns at L4 stage to 1.2 mm at Day 1 stage. These parameters were comparable in Day 1 and Day 2 stages (Extended Data Fig. 2-1G–I). Therefore, we decided to measure the changes in the parameters of body posture due to dendrotomy and regeneration at Day 1 and Day 2 stages, respectively.

We performed two types of dendrotomy experiments: (1) dendrotomy in major primary dendrites and (2) dendrotomy in both major and minor primary dendrites for both PVD (PVDL and PVDR, Fig. 2B). The trajectories of the worms that underwent dendrotomy appeared distinctly different as compared to the mock control (Fig. 2C). The postural parameters were decreased after dendrotomy of either major dendrites or major-plus-minor dendrotomy (Fig. 2D,F). The amplitude and bending angle were significantly decreased after the dendrotomy of the major primary (Fig. 2D,F). These parameters were reduced further significantly when both major and minor dendrites were cut, except the wavelength (Fig. 2D,F). The wavelength parameter was only perturbed with dendrotomy of major and minor dendrites (Fig. 2E). To see whether the defect in sinusoidal posture after laser injury is a secondary consequence of locomotion defect, we evaluated the baseline speed of the dendrotomized worms (Extended Data Fig. 2-1G). The speed of the worms remained unaltered postdendrotomy (Extended Data Fig. 2-1G). The gentle touch response was also unaffected in these worms (Extended Data Figure 2-1G, lower panel). When the PVD neurons were axotomized, we did not find any change in the postural parameters (Fig. 2G–I).

To ascertain if the dendrotomy-related drop in proprioception is nonspecific effect in one transgenic strain, we repeated this experiment in another transgenic strain expressing mScarlet diffusible reporter in PVD neurons (Extended Data Fig. 2-1H–L). The amplitude and bending angle both exhibited a reduction after dendrotomy of major as well as major-plus-minor dendrite in this strain (Extended Data Fig. 2-1I-L) whereas, the wavelength was dramatically reduced when both major and minor dendrites are severed (Extended Data Fig. 2-1K). We further assessed the relative contribution of PVDL and PVDR in the proprioception by injuring both major and minor dendrites on only one side (PVDL or PVDR) (Fig. 2J). When the injured PVD was in close-contact with the agar-surface of NGM plate, we observed a defect in the proprioception (Fig. 2K–M). Both the amplitude as well as wavelength parameters were affected in the one-side dendrotomy as compared to two-side dendrotomy experiments (Fig. 2K–M).

These experiments support the conclusion that the axons and dendrites of PVD neuron differentially regulate the harsh touch response and proprioception, respectively.

Axon regrowth of PVD neurons restores the impairment in harsh touch response

Previous studies have elucidated that axon regeneration from the injured cut stump leads to functional recovery in various model systems (Becker et al., 1997; Laha et al., 2017; Rasmussen and Sagasti, 2017; Basu et al., 2021). As PVD neurons also show axonal regeneration (Brar et al., 2022), we investigated whether the deficits in the harsh touch response due to PVD axotomy would be restored in subsequent time points during the course of axon regeneration. After the axotomy near the cell body (red arrow, Fig. 3A), we saw a retraction of the injured tip, followed by new neurite formation and regrowth toward the ventral side (Fig. 3A). We noticed that in 67% cases, the distal part of the axons apparently looked intact (Fig. 3A, labeled in red color) and the regrowing neurites reconnected with the distal counterpart and fasciculated toward the ventral nerve cord (Fig. 3A). The proportion of worms responding to harsh touch at 24 h after axotomy was higher than at 8–10 h after axotomy in both one-side and two-side axotomy paradigms (Fig. 3B). Similarly, we estimated an increase in HTRI at 24 h postaxotomy in a side-specific manner (Fig. 3C). This rapid recovery in touch sensation could be partly mediated by the fusion between the proximal and distal dendrite as described in gentle touch neuron (Ghosh-Roy et al., 2010; Basu et al., 2017).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Axon regeneration leads to recovery in harsh touch defect. A, Confocal images along with the schematics of PVD neuron is shown at 0, 6, and 24 h after axotomy representing the axon in orange, new protrusions in green traces, and orange arrowheads indicating the new regrowing branches. The red arrow marks the injury site. B,C, Quantification of harsh touch as a percentage of worms responding (B) and harsh touch response indices (HTRI) (C) of mock (24 h), axotomized-side and nonaxotomy-side in one-side axotomy (8 and 24 h), and two-side axotomy (8 and 24 h) in wdIs52 (pF49H12.4::GFP) worms. 24 < n < 32, N = 3 (B). 21 < n < 35, N = 3 (C). The functional recovery at 24 h postsurgery was also followed up in case of the ablation of PVD (Extended Data Fig. 3-1). D, The confocal images of regrowing axon in the wild-type and mlk-1(0) background at 24 h after injury. The red arrow marks the site of injury. E,F, Worms responding to harsh touch as percentage (E) and harsh touch response indices (F) were quantified and plotted for mock at 8 h, axotomy at 8 h, and axotomy at 24 h after axotomy of both the PVD neurons in wdIs52 and mlk-1(0); wdIs52 background. 18 < n < 35, N = 3 (E). 18 < n < 35, N = 3 (F). The statistical analysis B, and E, is Fisher's exact two-tailed test, for C, and F, is one way ANOVA with Tukey's multiple comparison tests with p < 0.05*, 0.01**, and 0.001***. Violin plots represent the median and population distribution. ns stands for not significant, N stands for the number of independent replicates, and n stands for the number of worms taken for analysis.

