Figure 1. 24-Hr Combined Progesterone-Device Treatment Changes Early Cellular Response, Decreasing the Early Leukocyte Invasion and
Leading to a Scar-free Wound Healing with Higher Nerve Supply at Later Stages
(A) Schematic showing the experimental design and integrative methodology for assessment of the regenerative potential. The orange dashed line indicates
amputation plane. The black dashed lines delineate gross morphological outcomes for the two experimental groups.
(B–E) Expression of progesterone receptor (PR) on frog limbs revealed by fluorescent ligand binding assays and hormone displacement studies.
(B) Low-magnification image of a cross-section of intact frog hindlimb after incubation with the fluorescent ligand for PR (progesterone 3-(O-carboxymethyl)
oxime:BSA-fluorescein isothiocyanate conjugate [PFITC]) and counterstained with DAPI. In the center of the bone is the marrow containing PR+ mesenchymal
cells (white arrowheads).
(C) High-magnification image through the bone marrow of an intact frog hindlimb showing nuclear colocation for PFITC and DAPI labeling (white arrowheads).
(D) Parallel section to (B) after PFITC co-incubation with a saturating concentration of the physiological agonist progesterone. No positive cells were detected.
(E) Cross-section of frog hindlimb after PFITC co-incubation with an unspecific hormone, hydrocortisone. Since no competitive binding occurs, the fluorescent
ligand reaction showed a similar positive pattern to the single PFITC incubation, as detected when compared to (B). Scale bar, 100 mm.
(F–H) H&E staining at 0.5 months post-amputation (mpa), showing the overall organization of the soft tissue after no subsequent treatment (Ctrl group, F) and after
treatment with the silk-hydrogel device loaded with progesterone (Prog-device group, G). Two weeks after amputation, differences in the fibroblastema region
(asterisk) were present between the two groups—mostly collagen accumulation in untreated animals—and frequent areas of irregular cells (asterisk) in
Prog-device treated animals. In (H), a scheme with the main histological elements, as seen on sections along the longitudinal axis, is presented.
(I–K) Immune infiltration (after XL2 immunofluorescence) in the early fibroblastema of Ctrl (I) and Prog-device (J) animals. The number of leukocytes was
significantly lower (K) in the apical region (or distal from the amputation plane, blue dashed line) of the blastema in the treated animals.
(L–N) The number of H3P-expressing proliferative cells (white arrows) normalized to the total number of cells (DAPI labeled, blue) between Ctrl (L) and Prog-device
(M) animals at late-blastema stage (3 mpa) showed no significant differences (N).
(O–Q) The nerve patterning, as detected after tubulin (Tub) immunofluorescence on longitudinal sections, was less organized in Ctrl (O) group than after Progdevice
treatment (P). The former was characterized by the presence of specialized groups or bundles of nerve fibers, with higher area per positive unit (Q).
(I, J, L, M, O, and P) Scale bars, 100 mm. (I and J) Amputation plane is right, medial is up. (L and M) Amputation plane is up, medial is left. (O and P) Amputation
plane is left, lateral is up. (F–Q) Micrographs show longitudinal sections (see Figure S5A for scheme). (K, N, and Q) Values are represented with scatterplots, in
which each dot represents one histological section and each dot style represents one animal (with at least three histological sections). Statistical analysis was
performed on the pooled individual sections. Horizontal line indicates mean. p values after t test (equal variances; N) or Mann-Whitney test (unequal variances;
K and Q) are indicated as *p <0.05, **p <0.01, ns, p > 0.05.
Figure 2. Gross Morphological Outcomes and Bone Reorganization Show Tendency to Pattern Formation in Animals Treated with the
Combined Drug Device
(A and B) Soft-tissue patterning for Ctrl (top) and Prog-device treated (bottom) animals during a 9.5-month regeneration period.
(A) Amputation plane is indicated with an orange line. Lateral panels show the appearance of the regenerate at 2.5 mpa, when differences between groups
become obvious. White arrows point out the typical unpigmented area covering wider regenerates. Scale bar, 1 cm.
(B) Data represent the mean and SD (Ctrl group n > 6 animals for 0.5–2.5 mpa, n = 3 for 5–9.5 mpa; Prog-device group n > 6 animals for 0.5–9.5 mpa). p values after
two-way ANOVA are indicated as *p <0.05, **p <0.01, ns, no significant difference.
(C–E) Reorganization of the regenerated bone and remaining tissues after treatment and comparisons to the naturally occurring joint formation between femur
and tibiofibula bones in intact limbs.
