Professional Documents
Culture Documents
net/publication/332801637
CITATIONS READS
30 454
6 authors, including:
All content following this page was uploaded by Anish Hiresha Verma on 02 May 2019.
Bone cancer or osteosarcoma is an aggressive cancer affecting the long bones and is treated by
a combination of surgery and chemotherapy. Local drug delivery directly to the site of bone cancer
and the use of plant-based drugs has been explored towards improving the efficacy and decreasing
the toxicity of the anti-cancer drugs. Curcumin, derived from turmeric is highly effective against can-
cer cells and shows very low toxicity against
IP: 14.139.160.231 On:normal
Thu, cells.
02 MayBone repair
2019 is facilitated by use of bone
05:39:19
substitutes such as bioceramics,
Copyright: American Scientific Publishers (CA) nanocarriers closely
amongst which the carbonated apatite
mimic the natural bone mineral. In the current work,
Delivered we have developed CA nanocarriers based
by Ingenta
local delivery of curcumin as an adjunct treatment for bone cancer. CA nanocarriers with 6 wt.%
carbonate were prepared by wet chemical synthesis using synthetic derived (6SWCA) and eggshell
derived (6EWCA) precursors along with hydroxyapatite (WHA) as a control. The X-ray diffraction
(XRD) patterns showed the CAs to be phase pure with a mean crystallite size of 17 nm. The Fourier-
transform infrared spectroscopy (FTIR) analysis of both CAs indicated the carbonate substitution
as B-Type. The amount of carbonate substitution was observed to be around 6 wt.% using FTIR
and CHNO elemental analyzer. The 6EWCA showed a greater loading (36%) and release (66%) of
curcumin than 6SWCA and WHA nanocarriers. The bovine serum albumin (BSA) protein denatura-
tion assay showed the curcumin loaded CAs to be highly anti-inflammatory while their anti-cancer
activity was confirmed by the high cytotoxic activity against MG-63 human osteosarcoma cells. Con-
clusively, an eggshell derived apatite drug delivery system was found to be very suitable to cure
osteosarcoma, prevent post-cancer inflammation and modulate bone repair and regeneration.
Keywords: Osteosarcoma, Eggshell, Carbonated Apatite, Curcumin, Local Drug Delivery,
Anti-Inflammatory, Anti-Cancer.
1. INTRODUCTION femur, 23% in the proximal tibia and 10% in the humerus.1
Osteosarcoma (OS) is defined as the most common type of These cancer cells often show a metastatic growth in the
cancer, which starts in the bone, mainly affecting children pulmonary area leading to failure of the respiratory system
and young adults. It is an aggressive malignant neoplasm and subsequent death.2 The conventional treatment for OS
that exhibits uncontrolled osteoblastic differentiation and is is a combination of neoadjuvant chemotherapy, followed
characterized by spindle cells of mesenchymal origin pro- by surgery and then adjuvant chemotherapy.3 Neoadjuvant
ducing malignant osteoid. It exhibits high affinity towards chemotherapy is defined as a chemotherapy given before
the metaphysis of long bones, with 43% in the distal any surgical treatment whereas adjuvant chemotherapy
involves therapy post-surgery. Chemotherapy involves the
∗
Author to whom correspondence should be addressed. use of synthetic anti-cancer drugs such as methotrexate,
doxorubicin, cisplatin and ifosfamide to kill the cancer comprising carbonated apatite nanocrystals (∼3–6 wt.%
cells by interfering with their cell division cycle.3 Despite CO2−3 of platelet-shaped morphology embedded in an
the chemotherapy, OS has one of the lowest survival rates organic matrix composed largely of type-1 collagen with
of any pediatric cancer.4 This is because of the adverse minor amounts of other collagens, glycoproteins, proteo-
effects of the chemo-drugs such as nausea, fatigue, hair glycans and sialoproteins.23 The carbonated apatite (CA)
loss, anaemia, diarrhoea, constipation, nerve and muscle can be classified depending on the location of carbon-
ailments, appetite and weight reduction, urinary infections, ate substitution as A-type and B-type.23 A-type carbonated
5 apatite has the carbonate ion substituting the hydroxyl sites
and fertility problems.