DLK/MLK pathway is a critical regulator of axon regrowth following injury (Yan et al., 2009; Nix et al., 2011; Brar et al., 2022). To study if DLK/MLK mediated axonal regrowth is required for functional recovery, we assessed the harsh touch response in the mlk-1 mutant (Fig. 3D). After axotomy, regrowth toward the ventral nerve cord was diminished in the mlk-1 mutant (Fig. 3D). We noticed that the percentage of mlk-1(0) worms responding to harsh touch at 24 h did not increase as compared to 8 h after injury (Fig. 3E,F). The HTRI values indicated a recovery in harsh touch sensation at 24 h postinjury in the wild-type worms while this recovery was absent in the mlk-1(0) (Fig. 3F). This indicated a strong correlation between the axon regeneration to the recovery of harsh touch sensation in the PVD neurons. To ensure that the recovery in harsh touch response is indeed due to rewiring event, not due to some other phenomenon, we tested the recovery profile upon the complete removal of PVD in the wild-type worms (Extended Data Fig. 3-1A,B). When PVD was ablated, the harsh touch response did not improve at 24 h postablation as compared to response at 8 h postablation (Extended Data Fig. 3-1A,B). These results suggest that the axon regeneration in PVD neuron that is mediated by DLK/MLK pathway enables functional recovery of harsh touch sensation.

Functional restoration during dendrite regeneration is correlated to successful events of arborization and fusion

Our previous study showed that the dendrites of PVD neurons can regenerate following a laser-assisted injury (Brar et al., 2022). However, it is unclear whether the relative change in the parameters involving proprioception would correlate with the events of successful dendrite regrowth. Dendrotomy of either one or both PVD neurons resulted in a drop in the amplitude and wavelength of the sinusoidal traces made by the moving worm (Fig. 4A,B). Qualitative inspection of the traces of dendrotomized worms showed a marked improvement at 24 h postdendrotomy with lesser irregularities in the sinusoidal movement as compared to the traces observed at 8 h postdendrotomy (Fig. 4B). All the parameters of proprioceptive behavior including amplitude, wavelength, and bending angle were significantly increased at 24 h postdendrotomy as compared to the values at 8 h (Fig. 4C–E).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Dendrite regeneration in PVD neurons leads to recovery in posture defect of worms. A, Schematics representing the one-side and both-side dendrotomy with both the major and minor dendrites were injured on one or both sides, respectively. B, The images of the worm trajectories on the bacterial lawn are shown at 8 and 24 h after injury in mock, one-sided dendrotomy (cut-side down where the injured PVD is touching the agar media) and both-side dendrotomy conditions. The scale bar represents 500 microns. C–E, Quantification of amplitude (C), wavelength (D), and bending angle (E) at 8 and 24 h after injury in mock, one-side dendrotomy (cut-side down, i.e., injured PVD neuron is touching the agar media) and two-side dendrotomy are plotted with the absolute values in a violin plots. The red dotted lines represent the median of population occurrence. 27 < n < 30, N = 3. The posture parameters were also correlated with the territory covered by regenerated dendrites and plotted in regression plot (Extended Data Fig. 4-1A–C). F, Representative confocal images along with the schematics 24 h after dendrotomy with the menorah fusion highlighted by a red faded box, the green arrowheads marking the reconnection phenomena, and the red arrows representing the site of injury. The regenerated dendrites are represented in green, the distal part in grey, and the axon is represented in red in the schematic. The yellow dotted lines represent the longest regenerating dendrite corresponding to the territory covered by regenerated dendrites. Three different cases are shown, i.e., menorah fusion, menorah–menorah fusion plus reconnection, and no fusion events. G–I, The quantification of postural parameters, i.e., amplitude (G), wavelength (H), and bending angle (I) are shown at 8 and 24 h after dendrotomy which are classified into three different cases which are menorah fusion, menorah fusion plus reconnection, and no fusion are plotted. 15 < n < 22, N = 3. The posture parameters were also quantified after the dendrotomy of wild type and dlk-1(0);mlk-1(0) worms (Extended Data Fig. 4-2A–D). The statistical analysis for C–E, G–I is Tukey's multiple comparison tests, with p < 0.05*, 0.01**, and 0.001***. The violin plots represent the median and population distribution. ns stands for not significant, N stands for the number of independent replicates, and n stands for the number of worms taken for analysis.

