Understanding the Incorporation and Release of Salicylic Acid in Metal-Organic Frameworks for Topical Administration

Article abstract of DOI:10.1002/ejic.202001134

Although metal-organic frameworks (MOFs) have been widely demonstrated to be great candidates for drug delivery applications, they have been mainly proposed for the intravenous route. Here, eight highly porous benchmarked MOFs, with different topologies a

Full text of DOI:10.1002/ejic.202001134

Full Papers
doi.org/10.1002/ejic.202001134
Understanding the Incorporation and Release of
Salicylic Acid in Metal-Organic Frameworks for Topical
Administration
Sara Rojas[a] and Patricia Horcajada*[a]
Although metal-organic frameworks (MOFs) have been widely
demonstrated to be great candidates for drug delivery
applications, they have been mainly proposed for the intra-
venous route. Here, eight highly porous benchmarked MOFs,
with different topologies and structures, are proposed for the
topical delivery of salicylic acid (SA), an important, but highly
reactive and unstable, therapeutically active metabolite of
aspirin. The microporous Zr aminoterephthalate UiO-66-NH2
was selected as the most promising SA carrier, achieving
important loadings (12.1 wt.%). Finally, the SA delivery process
was studied under simulated cutaneous conditions (aqueous
°
media at 37 C), reaching a plateau in 6 h (with ~64% or
~105.6 mg of released SA per g of UiO-66-NH2). These results
demonstrate the suitability of UiO-66-NH2 for the topical
controlled release of SA, making this formulation a promising
candidate for the development of new devices for skin treat-
ment.
Introduction
gases),[10–12] with controlled releases under physiological con-
ditions.
During the last decades, the development of controlled drug
delivery systems (DDS) for the treatment of human diseases is
growing very rapidly. DDS are designed to get a specific
transport and progressively release of active pharmaceutical
ingredients (APIs) for a long timeframe, protecting them from
potential biodegradation and modifying their physicochemical
properties (e.g., solubility). Polymer-based drugs and DDS
emerged from the laboratory bench in the 1900s as a promising
therapeutic strategy for the treatment of certain devastating
illnesses.[1] Among them, metal-organic frameworks (MOFs) are
considered as great candidates for drug delivery applications,
not only providing progressive release of the APIs but also
modifying their biodistribution.[2–4] These hybrid materials,
comprised of inorganic nodes and organic polycomplexant
linkers that assemble into multidimensional periodic lattices
through coordination bonds,[5] present several advantages as
DDS: i) exceptional porosity associated with high drug cargoes;
ii) highly versatile structures and compositions, potentially
tunable depending on target API/site; iii) appropriate biological
compatibility, stability and particle size; and iv) amphiphilic
internal microenvironment, adapted to a large number of APIs,
among others.[6] MOFs accomplished remarkable loadings (often
much higher that traditional carrier systems) of a large variety
of active molecules (drugs,[3,7,8] cosmetics,[9] biological
MOF-based DDS have been mainly proposed for the intra-
venous administration route, remaining other routes less
studied Particularly, the skin, the largest organ in the body, is
an attractive route for the delivery of therapeutics and
represents a proven and smart option for treating a variety of
cutaneous diseases (e.g., acne, lymphoma, leishmaniosis). For
instance, the topical drug delivery market is projected to reach
110 billion euros by 2025 (from 80 billion euros in 2020), at a
compound annual growth rate (CAGR) of 6.4% during the
forecast period.[13]
Although there have been many innovations in the
development of new DDS, the number of effective topical
cutaneous drugs remains small, primarily because of their
instability within the formulation or skin, their uncontrolled
release and the poor control of their skin permeation.[14]
Nonsteroidal anti-inflammatory drugs (NSAIDS) are a class
of Cox-1 and/or Cox-2 inhibitors that suppress inflammation.[15]
Particularly, the anti-inflammatory salicylic acid (SA), an impor-
tant active metabolite of aspirin, has received significant
attention because of its additional antibacterial, antifungal, and
antipyretic activities. It has also shown promising results in the
prevention of cardiovascular diseases and cancer.[16,17] Thus, an
adapted release of SA may be desired for cutaneous antimicro-
bial and anti-inflammatory applications, such as treatment of
infections or, even further, useful in biomedical devices as
coatings. However, as a consequence of its highly reactive
nature, its instability (half-life of 2–30 h)[18] is a major concern
that needs to be addressed in order to use it as a potent topical
drug.
[a] Dr. S. Rojas, Dr. P. Horcajada
Advanced Porous Materials Unit (APMU)
IMDEA Energy Institute
Av. Ramón de la Sagra 3, 28935 Móstoles-Madrid, Spain
E-mail: patricia.horcajada@imdea.org
Motivated by our prior studies on the use of MOFs for the
cutaneous delivery,[9,19] and as impressive adsorbents of SA in
gastrointestinal conditions,[20,21] here we propose the use of
MOFs for the SA delivery. To reach a maximum SA cargo while
Supporting information for this article is available on the WWW under
https://doi.org/10.1002/ejic.202001134
Part of the joint “Metals in Medicine” Special Collection with ChemMed-
Chem.