Natural products have been an overwhelming success in whereas in B-type, carbonate ion substitutes at the phos-
our society and are the upcoming trend for the replace- phate site. It has been reported that only the B-type carbon-
ment of the synthetic chemo drugs due to the adverse ated apatite is present in bones and teeth.23 Nanocrystalline
effects. Amongst these products, curcumin has gained spe- carbonate-apatites have been studied for the release of anti-
cial attention for the treatment of cancer due to its anti- cancer platinum bisphosphonates.24 But the carbonate was
6
cancer potency. Curcumin is a phytochemical derived on the surface of the materials by adsorption during the
from the rhizomes of a very popular Indian spice, turmeric storage and was mainly retained to mimic the biological
(Curcuma Longa L.). Curcumin has already been used apatite. Recently, carbonated apatite (CA) nanoparticles
in medical applications as an Alzheimer’s therapeutic,7 with a molecular formula of Ca10 (PO4 6−X (CO3 X (OH)2
contraceptive extract,8 and infertility cure.9 Curcumin is loaded with paclitaxel showed enhanced efficacy against
reported to act as a powerful anti-oxidant with strong cancer cells both in a cell culture and an animal model
anti-carcinomic and anti-inflammatory abilities due to the compared to free paclitaxel.25 Surface modification with
presence of numerous reactive functional groups.10 These poly(ethylene glycol) or citrate has also shown more effec-
groups can interact with various molecular targets such as tive uptake of drug-loaded CA nanoparticles by the tumor
transcription and growth factors, protein kinases, cytokines cells compared to the uncoated particles and the free
and enzymes leading to multiple cell signalling resulting drug.26 27
in apoptosis of the cancer cell.10 Moreover, it is highly Eggshells are mainly composed of calcium carbonate
biocompatible to normal cells and even at high dosage, with traces of strontium, magnesium, fluorine, silica and
28
ranging from 1.12 to 2.5 g per day of curcumin, has been sodium. The synthesis of CAs from eggshells by par-
IP: 14.139.160.231
11 On: Thu, 02
tially May 2019its05:39:19
retaining carbonate content has been demonstrated
reported to be non-toxic to normal tissues. Hence,
Copyright: cur- Scientific Publishers
American 28
cumin is as an ideal choice as a potential anti-cancer earlier
drug. by Ingenta by our group. Inspired by the advantages of CA
Delivered
However, owing to its poor solubility (11 ng/mL), stability for cancer drug delivery applications, we have studied the
and rapid metabolism, efficient delivery of curcumin still potential of curcumin delivery by eggshell derived CA
remains a challenge in achieving therapeutic outcomes. 12 (6EWCA) nanoparticles for the first time to the best of
A localized nano-drug delivery system (NDDS) can available literature in comparison with synthetic derived
deliver the hydrophobic drug in high payload with CAs. The curcumin loaded CAs was also evaluated for
their anti-inflammatory activity and anti-cancer activity by
enhanced efficacy to the cancer cells directly at the
in vitro cell culture studies. As bone substitutes are nec-
target site, thereby minimizing the drug plasma con-
essary after the surgical removal of the bone tumor, the
centration in the bloodstream or any other organs and
curcumin-loaded 6EWCA is anticipated to perform both
ultimately aiding patient’s comfort.13 Many materials
bone regeneration and anticancer drug release features to
have been reportedly used as an NDDS with the lat-
inhibit tumor re-growth.
est including hybrid polymer-drug conjugates,14 meso-
porous biosilica,15 nano-lipids,16 magnetic nanoparticles,17
electrospun nanofibers,18 hydrogels19 and graphene-based 2. MATERIALS AND METHODS
nanomaterials.20 However, the ideal carrier for any local 2.1. Materials
bone therapeutics will be calcium phosphate (CaP) Precursors for preparation of CAs like diammonium
nanocarriers due to their composition mimicking the hydrogen phosphate [(NH4 2 HPO4 ], calcium hydroxide
bone mineral, availability of various binding sites, high [Ca(OH)2 ] and calcium carbonate (CaCO3 were obtained
biocompatibility and biodegradability.21 Additionally, the from Merck, Mumbai. The drug curcumin (from Curcuma
localized CaP based delivery system provides an added longa) was purchased from Sigma Aldrich, India. Ethanol
advantage of bone repair and regeneration due to its osteo- (99.9%), used as a drug loading media, was obtained
genesis ability, which is very much desirable for post- from Changshu Hong sheng Fine Chemical Co. Ltd.,
cancer healing.22 China. Chemicals for preparation of phosphate buffer solu-
Amongst the various CaPs, carbonated apatite is a suit- tion (PBS) like sodium biphosphate dihydrate (NaH2 PO4 ·
able choice since its composition very closely mimics the 2H2 O) and disodium hydrogen phosphate anhydrous
biological composition of human bone and teeth. Both (Na2 HPO4 were obtained from SD Fine Chemicals,
bone and teeth typically consist of an inorganic phase Mumbai and Merck, Mumbai respectively. Chemicals used
for anti-inflammatory assay like bovine serum albumin dispersed and sonicated in water media, followed by anal-
(BSA) and glacial acetic acid (C2 H4 O2 were purchased ysis at 25 C.
