Neurotrophic control of size regulation during axolotl limb regeneration

The mechanisms that regulate the sizing of the regenerating limb in tetrapods such as the Mexican axolotl are unknown. Upon the completion of the developmental stages of regeneration, when the regenerative organ known as the blastema completes patterning and differentiation, the limb regenerate is proportionally small in size. It then undergoes a phase of regeneration that we have called the “tiny-limb” stage, that is defined by rapid growth until the regenerate reaches the proportionally appropriate size. In the current study we have characterized this growth and have found that signaling from the limb nerves is required for its maintenance. Using the regenerative assay known as the Accessory Limb Model, we have found that the size of the limb can be positively and negatively manipulated by nerve abundance. We have additionally developed a new regenerative assay called the Neural Modified-ALM (NM-ALM), which decouples the source of the nerve from the regenerating host environment. Using the NM-ALM we discovered that non-neural extrinsic factors from differently sized host animals do not play a prominent role in determining the size of the regenerating limb. We have also discovered that the regulation of limb size is not autonomously regulated by the limb nerves. Together, these observations show that the limb nerves provide essential and instructive cues to regulate the final size of the regenerating limb.

The ratio of limb to body length in regenerating and unamputated limbs was measured over time (10cm animals; n=10). We have separated the growth of the limb regenerate into three stages: the blastema stage (dark grey), the early tiny limb stage (medium grey), and the late tiny limb stage (light grey). Error bars = standard error of the mean. T-Test was used to evaluate significance between the regenerating and uninjured limb size at each time point. All data points not marked with N.S. had p-values less than 0.005. (C) Histogram showing the average amount of time in days that the regenerating limb is in each growth stage for animals of different sizes (4 cm, 10 cm. and 20 cm in length). each of these processes during the different phases of regeneration, relative to uninjured limbs, 147 to determine which could be contributing to growth of the tiny limb. Additionally, we speculated 148 that the contribution of these cell processes could vary in the different tissue types in the 149 regenerating limb. Rather than quantifying the above-described processes globally, we analyzed 150 the epidermis, soft tissue (including all tissues except for skeleton and epidermis), and skeletal 151 tissue (bone and cartilage) separately.

152
Regenerating limbs at the different stages of growth were sectioned transversely mid-153 zeugopod, and cell proliferation and cell death were each analyzed by either EdU incorporation 154 or TUNEL staining, respectively. We observed significantly more cell proliferation and significantly 155 less cell death in all tissues analyzed in the early and late tiny limb staged regenerates compared 156 to the unamputated and fully regenerated limbs (Figure 2A and B). Interestingly, the fold increase largest increase in cell proliferation was observed in the epidermis (3.7-fold increase, Figure 2A), 159 while soft tissue and skeletal tissue had slightly more modest increases (3.2 and 2.4-fold 160 increases respectively, Figure 2A). The largest decrease in apoptosis was seen in the skeletal 161 tissue (20.8-fold decrease, Figure 2B), while the soft tissues and epidermis exhibited more 162 moderate decreases of 3.0 and 3.3-fold, respectively ( Figure 2B).   Figure 2A). We observed that cell size was significantly smaller in the regenerating tissue than 169 the uninjured tissue and increased as regeneration progressed ( Figure 2C). This was most 170 profound in the muscle (4.5-fold smaller) and least in the epidermis (1.4-fold smaller, Figure 2C).

171
The ECM area was calculated indirectly by subtracting the total cellular area from the tissue area 172 and dividing by the tissue area (Supplemental Figure 2B) (more detail in materials and methods).

173
However, we did not observe any significant differences in the extracellular compartment of limbs,

174
indicating that ECM deposition does not play a significant role in growth of the tiny limb ( Figure   175 2D).

176
Together, this data indicates that a combination of increased cell proliferation, decreased 177 apoptosis, and increased cell size contributes to the growth of the tiny limb staged regenerate.