Eur. J. Inorg. Chem. 2021, 1325–1331
1325
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/ejic.202001134
understanding the adsorption process, we have rationally
on zirconium(IV) oxoclusters and terephthalate anions (H2BDC-
X: 1,4-benzenedicarboxilic acid, where X=H, NH2) UiO-66 and
UiO-66-NH2 ([Zr6O4(OH)4(C8O4H3X)6]·nH2O);[28,29] 2) the flexible
porous MIL-163 or [Zr2(C26H6O6N4)2]·(DMA)5(H2O)14; (DMA=N,N-
dimethylamine) based on zirconium(IV) cations coordinated by
1,2,3-trioxobeneze groups;[30] 3) the microporous MIL-127
structure or [Fe3O(OH)0.88Cl0.12(C16N2O8H6)1.5(H2O)3]·nH2O based
on iron(III) octahedra trimers and 3,3’,5,5’-azobenzenetetracar-
boxylate anions;[31,32] 4) the mesoporous MIL-100 or [Fe3O(H2O)
OH(C9H6O6)]·nH2O (H3BTC: 1,3,5-benzenetricarboxylic acid or
trimesic acid) based on iron(III) octahedra trimers and trimesate
anions;[33] 5) two flexible iron(III) terephthalates denoted as MIL-
53 and MIL-53-(OH)2 ([Fe(OH)(C8O4H2X2)], where X=H and OH,
respectively), which are 3D porous solids built up from chains of
corner-sharing [FeO4(OH)2] octahedra, connected through ter-
ephthalate linkers with potential structural adaptability for
guest molecule diffusion;[34] and 6) the microporous zeolitic
imidazolate framework ZIF-8 or [Zn(C4H5N2)2], based on zinc(II)
interconnected with 2-methylimidazole ligands (2-mIm) leading
to a sodalite topology (see detailed porous properties in the
Table 1).[35]
selected 8 highly porous benchmarked MOFs with different
topologies and structures. First, a deep screening of the SA
adsorption capacity of the MOFs and their stability was carried
out, leading to the selection of the Zr aminoterephthalate UiO-
66-NH2. A deep study on the SA adsorption of UiO-66-NH2 was
performed using a large variety of techniques (powder X-ray
diffraction-PXRD, thermogravimetric analysis-TGA, Fourier trans-
form infrared-FTIR, N2 sorption, elemental analysis-EA, high
performance liquid chromatography-HPLC, etc.). Finally, the SA
release from the loaded SA@UiO-66-NH2 was evaluated under
simulated cutaneous conditions.
Results and Discussion
Screening of the SA adsorption capacity and MOF stability
In order to obtain an efficient SA formulation for topical
treatment, a high drug loading capacity rate was achieved
through the selection of the best SA adsorbing material from a
series of 8 MOFs. The selection of the MOF candidate for SA
adsorption was based on the following criteria: i) exceptional
porosity, compatible with the SA dimensions (7.5×6×2 Å3;
estimated by Vesta considering Van der Waal radii, Figure 1),[22]
ii) their biocompatibility (e.g., no sign of toxicity after the in vivo
intravenous administration of very high doses of MIL-100;[23]
after the oral administration of 1 gKgÀ 1 of MIL-127;[20] the
exposure of zebrafish embryos with 200 μM of UiO-66 or MIL-
100;[24] UiO-66, UiO-66-NH2, MIL-53 or ZIF-8 against different cell
lines),[25–27] or biosafety when using in humans cutaneous
treatments,[19] and iii) a priori remarkable aqueous stability,
which will avoid the possible toxicity of their degradation
products (linkers and metals) during the drug delivery process.