from SRL Pvt. Ltd., Mumbai. Chemicals for cell culture
studies like thiazolyl blue tetrazolium bromide or MTT 2.4. In Vitro Curcumin Loading and Release Studies
(C18 H16 BrN5 S), dimethyl sulphoxide or DMSO (C2 H6 OS), The in vitro curcumin loading and release studies were per-
trypan blue dye and minimum essential medium (MEM) formed similar to the method as reported earlier.29 50 mL
were purchased from Himedia Lab, Mumbai. of 1 mg/mL solution of curcumin-ethanol was prepared to
which 50 mg of nanocarriers were added. The resultant
2.2. Synthesis of CAs solution was left undisturbed for 24 h at 4 C. After 24 h,
CAs were prepared from synthetic sources as well as natu- 2 mL of the supernatant was removed for the estimation
ral eggshell sources as reported28 using wet chemical syn- of the residual drug concentration using a UV-Vis spec-
thesis method. The carbonate substitution in CAs was kept trophotometer (Lambda 35, PerkinElmer, USA). The peak
at 6 wt.% since it mimics the natural bone composition. at 424 nm was used to determine the curcumin concentra-
The synthetic precursors used were Ca(OH)2 , CaCO3 and tion, and thereby calculating the drug loading percentage.
(NH4 2 HPO4 . 0.24 M (NH4 2 HPO4 was added dropwise The remaining solution was then filtered and the drug-
to a 0.4 M Ca(OH)2 and 6 wt.% CaCO3 solution to obtain loaded nanocarriers were obtained. The amount of drug
calcium- to-phosphate ratio (Ca/P) of 1.67. The resultant loaded onto the nanocarriers is calculated by the below
mixture was kept for 2 days for proper aging and then equation:
filtered. The precipitate obtained was then washed with
% Drug Loading = Ic − Fc /Ic × 100 (1)
distilled (DI) water to remove any unwanted ions and oven
dried at 100 C. Finally, the precipitate was grinded using where Ic and Fc are the initial and the final concentration
a mortar and pestle to obtain a fine powder. of curcumin in the solvent.
Eggshell derived CAs were synthesized using hen’s The release studies were performed by dispersing the
eggshells as a calcium source. Hen’s eggshells were col- drug-loaded nanocarriers in a specially designed release
lected and their surface was manually cleaned by strip- media of pH 7.4 in a constant temperature water bath at
ping the inner surface membrane. They were then washed 37 C. The release media consisted of PBS and ethanol in
with water followed by overnight oven drying at 100 C
IP: 14.139.160.231 On: Thu,the 02ratio
Mayof 4:1. 05:39:19
2019 A standard calibration curve was plotted
and then powdered using a mortar andCopyright:
pestle. These pow- between the absorbance
American Scientific Publishers and the various concentration of
ders were then heated to 805 C for 1 h in theDelivered
furnace, bycurcumin
Ingenta (g/mL) in the release media. About 2 mL of
followed by immediate air quenching to obtain Ca(OH)2 the supernatant was removed from the release system for
with 6 wt.% arrested residues of carbonate content.28 The UV-Vis characterization and was replaced by 2 mL freshly
product was then washed and powdered similar to syn- prepared release media. The readings were noted at a reg-
thetic CAs. Additionally, a batch of synthetic hydroxya- ular interval of 2 h up to 12 h and then at 24th and 48th h.
patite (WHA) sample was also prepared using the wet The experiment was done in triplicates. The drug release
chemical as a reference. The synthetic CA and eggshell was calculated by:
derived CA samples were coded as 6SWCA and 6EWCA
respectively. % Drug release = Fr /Ir × 100 (2)
where Qt is a fraction of drug released at a time ‘t’, k is were solubilized in DMSO and the absorbance was mea-
the release rate constant (or Higuchi’s constant). sured at 570 nm using a multimode plate reader (EnSpire,
PerkinElmer, Singapore).