(C) Left: X-ray image of right leg of intact or uncut animals. Typical deviation angle (a, close to 90) for joint formation between femur and tibiofibula bones is
indicated. An orange dashed line indicates the plane where amputation is performed. Right: X-ray images for Ctrl and Prog-device animals at 7.5 mpa, to evaluate
the reintegration of the regenerated bone and remaining tissues (old bone and soft tissue) during limb regeneration (for schematic representation and axis
explanation, see Figure S2B).
(D) Deviation angle created by the new bone growth for Prog-device animals is significantly more open than the one for regular spikes in Ctrl or no-device animals.
(E) Graph representing the old-bone displacement (d) during the course of the regeneration, from the longitudinal axis (0) to the medial edge of the soft tissue
(0.5). Values are represented with scatterplots, in which each dot represents one animal. p values after t test (D) or post hoc Bonferroni’s test (E, p <0.01 for oneway
ANOVA) are indicated as *p <0.05, **p <0.01, ns, no significant difference.
Figure 3. Instead of the Typical Spikes Formed
in the Absence of Treatment, Combined ProgDevice
Treated Animals Regenerate Complex
Patterned Paddle-like Structures
(A–D) Anatomical outcome (left) and X-ray images
(right) of regenerates formed in adult Xenopus hindlimb
amputation after no treatment (Ctrl, A) and after 24-hr
combined treatment of drug-loaded device (Progdevice,
(A) Red arrow indicates the hypomorphic cartilaginous
structure, spike, lacking specialized tissues.
(B–D) Green arrows on soft tissue images (left panels)
indicate the dense and sprouted vascularization, easily
visualized due to unpigmented epithelium covering the
regenerates of treated PD7 (B), PD5 (C), and PD1 (D)
animals. X-ray images (right panels) show the new
bone growth within the regenerates of treated PD7 (B),
PD5 (C), and PD1 (D) animals (weakly ossified bone
close to the amputation plane, green arrows; nonossified
bone or regenerated cartilage traveling toward
the distal part, turquoise arrows in B–D; magenta arrow
in A indicates the absence of bone regeneration in the
distal spike). Individual identification of each animal is
indicated in white (Ctrl: untreated control, PD: Progdevice
treated, followed by the individual number).
(E) Shape profiles, as obtained after MorphoJ software
measurements (see Figure S2B for details), for untreated
(red) and Prog-device treated animals (green).
Each circle indicates the average position for each
landmark, after computations for n = 3 and n = 7 animals
in Ctrl and Prog-device groups, respectively.
Quantification and statistical analysis of these shape
states showed significant differences in the Mahalanobis
distance (p <0.05) for the profile of the shape of
(F) Application of RI to regenerates at both 2.5 and 9.5
mpa (two-way ANOVA, p <0.01 for ‘‘group’’ factor).
Scatterplots are presented in which each dot represents
(G and H) Quantification of regenerated limb length
(a0 + b0
) relative to uncut limbs (a + b) revealed that the
regenerating hindlimb stops growing at a point at
which, under normal growth conditions, the animal’s
digits appear (yellow arrows) (G). Values in (H) are
represented with scatterplot, in which each dot represents
Figure 4. Prog-Device Treated Animals
Show Significantly Greater Innervation and
Vascularization than Untreated Animals at
the Most Distal Part of the 9.5-mpa Regenerate,
with Patterns Closer to Uncut (Intact)
(A–D) Low-magnification immunofluorescence
images of DAPI-counterstained cross-sections
revealing innervation (anti-acetylated a-tubulin
antibody [Tub]) and vascularization (anti-smoothmuscle
actin antibody [SMA]) of amputated untreated
(Ctrl) and amputated plus Prog-device
(Prog-device) treated animals. The overall area
occupied by Tub+ fibers is significantly greater in
treated tips (turquoise double-headed arrow in B,
compared to magenta double-headed arrow in A).
Similarly, the density and extension of the blood
vessels for Prog-device treated animals (D, turquoise
arrows) is higher when compared to untreated
animals (C, magenta arrows). gl, SMA+
epidermal glands were not included in the analysis.