The series comprises: 1) the microporous UiO-66 series based
First, SA encapsulation was studied in the selected porous
MOFs by using two different MOF:SA molar ratios (1:2 and
1:0.5, see Experimental Section for further details-adsorption
kinetics, fitting, and stability studies-, Supporting Information
S1) to evaluate the influence of the drug concentration in the
adsorption process. Although the SA incorporation was fast in
all the studied MOFs, reaching a plateau after only 1 or 4 h, the
amount of adsorbed SA strongly depends on the MOF and the
MOF:SA ratio used. The amount of adsorbed SA (mggÀ 1) in the
tested materials follows similar order for each studied ratio,
decreasing as follows: UiO-66-NH2 >ZIF-8>MIL-53�UiO-66>
MIL-53-(OH)2 �MIL-100�MIL-163�MIL-127 for MOF:SA ratio
1:0.5; and MIL-53�UiO-66-NH2�MIL-53-(OH)2 >UiO-66>MIL-
163>ZIF-8>MIL-100�MIL-127 for MOF:SA ratio 1:2. Although
the SA (7.5×6×2 Å3) is accessible to all the selected materials
(see channels and windows sizes in Table 1), we can tentatively
relate the SA adsorption capacity with the porosity of the
framework (pore size/windows dimensions). Therefore, it is
reasonable to suggest that high SA adsorption will be achieved
in highly porous structures. However, the SA insertion experi-
ments highlight a remarkable SA adsorption capacity in MOFs
with a moderate porosity. For example, UiO-66-NH2 shows a
lower surface area than other studied MOFs (e.g., lower that
UiO-66), but a relatively high SA adsorption capacity (4-fold
higher that UiO-66). Therefore, other factors may influence on
SA capacity. In parallel, the chemical and structural stability of
the MOFs during the SA adsorption experiment was tested by
means of the leaching of the constitutive ligands by HPLC, and
the crystallinity of the structure by PXRD, respectively. In SA
aqueous solution, the chemical stability of the frameworks
decreased in the following order: MIL-163>MIL-53�UiO-66-
NH2 >MIL-127>UiO-66>MIL-100>ZIF-8�MIL-53-(OH)2
(Ta-
Figure 1. Schematic view of the structure of UiO-66-NH2 (zirconium poly-
hedra, nitrogen, oxygen, and carbon atoms are represented in green, blue,
red, and brown, respectively; the hydrogen atoms are omitted for clarity),
its formula, and the description of its structure and porosity. SA chemical
structure and its dimensions (size calculated from Vesta free software
considering van der Waals radii) are also given.
ble 1). Further, PXRD patterns of the SA-containing MOFs
evidenced that the SA loading process do not alter the
crystallinity of the MOFs, except for the MIL-53 and MIL-53-
(OH)2, with a partial loss of crystallinity (Supporting Information,
Eur. J. Inorg. Chem. 2021, 1325–1331
www.eurjic.org
1326
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/ejic.202001134
Table 1. Pores and molecules dimensions (Å), Brunauer, Emmett and Teller surface areas (SBET, m2 ·gÀ 1), pore volume (Vp, cm3 ·gÀ 1), MOF:SA ratio, adsorbed
SA (mol·molÀ 1, and mg·gÀ 1), and MOF degradation (%) for all studied materials.[a]
Structure
Pore or molecule size [Å]
Starting solid
Initial ratio
(MOF:SA)
Adsorbed SA
Adsorbed SA
MOF degradation
[%]
SBET [m2 ·gÀ 1
]
[mol·molÀ 1
]
[mg·gÀ 1
]
Vp [cm3 ·gÀ 1
]
SA
7.5×6×2 Å3
–
–
–
–
–
UiO-66
Td (8 Å) & Oh (11 Å) cages accessible
via triangular windows (5–7 Å)
1160
0.50
950
1:2
1:0.5
1:2
1:0.5
1:2
1:0.5
1:2
1:0.5
1:2
1:0.5
1:2
1:0.5
0.42
0.15
1.54
0.75
0.06
0.01
0.01
0.01
0.63
0.03
0.14
0.01
0.10
0.02
34.9
12.8
121.42
59.0
17.9
3.6
1.7
1.6
367.9
14.5
71.8
7.4
17.44�0.13
20.57�0.53
5.49�0.03
4.07�0.07
0.00�0.00
0.00�0.00
4.17�3.48
11.68�4.58
0.78�0.0
UiO-66-NH2
MIL-163
0.34
Square channels 12×12 Å2
90–170[a]
MIL-127
1D channels (6 Å) & cages (10 Å)
accessible via 3 Å apertures
7.5 Å (hydrated form)
1300
0.70
–
MIL-53
6.30�0.01
41.94�0.12
26.88�0.05
23.75�0.35
21.12�0.37
MIL-53-(OH)2
MIL-100
9.1 Å (hydrated form)
–
cages (25 & 29 Å)
accessible via microporous windows
(4.8–5.8 & 8.6 Å)
1650
0.75
1:2
1:0.5
2.0
3.8
ZIF-8
11.6 Å accessible through
3.4 Å windows
1135
0.54
1:2
1:0.5
0.01
0.03
7.8
20.5
43.53�0.10
23.68�0.08
[a] Flexible matrix confirmed by a low N2 adsorption compare to the expected from the structural analysis; Td, tetrahedral; Oh, octahedral.