2.6. In Vitro Anti-Inflammatory Assay The percentage of viable cells was calculated as the per-
The anti-inflammatory studies of curcumin loaded CAs centage relative to the control using the equation:
were carried out by inhibition of albumin denaturation
% Cell viability = Sample OD/Control OD×100 (6)
assay.30 1% aqueous solution of BSA was prepared and the
pH was adjusted to 6.5 using the glacial acetic acid. 10 mg where OD signifies optical density measured by the mul-
of the curcumin-loaded nanocarriers (WHA, 6SWCA and timode plate reader.
6EWCA) were suspended in 10 ml of release media and
incubated at 37 C for 24 h. 1 mL of the supernatant from 2.8. Cellular Uptake Studies
the curcumin-loaded nanocarriers was added to 3 mL of
1 mg of CA nanocarriers were added to 1 mg/mL of rho-
BSA solution. The solution was then heated to 52 C for
damine B and sonicated for 30 min and then stirred at
20 min, and the solution’s absorbance was then measured
50 rpm for 24 h. The samples were then centrifuged and
at 660 nm by UV-Vis spectroscopy. Pure samples (no drug)
reconstituted with MEM and sonicated for 30 min in a
were taken as control. The experiment was performed in
bath sonicator. MG-63 cells were seeded in a petri dish
triplicates. The percentage inhibition of protein denatura-
with a density of 106 cells and incubated at 37 C for 24 h.
tion was calculated as follows:
The rhodamine tagged samples were added and incubated
% inhibition = Ac − As /Ac × 100 (5) for 24 h after which the medium was removed and the
cells were washed three times with PBS to remove the
where Ac is the absorbance of control and As is the excess dye. The cells were then imaged using fluorescence
absorbance of samples. microscopy (Olympus IX51-URFLT, Japan).
of B-type carbonated apatites. Additionally, the absence of Table III. Mean particle size and zeta potential calculations of WHA,
−1 6SWCA and 6EWCA samples.
CO2−3 at 2349 cm indicates that no residual carbonate is
present and all the CO2−3 have been substituted in place of Samples Size of particles (dia. nm) Zeta potential (mV)
PO3− . On further analysis, it was found out that the relative
4 WHA 208 ± 14.7 −17.5 ± 3.4
intensities at 962 and 1091 cm−1 for 6EWCA and 6SWCA 6SWCA 156 ± 10.6 −21.8 ± 7.9
samples were reduced when compared to WHA sample as 6EWCA 106.2 ± 2.8 −21.5 ± 4.6
3−
a result of CO2−3 replacing PO4 in the hexagonal lattice
structure. Moreover, for CA samples, the OH− vibration
peak at 632 cm−1 was found to reduce in intensity signify- 3.2. Curcumin Loading and Release Studies
3− −
ing the CO2− 3 substituting for PO4 and formation of OH
The loading of curcumin on CAs and HA are shown in
vacancies for charge compensation. Figure 3. The WHA has a loading of 20.8%, 6SWCA has
The CO2− 3 substitution in the B-type carbonated apatite
32.7% and 6EWCA has 35.6%. Both the CA samples have
was also quantified by FT-IR spectra33 and CHNO- higher curcumin loading than HA. As discussed earlier,
elemental analyzer, as shown in Table II. Both confirmed a the carriers with lower crystallinity index (CI) will have
6 wt.% substitution in the 6EWCA and 6SWCA samples. high drug absorption capacity. Hence, CAs with lower
The formula used to calculate the wt.% of carbonate from CI exhibits higher loading, as expected. Amongst the two
the FTIR spectra is shown below:33 CA samples, 6EWCA has a slightly higher loading than
6SWCA. It may be due to the presence of ions in 6EWCA
wt % CO3 = 28 62 × rc/p + 0 0843 (13) inherited from the eggshell.