Scale bars, 250 mm. The cartilage core of the
9.5-mpa regenerates is indicated in (A). For a
schematic representation of the plane cut, see
(E and F) Quantitative results of number of positive
units per square millimeter and area per unit for Tub
(E) and SMA (F) immunofluorescence in Ctrl (red)
and Prog-device groups (green). Values are represented
with scatterplots, in which each dot represents
one histological section, and each dot style
represents one animal (with at least three histological
sections). Statistical analysis was performed
on the pooled individual sections. Horizontal line
indicates mean. p values after Mann-Whitney test
(unequal variances) or t test (equal variances; only
for E, left) are indicated as **p <0.01.
(G–N) High-magnification images revealed that
while in untreated animals, the typical nerve
patterning is composed of individual and unpatterned
nerve fibers (magenta arrows in G), Tub+
axons in Prog-device animals show a tendency to
group organization, forming bundles (turquoise
line-encircled areas in H), similar to the nerve organization
in intact or uncut limbs (turquoise lineencircled
areas in I and J). Both the number and
morphology and area of the SMA+ vessels for
Prog-device tips were clearly closer to the intact
limb (turquoise arrow and asterisks in L, M, and N)
than to the Ctrl untreated limbs (small vessels
indicated by magenta arrows in K). See Table S4
for quantitative data. Scale bars, 100 mm.
Figure 5. Locomotor Activity in Prog-Device
Treated Animals Resembles More Closely
that of Uncut Animals
(A) Quantifications of efficient activity levels,
measured by counting the number of active movements
(with trajectory displacement). Values are
represented as mean ± SD (n = 4 animals per group).
One-way ANOVA p <0.01. p values after Bonferroni’s
post hoc analysis are indicated as *p <0.05, **p <
0.01, ns, p > 0.05.
(B) Surfacing frequency or number of visits to the
water-air surface performed by each experimental
group per time unit (minute). Values are represented
as mean ± SD (n = 4 animals per group). One-way
ANOVA p <0.01. p values after Bonferroni’s
post hoc analysis are indicated as *p <0.05, **p <
0.01, ns, p > 0.05.
(C) Graphic representation of the percentage of time
spent in each quadrant of the tank per experimental
group. Unbiased exploration (black line) is shown as
reference (25% of time in each quadrant). Analysis of
the time distributions within each experimental group
showed clear significant differences (p <0.01 for
(0.05,9)). Data represent the pooled distribution of
animals per group (see Figure S6A for contingency
(D) Mean speed (expressed in centimeters per second)
at which each animal (represented by individual
dots) swims around the tank. Orange squares, uncut
or intact animals; green circles, amputated animals
followed by Prog-device combined treatment; red
triangles, amputated animals without additional
Figure 6. Subnetwork Enrichment Analysis of All Three (Ctrl, Sham, and Prog-Device) Blastema Datasets Showing Quantitatively and
Qualitatively Differences for Regulated Transcripts and Cell Processes after Combined Prog-Device Treatment
(A and B) Venn diagram comparing genes (A) and subnetworks (B) regulated in blastema for both Sham and Prog-device interventions.
(C) Pie chart of the functional classification of the pathways exclusively regulated after Sham intervention (top, amputation followed by only hydrogel device) or
after the combined Prog-device treatment (bottom, amputation followed by hydrogel device + drug).
Figure 7. Potential Targets (Gene Networks and Cell Processes) of the Combined Treatment (Device + Drug)
(A) Unweighted principal coordinate analysis (PCoA) demonstrated a clear separation of red, yellow, and green clusters representing Ctrl (or ‘‘no device’’), Sham
(or ‘‘only device’’) and Prog-device (or ‘‘device + drug’’), respectively
Supplementary Fig. S1. Early response to device attachment: Progesterone levels and proliferation &
innervation patterns, Related to Figure 1.
A-C. Progesterone levels in the three experimental groups 24 hours after amputation and device
attachment, measured via ELIA both at remote tissues (A&B, blood and brain) and at injury site (C, blastema).
Control (untreated after amputation), Sham (only-device treated after amputation) and Prog-device
(combined progesterone-loaded device treated after amputation) animals are represented in red, orange and
green, respectively. Values are represented with scatter plots, where each dot represents the average value of
three biological replicates (n=6 animals per replicate and experimental group). Horizontal lines indicate
mean ± sd. P values after Bonferroni’s post-hoc test (One-way ANOVA P >0.05 for A&B, and P <0.01 for C) are
indicated as ** P <0.01, ns: no significant difference. D-F. Cell proliferation (after H3P immunofluorescence)
in the early fibroblastema (0.5 mpa) of untreated Control (D) and treated Prog-device (E) animals. At this
stage, the proliferative response after amputation is weakly starting and no significant differences in the
number of H3P-positive cells were detected between groups. In D, the bottom-right insert corresponds to the
dashed-white line. Amputation plane is indicated with an orange-dashed line. G-I. The presence of organized
blood vessels, following the longitudinal axis, was a landmark for Prog-device late blastemas (3 mpa; white
asterisk in H compared to G), although not significant differences were obtained after group comparisons. Scale bar = 100 μm. G,H: Amputation plane is left, lateral is up. F, I: Values are represented with scatter plots,
where each dot represents one histological section, and each dot style represents one animal. Horizontal line
indicates mean. P values after t-test are indicated as ns: no significant difference.