Figure S2). This phenomenon could be related with the replace-
ment of the carboxylate groups of the H2BDC ligand by the
carboxylic acid group of SA, leading to the framework
degradation. This degradation is further confirmed by the
Considering that UiO-66-NH2 leads to the highest SA cargo
(5.9 and 12.1 wt.% at MOF:SA 1:0.5 and 1:2 ratios, respectively),
an impressive chemical and structural stability, and proper
polarity, we selected this MOF for the subsequent cutaneous
release of SA under simulated cutaneous conditions. In fact, SA
content in the UiO-66-NH2 is comparable or even higher than
the one obtained with other materials currently used in the SA
cutaneous delivery (e.g., 2 wt.% in Carbopol®,[43] 5 wt.% cuta-
neous microdialysis,[44] 12 wt.% in chitosan-based nanoemul-
sion-films,[45] 15 wt.% in SA patches,[46] from 2 to 16.7 wt.% in 6
different topical formation,[47] 5–25 wt.% in self-emulsifiable
formulation,[48] or 30 wt.% in polyethylene glycol).[49]
°
appearance of a diffraction peak at 2θ=22.2 , corresponding to
the iron oxide formation,[36] ruling out the use of the MIL-53
series in the SA encapsulation and release.
Aside from the MOF:drug ratio, the porosity/accessibility
and stability of the frameworks, other factors such as the
physico-chemical properties of the MOFs (e.g., hydrophobicity,
polarity, electron polarizability, and size) might strongly
influence on the SA loading.[37] First, SA is considered a
lipophilic and hydrophobic molecule with a partition coefficient
(log Pow) at room temperature of 2.26[38] and therefore, a priori
it will be better adsorbed in hydrophobic matrices. Although
MOFs possess both a polar part (metal clusters) and a more or
less non-polar fraction (aromatic ligands), from all tested
materials, ZIF-8 is the most hydrophobic material, UiO-66 and
UiO-66-NH2 show a moderate hydrophobicity, MIL-127 can be
considered a hydrophilic/hydrophobic structure, and MIL-163,
MIL-100, MIL-53 and MIL-53-(OH)2 could be regarded as more
hydrophilic materials.[39–41] These consideration fall mainly in the
broad that low SA capacity is reached in hydrophilic MOFs
while better total SA cargo is obtained in the hydrophobic
ones. In a further step, and considering only the hydrophobic
MOFs, UiO-66-NH2 shows the highest SA loading rate when
using both MOF:SA molar ratios.
UiO-66-NH2@SA: a detailed study
Considering the previous promising encapsulation results, we
further characterized the SA loaded SA@UiO-66-NH2 (with the
maximum SA amount of 12.1 wt.%). As mentioned before, PXRD
patterns evidence that the drug-loading process does not alter
the crystalline structure of the UiO-66-NH2 (Supporting Informa-
tion, Figure S2). In addition, the absence of Bragg peaks
corresponding to free SA rules out the presence of free
recrystallized drug out of the pores. The drug content,
estimated by HPLC, was confirmed by a combination of
thermogravimetric analysis (TGA) and elemental analysis (EA,
Supporting Information, Figure S3 and Table S2). The high SA
loading (12.1 wt.%) corresponds to 1.54 mol of drug per mol of
material. This value is in total agreement with the encapsulation
rates obtained for pure UiO-66-NH2 or its composites with other
molecules (i.e., 4.9 wt.% cisplatin,[50] and 27 wt.% 5-fluorouracil
in UiO-66-NH2;[51] or 8.7 wt.% of SA in Fe3O4@SiO2@UiO-66-
NH2).[52] Notably, the incorporation of SA within the porosity of
UiO-66-NH2 was confirmed by the reduction of the N2 sorption
This could be related with the presence of the amino
functional group, as it can efficiently interact with SA by polarity
matching and the potential formation of hydrogen bonds
between the amino groups of the MOF and the phenolic
hydroxyl and carboxyl groups of SA (see FTIR study in next
section).[42]
Eur. J. Inorg. Chem. 2021, 1325–1331
www.eurjic.org
1327
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/ejic.202001134
Figure 2. (a) N2 sorption isotherms, and (b) pore distribution with Horvath-Kawazoe (KH) method of activated UiO-66-NH2 (red) and SA@UiO-66-NH2 (blue).
Solid and empty symbols indicate adsorption and desorption branches, respectively.
capacity of the MOF (Figure 2; from the original UiO-66-NH2
with SBET =950 m2 ·gÀ 1 and Vp =0.34 cm3 ·gÀ 1, to the SA@UiO-66-
NH2 SBET =365 m2 ·gÀ 1, Vp =0.11 cm3 ·gÀ 1) and the reduction of
the pore size (Figure 2b). To shed some light on the influence
of porosity on the drug adsorption, we estimated the volume
occupied by one SA molecule inside the MOF by considering
the variation of the MOF pore volume after the drug
encapsulation (ΔVp =0.23 cm3 ·gÀ 1) and the total amount of
loaded drug (1.54 mol·molÀ 1). The larger occupancy volume of
SA in UiO-66-NH2 (257 Å3) compared to the free drug molecular
volume (98 Å3; estimated under vacuum) might be due to the
partial occupancy of the porosity, as confirmed by the
important remaining porosity after drug insertion. In this
regard, one has to consider the correlation between size and
shape of both the pores with the drug, determining whether
the drug can be adsorbed. We note that the triangular windows
of UiO-66-NH2 (ca. 5–7 Å), giving access to both octahedral (Oh)
and tetrahedral (Td) cavities (~11 and ~8 Å, respectively), seem
to be large enough to allow the incorporation of the drug (7.5×
6×2 Å3) in both cages. Therefore, the non-perfectly efficient
packing of the drug into the pores might lead to this important
remaining porosity.
the preferential adsorption of SA in the UiO-66-NH2 when
compared with its non-functionalized UiO-66 analogue.