Dendrite regeneration in the PVD neurons is accompanied by regrowth from the severed end, reconnection between distal and proximal counterparts, and fusion of the higher-order branches (Brar et al., 2022). The higher-order branches often fuse with each other to bypass the disjointed primary dendrites, which is termed as menorah–menorah fusion (Kravtsov et al., 2017; Oren-Suissa et al., 2017; Brar et al., 2022). In these experiments, we noticed a similar branching with reconnection (green arrowheads) and menorah–menorah fusion (red faded box) events (Fig. 4F). The regrowing dendrites span the gap created by the injury. This territorial expansion of the regrowing dendrites is estimated by the length of the longest neurite measured from the cell body (Yellow dotted line, Fig. 4F). We divided the events into three categories, “menorah–menorah fusion,” “menorah–menorah fusion plus reconnection,” and “no fusion/reconnection” (Fig. 4F). Based on the amplitude, wavelength, and bending angle measurements of the sinusoidal movement, significant increase in wavelength parameter was correlated to the “menorah fusion” events (Fig. 4G–I). In cases with “menorah fusion plus reconnection,” both amplitude and wavelength values were significantly increased at 24 h as compared to 8 h postdendrotomy (Fig. 4G,H). On the other hand, we did not observe any functional recovery in cases of “no fusion/reconnection” events (Fig. 4G,H). Interestingly, the bending angle did not change during dendrite regeneration except in cases of “no fusion/reconnection,” which showed a decrease (Fig. 4I). It is possible that loss of dendrites upon injury led to decline in the muscle or neuronal strength. This type of positive correlation between functional recovery and the “fusion” events is also observed after axonal injury in the gentle touch neurons (Abay et al., 2017; Basu et al., 2017). Additionally, there was a positive correlation between “territory length” and body posture parameters involving “amplitude” and “wavelength” (Extended Data Fig. 4-1A–C).

The proprioception parameters were also checked for both sides “major” and “major plus minor” dendrotomy experiments in worms with different transgene labeling PVD neurons with mScarlet. The trajectories on the OP50 lawn showed a visible recovery 24 h postinjury (Extended Data Fig. 4-1D). The regeneration events including reconnection and menorah fusion were also seen in these worms at 24 h postinjury (Extended Data Fig. 4-1E). We found that the affected parameters, i.e., amplitude and bending angle in case of major dendrotomy and all three parameters in case of major plus minor dendrotomy experiments showed an increase in value at 24 h after injury (Extended Data Fig. 4-1F–H).

Interestingly, the recovery in posture parameters was also observed in the double mutant for dlk-1 and mlk-1 (Extended Data Fig. 4-2). Overall, our analysis indicated that the worms can regain their posture that is lost due to dendrotomy through the regeneration of dendrites. This process can occur independent of DLK-/MLK-1 pathway.