Supplementary Fig. S2. Assessment of the regenerated soft tissue, Related to STAR Methods and Figure 2.
A. Diagram of a prototypical 2-mpa regenerate showing the biological meaningful elements and
morphometric parameters used to evaluate the limb outcome over time: percentage of width change (W2
respect to W1; WID), percentage of unpigmented area (respect to the total regenerated area; UNA), and
maximal length of the regenerate (LEN). WID: Before amputation, the limb has a constant width. This
situation is not maintained after amputation and tissues start to narrow while outgrowth is progressing. A
lower percentage of width change is, hence, an indicative of regenerative ability, as regenerate is closer to the
original morphology, previous amputation. We evaluate the decrease in width experienced by the limb as
consequence of the amputation by means of the percentage of change between two width values, before (W1,
width at amputation plane, orange-dashed line) and after (W2, width at the base of the regenerate, bluedashed
line) amputation, with the formula WID = (W2-W1)*100/W1. UNA: This variable quantifies the
differences in pigmentation pattern of the regenerate, by means of the percentage of total regenerated area
that is covered by unpigmented epidermis. To this, firstly, we calculated the total area regenerated (from the
amputation plane to the tip of regenerate; green and pink surfaces in diagram, respectively). Secondly, a
straight line indicating the demarcation of the totally unpigmented portion from the rest of the limb was
drawn. Then, the area of the regenerate under that line (unpigmented area) was divided by the total area,
with the formula UNA = pigmented area/ (pigmented+ unpigmented area)*100. LEN: We evaluate the
maximal length of the regenerate by calculating the distance from the amputation plane to the plane set by
the tip of the regenerate (distance between the two orange-dashed lines). To avoid size noise, this value was
normalized to the total animal length (or distance from snout to vent, TOT LEN). This morphometric analysis
was performed on regenerates belonging each experimental group (Control and Prog-device, respectively) at
five selected times for a 9.5-month period: 0.5, 1, 2.5, 5, 7.5 and 9.5 months post amputation (mpa).
Representative images and graphs are indicated in Fig. 2A, B. B. Diagram of a prototypical 7.5-mpa regenerate
including the positions of the twelve landmarks used for the geometric morphometrics analysis. To describe
and quantify the changes in shape of the regenerate between Control and Prog-device groups, we employed
MorphoJ software (Klingenberg, 2011), as extensively detailed in (Mondia et al., 2011). MorphoJ performs a
geometric morphometric analysis based on landmarks or points used to define the profile of a shape. For our
analysis, twelve landmarks defined each regenerate profile: the first six landmarks were biologically
meaningful positions: 1 and 2 define the amputation plane, 3 and 4 for the base of the regenerate, 5 and 6
define the points at the tip of the regenerate. The other six points were semi-landmarks, chosen by
successively finding the midpoints between the previous positions, as indicated by the numbers in the
diagram. Landmarks were placed on digital images using ImageJ. MorphoJ is freely available from
http://www.flywings.org.uk/MorphoJ_page.htm. A reference axis, passing through the middle point of a
straight line linking landmarks 1 and 2, was placed on each profile. This axis was used to evaluate the
curvature at the tip of the regenerate. This parameter was included as criterion for the regeneration Index.
Curvature was considered positive when the four most distal landmarks (9, 12, 5 and 6) were situated on the
same lateral plane (right or left) respect to the middle axis. In this example diagram, 9 and 5 are on left plane
(blue square), as 12 and 6 are on the right one (orange square). The tip morphology of the illustrative
diagram would be considered negative for the variable curvature.
Supplementary Fig. S3. Assessment of the bone growth and its reintegration with the remaining
tissues, Related to STAR Methods and Figure 2C.