Drug release
Once the SA@UiO-66-NH2 was fully characterized, we evaluated
its ability to deliver its active cargo to propose it as topical DDS.
°
The delivery of the AS was performed in distilled water at 37 C,
under continuous bidimensional stirring, to assess a potential
cutaneous administration, mimicking the skin hydration. The
release kinetics was determined by quantifying the amount of
delivered drug in the medium by HPLC as a function of time
(Figure 3). According to the drug release profiles, the SA@UiO-
66-NH2 system shows a quite fast and partial SA release. A
plateau in the SA release process from UiO-66-NH2 is reached
after 6 h (with a 64.3�2.1% or 105.6 mg of released SA per g of
UiO-66-NH2). After 24 h, the total amount of released drug is
66.8% (or 103.3 mg of SA per g of UiO-66, Supporting
Finally, the FTIR spectroscopic analysis clearly shows the
presence of the main bands of pure SA in the IR spectrum of
SA@UiO-66-NH2 (Supporting Information, Figure S4). The IR
spectrum of SA@UiO-66-NH2 confirmed a shift in the wave-
lengths in comparison with the bare UiO-66-NH2 (from 3480
and 3367 cmÀ 1 in the UiO-66-NH2, to 3494 and 3380 cmÀ 1 in the
SA@UiO-66-NH2, assigned to the νas(NH2) and νs(NH2) modes,
respectively; and from 1334 and 1256 cmÀ 1 in the UiO-66-NH2,
to 1338 and 1251 cmÀ 1 in the SA@UiO-66-NH2, assigned to the
ν(CÀ N) modes). These shifts between samples may be indicative
of the formation of specific interactions between the drug
moieties and the UiO-66-NH2 framework. One could tentatively
propose the formation of hydrogen bonds between the
carboxylic acid and hydroxyl groups in the SA with the NH2
groups present within the MOF. This result seems to agree with
Figure 3. Released SA from UiO-66-NH2 (blue) and MOF degradation (red)
°
under simulated cutaneous conditions (aqueous media at 37 C). The inset
shows the fitting of the data to a first order kinetic model.
Eur. J. Inorg. Chem. 2021, 1325–1331
www.eurjic.org
1328
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/ejic.202001134
information, Figure S5). The release profile and the time to Experimental Section
reach the plateau in play here is similar when further compared
All reactants were commercially obtained and used without further
to previously delivery studies of SA in water (e.g., 60% in 10 h
from halloysite nanotubes;[53] 50% in 15 min from the micro-
porous zeolite NaY, and the mesoporous SBA-15 and MCM-
41).[54]
The desorption data can be satisfactorily fitted to a first
order kinetic model according to Equation (1):
purification. The 5,5’-(1,2,4,5-tetrazine-3,6-diyl)bis(benzene-1,2,3-tri-
ol) ligand (H6TzGal),[30] and the 3,3’,5,5’-azobenzenetetracarboxylic
acid ligand (H4TazBz)[56] were prepared as previously described in
literature. Triethyl-1,3,5-benzenetricarboxylate (Et3BTC) was pre-
pared by heating under reflux 24 mmol (5 g) of trimesic acid
(H3BTC) in a solution of 2 mL of H2SO4 in 100 mL of ethanol
absolute and for 24 h. The obtained Et3BTC was filtered and washed
with water.
qEÀ qt ¼ qE � eÀ kt
(1)
Synthesis of MOFs
where qE and qt are the amounts of released SA per gram of
MOF (mmol·gÀ 1) at the equilibrium and at the time t(h),
respectively, and k is the first order kinetic constant. In a typical
first order kinetics, the drug release rate exclusively depends on
its concentration. The fitting of the data gives rise to a k value
of 0.68 hÀ 1 which correspond to a t1/2 for the release of SA of
1 h (Figure 3). This is indicative of the SA physisorption on UiO-
66-NH2, ruling out its chemical binding. To confirm the stability
of the MOF matrix, the leaching of the ligand to the water
medium was further monitored by HPLC, confirming the
stability of the MOF (ca. 2% of degradation after 4 days).
Taking into account that optimal cutaneous systems are
usually applied for a maximum of 8–24 h period,[55] the
structural stability of the obtained matrix after the SA delivery
(denoted as SA@UiO-66-NH2_del), and the evaluation of the
residual amount of SA were studied by XRPD, FTIR, TGA and
sorption measurements. FEG-SEM images denoted no signifi-
cant morphological changes among all the studied samples
(Figure S7).