2− The cumulative release profile of curcumin-loaded
where rc/p is the area ratio between the 3 CO3
−1 3− 6EWCA, 6SWCA and WHA were plotted as shown in
band (1550–1330 cm and the 1 3 PO4 band
Figure 4. The curcumin-release from WHA, 6SWCA and
(1230–900 cm−1 .
6EWCA was found out to be 45.48%, 58.57% and 66.24%
The particle size and zeta potential of the apatites are
respectively, confirming the role of CO2− 3 in the release
listed in Table III. It was found that both the CA samples
of the curcumin from the nanocarriers. Moreover, it was
were significantly smaller than the HA sample. The WHA
found that 6EWCA had a greater release than 6SWCA due
sample had a mean size of 208 nm whereas 6SWCA and
to the inherited trace elements present in 6EWCA.
6EWCA had a mean size of 156IP: and14.139.160.231
106 nm respectively.
On: Thu, 02TheMay 2019kinetics
release 05:39:19 of curcumin-loaded nanocarriers
The particle size obtained from DLS Copyright:
are typicallyAmerican
higher Scientific Publishers
were studied using both Korsmeyer-Peppas and Higuchi
since it reports hydrodynamic radius as an intensity dis- by Ingenta
Delivered
models. The release rate constants (K) were calculated
tribution. The 6EWCA had a hydrodynamic diameter less
from the slope of the plots using regression analysis. The
than that of 6SWCA. The particle size is said to play a
parameters of release kinetics are listed in Table IV. The
key role in drug circulation and absorption into tissue and
34 coefficient of correlation (R2 ranged in the domain of 0.97
cellular membranes. It is reported that the smaller are the
to 0.99, depicting a very good fit of the data with both
particles, higher will be the particle uptake by the cells.34
the mathematical models. The diffusion coefficient (n) is
Zeta potential is a term used to quantify the electroki-
an important entity that gives the drug release mechanism.
netic potential in a colloidal system. The zeta potential is
As per the model, if n < 0 5, it indicates Fick’s diffusion,
said to affect the pharma kinetics as well as the phagocy-
35
tosis of the drug-loaded nanocarriers in the bloodstream.
It has been reported that the highest cellular uptake was
found out to be in the samples having higher negative zeta
potential value.35 As shown in Table III, 6EWCA tends to
have a higher negative zeta potential value of −21.5 mV
whereas WHA shows the least, i.e., −17.5 mV. As the
smaller particles and higher negative zeta potential tend to
show a higher affinity for cellular membrane due to elec-
trostatic interactions,35 it is expected that 6EWCA will be
having a higher affinity towards osteosarcoma cell uptake
than any other sample.
Samples % CO2−
3 (FTIR) % CO2−
3 (CHNO-analyzer)
if 0 5 < n < 0 85 drug release follows non-Fickian diffu- Figure 5. Anti-BSA denaturation studies of curcumin loaded HA (Cur-
WHA) and CAs (Cur6SWCA and Cur6EWCA) nanocarriers (n = 3; val-
sion and if n > 0 85, it indicates a zero-order release.36 The ues shown as mean ± SD).
WHA sample followed two-step kinetics, i.e., non-Fickian
mechanism to 10 h, followed by a steady Fickian release
up to 48 h. On the contrary, the CAs followed three-step induces a greater anti-inflammatory activity. Amongst
kinetics. Till 4 h, it was non-Fickian and then followed by CAs, the 6EWCA showed a higher inhibition of 71.1%.
Fickian release to 48 h but with different release constants The significant difference between 6EWCA (71.1%) and
for 4 to 12 h and 12 to 48 h. 6SWCA (52.3%) may be due to the presence of trace ions,
which might themselves be involved in suppressing protein
IP: 14.139.160.231
3.3. In Vitro Anti-Inflammatory, Cell Viability and On: Thu, 02 May 2019 05:39:19
denaturation.
Copyright: American Scientific Publishers
The results of the MTT assay on MG-63 human
Uptake Studies Delivered byosteosarcoma
Ingenta cell lines performed at 5 mg/mL nanocar-
Figure 5 shows the result of anti-BSA denaturation study
by the curcumin-loaded nanocarriers. Denaturation of pro- rier concentration for 3 days are shown in Figure 6.
teins is a well-documented cause of inflammation26 and The nanocarriers (6SWCA and 6EWCA) were found to
it has been reported that even though if cancer cells are be highly biocompatible (>80%) and showed a greater
destroyed, prolonged chronic inflammation can lead to cell viability for all three days. On curcumin load-
regression.37 Therefore, the inhibition of protein denat- ing, the nanocarriers displayed a reduction in cell via-
uration was studied as a mean of evaluating the anti- bility. After the 3rd day, only 6.97% (in the case of
inflammatory activity of curcumin loaded nanocarriers.