178
Additionally, while all tissue types showed the same trends in all of the cell processes that we 179 analyzed, our data suggests that different cell processes contribute more or less to growth in 180 different tissue types. Future studies will be required to resolve these tissue-specific contributions 181 to growth in more detail.

183
Growth of the Tiny Limb is dependent on limb nerves 184 One interesting observation from the above-described characterization is that the 185 abundance of cell proliferation, cell death, and cell size all show similar trends regardless of the 186 tissue type assessed during each stage of growth in the regenerate. This suggests that there 187 could be a singular signal that coordinates these processes such that the highest growth-188 promoting signal is occurring during the early tiny limb stage when growth is most abundant and decreases as the growth rate slows during the late tiny limb stage. Thus, we next sought to 190 determine the source of the signal that regulates cell proliferation, death, and size during 191 regeneration.
regenerating limbs. The role and mechanism of neurotropic regulation during the early (blastema) 194 stages of regeneration has been widely studied (Farkas, Satoh, 2014;Singer, 1946Singer, , 1952Singer, , 1978Singer & Inoue, 1964). It has been well established that 197 nerve signaling is required for, and is a key driver of, blastemal cell proliferation (Brockes, 1984; miniaturized limbs through repeated removal of limb buds results in "mini limbs" that are hypo-

202
To test this idea, we first characterized the abundance of innervation during these post-203 developmental stages of regeneration. We collected early and late tiny limbs, as well as 204 unamputated and fully regenerated limbs for comparison, and sectioned them transversally 205 through the zeugopod. The sections were stained with an anti-acetylated tubulin antibody (nerve 206 Figure 3: The tiny limb staged regenerate is hyperinnervated. A) Fluorescent images were obtained of transverse sections of uninjured, early and late tiny limb stages, and fully regenerated limbs (DAPI = blue, Phalloidin (for actin filaments) = red, Acetylated-tubulin (for nerves) = green; scale bars are 1000um). B) Nerve area relative to total limb area was quantified from the sections represented in A (n=5). Error bars = SEM. P-values calculated by ANOVA and the Tukey Post-hoc test. *=p<0.05 **=p<0.005.
( Figure 3B). This quantification revealed significantly higher levels of relative innervation during 208 the early and late tiny limb stages compared to unamputated and fully regenerated limbs ( Figure   209 3B). Interestingly, the total abundance (not normalized to tissue area) of innervation increases 210 from the early to late tiny limb stages (Supplemental Figure 4). Thus, the decrease in relative 211 innervation during the transition from the early to late tiny limb stages is likely due to the 212 substantial increase in limb area of the later staged regenerate. The relative abundance of 213 innervation also correlates well with the growth rate in these tissues. The early tiny limb stage has 214 the highest relative abundance of innervation, followed by the late tiny limb stage ( Figure 3B).

215
The abundance of innervation of the completed regenerate has decreased to that of the uninjured 216 limb ( Figure 3B). We speculated that nerves could provide growth-promoting signals during 217 regeneration, which decreases as the relative abundance of innervation decreases, slowing the 218 growth of the regenerate as it reaches its final size.

219
To determine whether nerve signaling plays a functional role in determining size, we next 220 tested whether nerve signaling is required to maintain growth in the tiny limb staged regenerate.

221
Limbs were amputated and permitted to regenerate to the early tiny limb stage, at which point 222 nerve signaling was severed (blue line, Figure 4A) via denervation at the brachial plexus. To test 223 for a possible dose response or signaling threshold effect, we severed either one, two, or all three

228
We observed that nerve signaling is required for growth of the tiny limb. Fully denervated 229 early tiny limbs had a 9-fold slower growth rate than innervated early tiny limbs ( Figure 4B).