The presence of the main bands of pure SA in the IR
spectrum of UiO-66-NH2@SA_del suggests the partial release of
the drug (Supporting Information, Figure S6). These results
agree with the TGA, exhibiting a residue between the SA loaded
and the empty starting materials (40.3%, corresponding to the
release of the 72% of SA). The structural integrity of the
obtained SA@UiO-66-NH2_del in water after 3 days monitored
by PXRD indicates that the crystallinity is kept. This agrees with
the very low ligand leaching in water media, as estimated by
HPLC (ca. 2%). Therefore, one can conclude that UiO-66-NH2
remains stable during the whole delivery process, making this
safe material an excellent candidate for its use as cutaneous
DDS.
The synthesis of starting MOFs was performed following similar
procedures previously reported.
UiO-66 and UiO-66-NH2 or [Zr6O4(OH)(C8H3O4X)6] nH2O, X=H or
NH2.[29,57] 1 mmol (233.03 mg) of ZrCl4 and 1 mmol of dicarboxylic
linker (terephthalic acid - H2BDC, 166.13 mg; or 2-aminoterephthalic
acid – H2BDC-NH2, 181.15 mg) were dispersed in 3 mL of N,N-
dimethylformamide (DMF), placed in a teflon-lined autoclave and
heated for 12 h at 220 C (UiO-66) or 24 h at 100 C (UiO-66-NH2).
The resulting solid was recovered by filtration and washed with
deionized water and acetone. 200 mg of the solid were suspended
in 100 mL of DMF under stirring for 12 h. Then, the DMF-washed
solid was suspended in 100 mL of methanol (MeOH) under stirring
for 12 h, recovering the activated solid by filtration. UiO-66 and
°
°
°
UiO-66-NH2 samples were evacuated for 12 h at 200 and 150 C,
respectively.
MIL-163 or [Zr2(C14H2O6N4)2]·(DMA)5(H2O)14, DMA: N,N-dimethyl-
amine.[30] In a 25 mL teflon-lined steel autoclave, 0.2 mmol (47 mg)
of ZrCl4 was added to a solution of 0.4 mmol of the 5,5’-(1,2,4,5-
tetrazine-3,6-diyl)bis(benzene-1,2,3-triol) ligand (H6TzGal, 132 mg) in
DMA (2 mL) at RT. Deionized water (3 mL) was then added under
stirring, and the autoclave sealed and placed in the oven for 24 h at
°
130 C. After cooling to RT, the solution was filtered and MIL-163
was recovered as a dark-red fine powder. The solid was then
washed in DMF (10 mL) and ethanol (EtOH, 10 mL) for one night
with each, before being filtered and dried in air.
MIL-127 or [Fe3O(OH)0.88Cl0.12(C16N2O8H6)1.5(H2O)3]·nH2O.[32] The solu-
tion obtained by adding 0.927 g of NaOH to 26 mL of propan-2-ol
in
a 250 mL round-bottom flask, was stirred until complete
dissolution of NaOH. After the addition of 10.1 mmol (3.636 g) of
3,3’,5,5’-azobenzenetetracarboxylic acid ligand (H4TazBz), the result-
°
ing solution was stirred at 90 C. In parallel, 20 mmol (5.498 g) of
FeCl3 ·6H2O were dissolved in 14 mL of propan-2-ol in a flask at
°
70 C. In the next step, a mixture obtained by mixing the iron and
the ligand solutions was stirred under reflux for 24 h. A crystalline
powder was recovered by filtration, washed with 200 mL of distilled
water, 200 mL of propan-2-ol, and finally with 200 mL of ethanol.
°
MIL-127 was evacuated for 12 h at 150 C.
Conclusion
MIL-53 or [Fe(OH)(C8O4H4)]·nH2O.[3] 0.1 mol (27 g) of FeCl3.6H2O and
0.1 mol (16.6 g) of H2BDC were dispersed in 500 mL of DMF. The
mixture was placed in a round bottom flask and refluxed for 48 h at
The encapsulation of SA in MOFs is a promising alternative to
overcome its topical administration, hindered by its reactivity/
instability. Particularly, UiO-66-NH2 combines an important SA
cargo (12.1 wt.%) and an exceptional stability, with a release
kinetics adapted to the cutaneous route (64% in 6 h), fulfilling
the requirements of a topical DDS.
°
150 C under stirring. Then, the yellow solid was recovered by
filtration and washed with DMF. Then, the solid was suspended in
500 mL of water for 12 h. MIL-53 was evacuated for 12 h at 200 C.