A. Diagram of a prototypical 2-mpa regenerate, under X-ray image, showing the biological meaningful
elements and morphometric parameters used to evaluate the skeletal outcome over time: maximal length of
bone from the amputation plane (BLEN; Table 1) and maximal area occupied by new bone growth (mostly
non-ossified bone or regenerated cartilage, BAREA; Table 1). Responsive animals generated the same bone
growth pattern as displayed in the graphic. The skeletal regrowth origin began above the plane of
amputation, indicating the occurrence of a secondary or spontaneous amputation (as described in mice by
Muneoka’s group (Fernando et al., 2011)). The secondary amputation acts as a catalyst for bone regeneration
as it creates more bone resorption and degradation. The newly formed bone typically widened at its distal
end and deviated from the midline of the intact tibiofibula. The morphometric analysis was performed on
regenerates belonging each experimental group (Control and Prog-device, respectively) at five selected times
for a 9.5-month period: 0.5, 1, 2.5, 5, 7.5 and 9.5 months post amputation (mpa). Representative images and
graphs are indicated in Fig. 3 and Supplementary Fig. S4. B. Diagram of a prototypical 7.5-mpa regenerate,
under X-ray image, including the axis used to evaluate the reintegration of the regenerated bone and
remaining tissues during limb regeneration. Left, Reintegration of the regenerated new bone and the
remaining old bone. We evaluated the deviation angle (α) formed under the intersection between the middle
line of the new bone growth and the middle line of the remaining bone (longitudinal axis of the tibiofibula).
Right, The geometric intersection between the middle line of the intact tibiofibula bone and the plane set by
the base of the regenerate was taken for evaluating the old-bone displacement (d) over the course of the
regeneration, from the longitudinal axis to the medial edge of the soft tissue. This morphometric analysis was
performed on regenerates belonging each experimental group (Control and Prog-device, respectively).
Representative images and graphs are indicated in Fig. 2C-2E.
Supplementary Fig. S4. Bone patterning as seen under X-ray images over a 9.5-month regenerative
period, Related to Figure 2.
A. The morphometric analysis was performed on regenerates belonging each experimental group,
Control (top) and Prog-device (bottom), at five selected times for a 9.5-month period: 0.5, 1, 2.5, 5, 7.5 and 9.5
months post amputation (mpa). The lateral brown arrows indicate amputation plane. As soon as 2.5 mpa,
new bone growth above the amputation plane (red asterisk) is detected in treated animals, suggesting that
the necessary catalyst for bone regeneration (as it creates more bone resorption and degradation) is
enhanced by treatment. B, C. Overall, treated animals show a tendency for bigger area (B) and longer (C) new
bone growth. As might be expected with the variability of genetic background, individual animals are clearly
observed to be strong treatment ‘Responders’, while other treated animals behaved as ‘Non-responders’,
similar to individuals in Control group. See Supplementary Fig. S2 for details and meaning of the
measurements. Scale bar = 0.5 cm. Values are represented with scatter plots, where each dot represents one
Supplementary Fig. S5. Experimental design and section axis for histological processing of the
regenerates (A) and the uncut or intact frog hindlimb B), Related to Figure 1 & Figure 4
A. Top, Drawing represent a short-term regenerate, indicating plane for tissue harvesting (orangedashed
line indicates the original amputation site) and longitudinal sectioning, following the ventral-to-dorsal
axis. Bottom, The most distal portion of the 9.5-mpa regenerates (or tips) were recut approximately 2 cm
below the original amputation plane (dashed-blue line indicates plane for recutting and orange-dashed line
indicates the original amputation plane). Tips were sectioned along the transversal plane, obtaining cross
sections (dashed-blue line) for histological analysis. B. Tips of 9.5-mpa regenerates were morphologically
compared to the uncut or intact limb. Two different levels of the uncut limb were separately cross-sectioned,
immunostained and analyzed: proximal or distal respect to the regular level for the amputation plane
Supplementary Fig. S6. Behavioral observations of exploration and positional information, Related to
Figure 5. Contingency analysis for the mean percentage of time spent in each quadrant per experimental
group. P <0.01 for X2 (0.05, 9)=109.1.
Supplementary Fig. S7. Common down-regulated cell processes and gene networks for both Sham and
Prog-device treatments. Related to Figures 6 & 7. Subnetwork enrichment analysis of blastema exposed to
only hydrogel device (Sham group) or combined hydrogel plus drug (Prog-Device group) identified common
down-regulated pathways involved in (A) changes in membrane potential, (B) dopaminergic system and (C)
muscle physiology. Complete data are presented in Appendix 1. All genes within a pathway are located in
Appendix 2. Green = down gene, Red = up gene.