230
Likewise, cell proliferation was negatively impacted by denervation in all tissue types analyzed Lastly, cell size appears to only be significantly affected in the muscle, where there is a near 2-235 fold decrease in average cell size in the denervated early tiny limbs ( Figure 4E). When late tiny 236 limbs were fully denervated, only the growth rate and abundance of cell proliferation were 237 significantly decreased (Supplemental Figure 5). Thus, neurotrophic regulation of cell death and 238 cell size (in the muscle tissue) appears to be restricted to the early tiny limb stage of growth. on the tissue and growth characteristic quantified. A dose response is reflected by a linear 241 relationship between abundance of signal and the phenotype. A threshold response indicates a 242 specific abundance of a signal is required for a phenotype, for example, a specific level of 243 innervation is required for growth. We observed that cell proliferation in the soft tissue and 244 skeleton decreased significantly, to full denervation levels, with partial denervations indicating that 245 there is a high threshold of nerve signaling responsible for maintaining cell proliferation in these 246 tissues ( Figure 4C). Conversely, cell death in the epidermis had a strong dose response, with 247 significant incremental increases with increased denervation ( Figure 4D). These results indicate 248 that each tissue responds differently to nerve signaling to maintain growth. This could explain the 249 decreasing growth rate trend in partial denervations, which only becomes significant with the full 250 denervation ( Figure 4B). Together these results reveal the complexity of neuronal regulation of 251 growth and indicate that an evaluation of the tissue-specific responses to nerve signaling is 252 required for a complete understanding of growth and size regulation during regeneration.

254
Innervation abundance determines regenerate length 255 Having established that nerve signaling is required to maintain growth during limb 256 regeneration, we next wanted to determine whether we could positively and negatively manipulate 257 the size of the regenerate by altering the abundance of innervation. To test this, we performed a 258 grafting experiment between differently sized axolotl ( Figure 5A). Limb nerve bundles increase in 259 size as the animal grows, and quantification of the cross-sectional area of the nerve bundles 260 extending out from the limb Dorsal Root Ganglia (DRGs) reveals that the size of the bundle is 261 almost 2-fold larger in 14cm long animals compared to 7cm animals (Supplemental Figure 6).

262
Thus, blastema grafts onto the limbs of large hosts will be innervated by larger nerve bundles than grafts on small host animals.

287
We hypothesized that if nerve abundance can regulate size of the limb regenerate, we 288 would observe that the lengths of the grafted regenerates will correspond to the size of the host 289 environment rather than the donor. Thus, blastemas from the small donors will produce large 290 ectopic limbs when grafted to large hosts, and blastemas from large donors will produce small 291 ectopic limbs when grafted to small hosts. Alternately, if nerve abundance does not influence 292 regenerate size, then we would expect to see the blastemas from small donors on large hosts 293 produce ectopic limbs smaller than the control grafts from large animals, and vice versa.

294
We observed that nerve abundance influenced ectopic limb size. The blastemas (15 days 295 post amputation) from small donors were initially 2-fold smaller than those from large donors when 296 they were grafted onto the host environments ( Figure 5C, left panels, 0 days post graft (DPG)).
Interestingly, we observed that the grafted tissues were the same size by 18 DPG within each host type (Supplemental Figure 7A). By approximately 38 DPG, there was a significant difference 299 in size between the ectopic limbs on the large and small host animals, and this continued until 300 130 DPG, when the growth rate of the grafted limbs matched that of the host limbs ( Figure 5C, 301 right panels; Supplemental Figure 7A). In comparison, the control regenerates on the donor 302 animals remained significantly different in size throughout regeneration ( Figure 5D). These data 303 indicate that the sizing of the limb regenerate positively correlates with nerve abundance.

304
However, it does not rule out other potential influences from the host environments that may also 305 contribute to size. Thus, we next designed an experiment that decouples nerve abundance from 306 the size of the host to test whether non-neural signals contribute to the sizing of the limb 307 regenerate.

309
Non-neural extrinsic signals do not provide instructive cues on regenerate size.

310
To evaluate the potential role of non-neural sources of size regulation that may be present

336
To test whether non-neural extrinsic signals provide instructive cues that regulate the size 337 of the regenerating limb we performed the NM-ALM on differently sized host animals ( Figure 7A).