°
MIL-53-(OH)2 or [Fe(OH)(C8H4O6)].[34] 6.8 mmol (1.34 g) of 2,5-
dihydroxyterephthalic acid (H2BDC-(OH)2), 4.6 mmol (1.63 g) of
Fe(ClO4)3·H2O, 25 mL of DMF, 1 mL of a 5 M hydrofluoric acid
solution and 5 mL of a 5 M perchloric acid solution were placed in a
125 mL teflon-lined steel autoclave. The mixture stirred for 10
°
minutes, the autoclave sealed, and placed in an oven at 100 C for
Eur. J. Inorg. Chem. 2021, 1325–1331
www.eurjic.org
1329
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/ejic.202001134
16 h. Crude MIL-53-(OH)2 was recovered as a dark crystalline solid
by filtration, washed with ethanol and dried in air. MIL-53-(OH)2
zenetetracarboxylic acid (H4TazBz, rt=16.36 min, λ=311 nm),
H3BTC (rt=3.51 min, λ=225 nm), Hmim (rt=1.8 min, λ=205 nm),
and 5,5’-(1,2,4,5-tetrazine-3,6-diyl)bis(benzene-1,2,3-triol) (H6TzGal,
rt=3.24 min, λ=360 nm).
°
sample was evacuated for 12 h at 130 C.
MIL-100 or [Fe3O(H2O)2OH(C6H3(CO2)3)2] nH2O. First, triethyl-1,3,5-
benzenetricarboxylate (Et3BTC) was prepared by heating under
reflux 24 mmol (5 g) of H3BTC in a solution of 2 mL of H2SO4 in
100 mL of ethanol absolute and for 24 h. The obtained Et3BTC was
filtered and washed with water. MIL-100 was prepared from
hydrothermal reaction (125 mL vessel) of 10 mmol (2.7 g) of FeCl3
6H2O and 6.6 mmol (1.94 g) of Et3BTC in 50 mL of water for 3 days
Preparation of the phosphate buffered solution (0.04 M, pH=2.5):
0.02 mol (2.4 g) of NaH2PO4 and 0.02 mol (2.84 g) of Na2HPO4 were
dissolved in 1 L of Milli-Q water. The pH was then adjusted to 2.5
with H3PO4 (�85%).
Adsorption and stability studies. SA adsorption studies were
performed by suspending 30 mg of desolvated MOFs in 30 mL of
deionized water containing different amount of SA. The MOF:SA
molar ratios were 1:2 and 1:0.5, using SA concentration from 44.8
to 1369.8 mg·LÀ 1. The suspensions were stirred under bidimen-
°
at 130 C.
ZIF-8 or [Zn(C4H5N2)2] nH2O.[35] 2.5 mmol (0.744 g) of Zn(NO3)2 ·6H2O
was dissolved in 10 mL of deionized water and added to a solution
of 0.21 mmol of 2-methylimidazole (Hmim, 0.172 g) in 90 mL of
deionized water. The mixture was stirred at room temperature. The
solution quickly became cloudy and a suspension was obtained.
24-hours later, the suspension was centrifuged and washed with
methanol three times. The products were then dried for 24 h under
°
sional continuous stirring (80×80 rpm) at 25 C for 24 h. At different
incubation times, the suspensions were centrifuged (14500 rpm,
10 min), and filtered with a syringe filter (0.2 μm) to obtain clean
solutions. The obtained solids were characterized by PXRD to check
their crystallinity, while the liquid phases were analyzed by HPLC,
determining the amount of SA and the total amount of possible
MOF’s leached ligand in the solution. All the SA studies and leached
ligand determinations were performed by triplicate. Drug release.
10 mg of the drug-containing MOF (SA@UiO-66-NH2) was placed in
°
reduced pressure at 40 C.
Analysis and Characterization
°
10 mL of deionized water at 37 C under bidimensional continuous
stirring (80×80 rpm). At different incubation times, 5 mL of super-
natant was recovered by centrifugation (13000 rpm for 10 min) and
Physicochemical characterization. Fourier transform infrared (FTIR)
spectroscopic analyses were performed in a Thermo Scientific
Nicolet 6700. KBr pellets of the UiO-66-NH2 and SA@UiO-66-NH2
°
replaced with the same volume of fresh deionized water at 37 C.
°
were prepared and dried overnight at 100 C before the measure-
This procedure was performed in triplicate. The amount of released
drug and the possible leached ligand were determined by HPLC.
After the drug-delivery process, the sample was characterized by
PXRD, FTIR spectroscopy, and TGA.
ment. N2 isotherms were obtained at 77 K using Belsorp Max (Bel,
Japan). Before the measurements, UiO-66-NH2 and SA@UiO-66-NH2
°
were evacuated at 150 C for 12 h (see evacuation process for the
rest of MOFs in the Synthesis of MOF section). Specific surface area
was determined by applying Brunauer, Emmett & Teller equation
(BET) in the relative pressure interval p/p0 =0.01–0.3 (being p0 the
saturation pressure). Pore volume (Vp) and pore size distribution
were calculated by the non-localized density functional theory
(NLDFT) and the Horvath-Kawazoe (KH) methods, respectively.