The percentage of inhibition from WHA was found to
be 22.3%. The drug-loaded CA samples (6EWCA and
6SWCA) showed a higher inhibition than WHA sam-
ple. It may be concluded that the presence of carbonate
Table IV. Release kinetic profile displaying the diffusion coefficient (n),
release constant (K) and time-dependent release mechanism for curcumin
release from WHA, 6SWCA and 6EWCA nanocarriers.
17. X. Mou, Z. Ali, S. Li, and N. He, J. Nanosci. Nanotechnol. 15, 54 29. T. S. Sampath Kumar, K. Madhumathi, B. Rajkamal, S. Zaheatha,
(2015). A. Rajathi Malar, and S. Alamelu Bai, Colloids Surf. B: Biointer-
18. R. Sridhar, R. Lakshminarayanan, K. Madhaiyan, V. Amutha faces 123, 542 (2014).
Barathi, K. H. C. Lim, and S. Ramakrishna, Chem. Soc. Rev. 44, 790 30. K. Madhumathi, Y. Rubaiya, M. Doble, R. Venkateswari, and T. S.
(2015). Sampath Kumar, Drug Deliv. Trans. Res. 8, 1066 (2018).
19. S. Merino, C. Martín, K. Kostarelos, M. Prato, and E. Vázquez, ACS 31. R. S. Jayasree, J. Velkumar, and T. S. Sampath Kumar, Dent. Oral
Nano 9, 4686 (2015). Craniofac. Res. 4, 1 (2017).
20. Q. Zhang, Z. Wu, N. Li, Y. Pu, B. Wang, T. Zhang, and J. Tao, Mat. 32. Y. Sa, Y. Guo, X. Feng, M. Wang, P. Li, Y. Gao, X. Yang, and
Sci. Eng. C 77, 1363 (2017). T. Jiang, New J. Chem. 41, 5723 (2017).
21. S. Bose and S. Tarafder, Acta Biomat. 8, 1401 (2012). 33. A. Grunenwald, C. Keyser, A. M. Sautereau, E. Crubézy, B. Ludes,
22. K. Madhumathi and T. S. Sampath Kumar, Biomed. Mat. 9, 035002 and C. Drouet, J. Archaeol. Sci. 49, 134 (2014).
(2014). 34. A. Banerjee, J. Qi, R. Gogoi, J. Wong, and S. Mitragotri, J. Control.
23. Sampath Kumar, I. Manjubala, and J. Gunasekaran, Biomaterials Release 238, 176 (2016).
21, 1623 (2000). 35. S. Honary and F. Zahir, Trop. J. Pharm. Res. 12, 255
24. M. Iafisco, B. Palazzo, G. Martra, N. Margiotta, S. Piccinonna, (2013).
G. Natile, V. Gandin, C. Marzano, and N. Roveri, Nanoscale 4, 206 36. S. Racovita, A. L. Vasiliu, A. Bele, D. Schwarz, C. Steinbach,
(2012). R. Boldt, S. Schwarz, and M. Mihai, RSC Adv. 8, 23274
25. T. Fatemian and E. H. Chowdhury, Toxics 6, 12 (2018). (2018).
26. F. S. Mozar and E. H. Chowdhury, Pharmaceutics 9, 21 (2017). 37. S. M. Crusz and F. R. Balkwill, Nat. Rev. Clin. Oncol. 12, 584
27. S. M. Hossain and E. H. Chowdhury, Pharmaceutics 10, 32 (2018). (2015).
28. R. Jayasree, K. Madhumathi, D. Rana, M. Ramalingam, R. P. 38. M. M. Germaini, R. Detsch, A. Grünewald, A. Magnaudeix,
Nankar, M. Doble, and T. S. Sampath Kumar, J. Nanosci. Nanotech- F. Lalloue, A. R. Boccaccini, and E. Champion, Biomed. Mater.
nol. 18, 2318 (2018). 12, 035008 (2017).