338
If non-neural extrinsic signals play a role in regulating the size of the regenerate, then we expected 339 to observe that the regenerates that grew in the NM-ALMs on the large (14 cm) hosts would be 340 larger in size than the ones that grew on the small (7 cm) hosts ( Figure 7B). Alternately, if 341 signaling from the nerves play the major instructive role then we would not expect to see

372
One of the last steps of regeneration required to generate a fully functional limb is the 373 growth of the regenerate to a size that is proportionally appropriate to the animal. However, little and small (~7cm, n=16) GFP+ animals were grafted into wound sites on host animals(~14cm) followed by mid-bud blastemas from (~7cm) donor animals. B) If the ability to regulate size is autonomous to the DRGs, the DRGs from large animals will produce larger ectopic limbs than those from small animals. If size regulation is not autonomous, there will be no difference in ectopic limb size between grafts supplied by large or small animal DRGs. C) The ectopic limb lengths were the same size at 0DPG and remained the same throughout regeneration (135DPG). Error bars=SEM. P-values calculated by T-tests. data had previously been collected on the post-blastema stages of regeneration and how growth 375 and size are regulated. This study constitutes the first thorough investigation of how the 376 regenerating limb grows to the proportionally appropriate size. We have discovered that there are 377 three distinct phases of growth prior to the limb completing regeneration; the blastema phase, the 378 early tiny limb phase, and the late tiny limb phase (Figure 1). During the tiny limb phases of growth, 379 the regenerate grows through increased cell proliferation, survival and cell size (Figure 2), and 380 innervation is required to maintain this growth (Figure 3 and 4). Our data indicates that nerves 381 play the instructive role in size regulation ( Figure 5 and 7), and that factors from the native neural        This study provides foundational knowledge on the later stages of limb regeneration to 497 understand the how size and proportionality becomes reestablished in a continually growing 498 system. Our data indicates that nerve signaling plays an instructive role in determining regenerate 499 size, and future studies will focus on identifying the molecular mechanism of size regulation.

500
Furthermore, our data suggests that there is an upstream driver of size regulation, potentially in 501 the DRG's endogenous environment or from the CNS. Lastly, as previously stated, our studies 502 have not ruled out the likely intrinsic factors involved in size regulation. In total, there are still many how size and proportionality become reestablished during axolotl limb regeneration. Furthermore, 505 as regenerative medicine seeks to tap back into the developmental mechanism in order to regrow 506 a fully functional limb, it will be important to study size regulation in multiple species to identify the 507 shared mechanisms regulating this process.

518
Mexican axolotls (Ambystoma mexicanum). Animal sizes are measured snout to tail tip and 519 described in the text for each experiment. They were housed in 40% Holtfreters on a 14/10-hour 520 light/dark cycle and fed ad libitum. Animals were fed every day or three times a week depending 521 on the size of the animal. Animals were anesthetized in 0.1% MS222 prior to surgery or imaging.

522
Live images were obtained using a Zeiss Discovery V8 Stereomicroscope with an Axiocam 503 523 color camera and Zen software (Zeiss, Oberkochen, Germany).

525
To generate large and small animals, larval animals were either housed at 19 o C or 4 o C, which 526 slows their growth rate. Animals were grown at these temperatures until their body lengths were 527 approximately two-fold different, at which point the smaller animals were moved to 19 o C for two 528 weeks prior to any surgical manipulation.
were recorded prior to experimentation and weekly following surgical manipulation. After 5 weeks 536 measurements were biweekly, and after 10 weeks they were taken triweekly.

648
To analyze ECM size for the epidermis and muscle, the area of all WGA internal cell spaces (with 649 and without DAPI) were quantified. For skeletal tissue, the area of the Alcian blue negative cell 650 spaces were quantified. The sum of these areas was calculated to obtain the "total cellular area".

651
The area of the complete tissue "total tissue area" was then quantified (Supplemental Figure 3B).
Since the tissue sizes can vary, percent ECM was determined through the following equation: