Angewandte
Communications
Chemie
specificity, FNP1 activation was tested against FAPa in the
19.2 and 2.23-fold higher than that in HaCaT and NDF,
respectively (Figure 2c). Moreover, NDF cells were stimu-
[
19]
presence of its inhibitor Val-boroPro (talabostat) or other
enzymes relevant to skin diseases including dipeptidyl
peptidase IV (DPP IV), matrix metalloproteinase (MMP)-1,
MMP-2, MMP-13, caspase-3, and tissue plasminogen activa-
tor (tPA). As shown in Figure 1 f and the Supporting
Information, Figure S4, the fluorescence intensity of FNP1
barely increased when FAPa was treated with inhibitor
talabostat or in the presence of other enzymes, further
validating its high selectivity towards FAPa.
With its high sensitivity and fast kinetics, FNP1 was then
applied to detect KF cells in culture along with the control
skin cells including immortalized keratinocytes (HaCaT,
epidermis origin) and normal dermal fibroblasts (NDF,
dermis origin). After a short incubation period (1 h), strong
NIR fluorescence was detected for KF (Figure 2a). Co-
staining studies confirmed that FNP1 was mainly localized in
the cytoplasm, including the cell lysosome. In contrast, weak
fluorescence was observed in other cells including HaCaTand
NDF (Figure 2b). Fluorescence quantification further
revealed that the NIR fluorescence of FNP1 in KF cells was
lated using transforming growth factor (TGF)-b1, which is
[
20]
well-known to increase the expression levels of FAPa. As
shown in Figures 2b,c, the NIR fluorescence of FNP1 in NDF
cells was enhanced by 4.15-fold after TGF-b1 stimulation,
confirming that the NIR signal of FNP1 was indeed correlated
with the expression level of FAPa. To further validate this
correlation, reverse-transcription–quantitative polymerase
chain reaction (RT-qPCR) was conducted to quantify the
gene expression of FAPa for all these cells (Figure 2d). The
expression values of HaCaT, KF, NDF + TGF-b1 were 0.03,
4.32, and 23.2-fold relative to NDF expression, respectively.
Such an expression trend obtained from gene expression
analysis was consistent with that for the fluorescent signals.
Thus, these data show that FNP1 could be specifically
activated by FAPa, allowing for distinguishing KF cells
from other normal skin cells (that is, NDF, HaCaT).
Ability of FNP1 to detect KF cells was subsequently
evaluated in live, metabolically active human skin tissue
models containing diseased KF cells as a proof-of-concept. To
successfully detect the implanted KF cells, FNP1 was mixed
with Aquaphor ointment to form an emulsion to help it cross
the uppermost skin epidermal barrier to interact with dermis-
residing KF cells for topical application. Initial trials using
skin stripped of the epidermis (uppermost skin layer), showed
that the probe readily diffused throughout the skin dermis,
and detected KF cells within tissue at the depth of 1.4 mm at
least. This confirmed that FNP1 was suitable for imaging
keloid scars found less than 2 mm from the skin surface
(Supporting Information, Figure S5). However, when whole-
skin models with intact epidermis barrier was used, the probe
signal was mainly observed on the skin surface with negligible
signal in the skin dermis (Supporting Information, Figure S5).
These data showed that FNP1 was likely to be trapped in the
uppermost skin layer, failing to cross the epidermis.
To facilitate the transdermal penetration of the hydro-
philic probe, microneedles were employed to create micro-
channels (Figure 3a). Microneedle device (500 mm in height
per needle) is sufficient to insert into skin at the early stages of
scar formation, allowing FNP1 to traverse beyond the
epidermis barrier layer. The microneedles were weighted
down to deform skin at 18-fold pressure magnitude below that
[21]
required to break skin (that is, 300 kPa). After 5 min, the
microneedles were removed and FNP1 was topically applied
to the skin surface (Figure 3a) and incubated for 6 h before
imaging. As shown in Figures 3b,c, the NIR fluorescence of
FNP1 from KF-implanted skin was 14.5, 6.0, and 2.2-fold
higher than unmodified skin, HaCaT-implanted and NDF-
implanted skin, respectively. The specificity of FNP1 to KF
over NDF cells was further validated from skin tissue
histology (Supporting Information, Figure S6a) throughout
the entire skin tissue depth (1.2 mm; Supporting Information,
Figure S6b). These data demonstrated that the microneedle-
assisted topical application of FNP1 allowed it to cross the
epidermis layer to the dermis-residing cells for selective
detection of KF cells.
Figure 2. FNP1 probe in cell culture. a) Fluorescence microscopy of KF
cells after treatment with FNP1 (5 mm, purple) for 1 h and stained with
nucleus indicator (Hoechst 33342, blue) for 30 min and lysosome
indicator (LysoTracker, white) for 30 min. Scale bar: 20 mm. b) Fluores-
cence microscopy of HaCaT, NDF, KF and NDF cells stimulated with
À1
TGF-b1 (10 ngmL ) after treatment with FNP1 (5 mm, purple) for 1 h
and stained with nucleus indicator (Hoechst 33342, blue) for 30 min.
Scale bar: 100 mm. c) Quantification of fluorescence intensities of the
cells (HaCaT, NDF, KF, NDF+TGF-b1) after incubation with FNP1 in
Figure 2b using multiplate reader. The fluorescence intensities of
FNP1 were normalized by total cell nuclei signal (NucBlue). d) Relative
gene expression of FAPa in HaCaT, NDF, KF, NDF+TGF-b1 normal-
ized by GAPDH and NDF expression levels using qRT-PCR. Error bars:
standard deviation from three separate measurements. *:p<0.05,
To evaluate the sensitivity of FNP1 for detection of KF
cells in human skin models, skins implanted with different
*
*:p<0.01.
Angew. Chem. Int. Ed. 2018, 57, 1 – 6
ꢀ 2018 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!