Routine X-ray powder diffraction (PXRD) patterns were collected
using a conventional high resolution D5000 Siemens X’Pert MDP
diffractometer (θ–2θ) using λCu Kα1, and Kα2 radiation (λ=1.54051
Acknowledgements
S.R. and P.H. acknowledge the regional Madrid founding (Talento
2017 Modality 2, 2017-T2/IND-5149), MOFSEIDON project
(PID2019-104228RB-I00, MCI/AEI/FEDER, UE), and the Multifunc-
tional Metallodrugs in Diagnosis and Therapy Network (MICIU,
RED2018-102471-T). P.H. thanks the Spanish Ramón y Cajal
Programme (grant agreement 2014-16823). S. R. thanks the Juan
de la Cierva Incorporación Fellowship (grant agreement IJC2019-
038894-I).
°
and 1.54433 Å), from 3 to 35 or 60 (2θ), a step size of 0.02 , and
2 s·stepÀ 1 in continuous mode, and a conventional PANalytical
Empyrean powder diffractometer (PANalytical Lelyweg, Nether-
lands, θ–2θ) using the same λCu Kα1, and Kα2 radiation. Thermogravi-
metric analyses (TGA) were performed using
a Perkin-Elmer
Diamond TGA/DTA STA 6000 running form room temperature (RT)
to 600 C with a heating rate of 2 C·minÀ 1. Field emission guns-
scanning electron microscopy (FEG-SEM) images were acquired in a
FEI/Philips XL-30 Field Emission ESEM (Philips, Amsterdam, The
Netherlands) at 3 kV and 98–102 μA.
°
°
Conflict of Interest
High performance liquid chromatography (HPLC). The amount of
adsorbed SA, as well as the potential release of the corresponding
organic linker, were determined using a reversed phase HPLC Jasco
LC-4000 series system, equipped with a PDA detector MD-4015 and
a multisampler AS-4150 controlled by ChromNav software (Jasco
Inc, Japan). A Purple ODS reverse phase column (5 μm, 4.6×
150 mm, Análisis Vínicos, Spain) was employed. For the quantifica-
tion of all chemical species, isocratic conditions were used. The flow
rate was 1 mL·minÀ 1, and the column temperature was fixed at
298 K. In all cases, the injection volume was 30 μL. The mobile
phase was based on a mixture of 50:50 MeOH:phosphate buffered
solution (PBS; 0.04 M, pH=2.5) for SA and MOFs’ ligand analysis
was used with different retention time (rt) and absorption
maximum (λ) for each molecule: SA (rt=4.99 min, λ=231 nm),
H2BDC (rt=3.98 min, λ=240 nm), H2BDC-NH2 (rt=3.03 min, λ=
228 nm), H2BDC-(OH)2 (rt=2.92 min, λ=217 nm), 3,3’,5,5’-azoben-
The authors declare no conflict of interest.
Keywords: Metal-organic frameworks · Drug delivery · Salicylic
acid · Cutaneous treatment · Water stability
[1] V. P. Torchilin, Nanoparticulates as Drug Carriers, London, 2006.
[2] M. Giménez-Marqués, T. Hidalgo, C. Serre, P. Horcajada, Coord. Chem.
Rev. 2015, 307, 342–360.
[3] P. Horcajada, T. Chalati, C. Serre, B. Gillet, C. Sebrie, T. Baati, J. F. Eubank,
D. Heurtaux, P. Clayette, C. Kreuz, J.-S. Chang, Y. K. Hwang, V. Marsaud,
P.-N. Bories, L. Cynober, S. Gil, G. Férey, P. Couvreur, R. Gref, Nat. Mater.
2010, 9, 172–178.
[4] P. Horcajada, R. Gref, T. Baati, P. K. Allan, G. Maurin, P. Couvreur, G.
Férey, R. E. Morris, C. Serre, Chem. Rev. 2012, 112, 1232–1268.
[5] H. C. J. Zhou, S. Kitagawa, Chem. Soc. Rev. 2014, 43, 5415–5418.
Eur. J. Inorg. Chem. 2021, 1325–1331
www.eurjic.org
1330
© 2021 Wiley-VCH GmbH

Full Papers
doi.org/10.1002/ejic.202001134
[6] C. Doonan, R. Riccò, K. Liang, D. Bradshaw, P. Falcaro, Acc. Chem. Res.
[34] T. Devic, P. Horcajada, C. Serre, F. Salles, G. Maurin, D. Heurtaux, G. Clet,
A. Vimont, J. Grene, B. Le Ouay, F. Moreau, E. Magnier, Y. Filinchuk, J.
Lavalley, M. Daturi, I. Charles, G. Montpellier, J. Am. Chem. Soc. 2010,
132, 1127–1136.
[35] V. A. Online, K. Kida, M. Okita, K. Fujita, S. Tanaka, Y. Miyake,
CrystEngComm 2013, 15, 1794–1801.
2017, 50, 1423–1432.
[7] C. Orellana-Tavra, R. J. Marshall, E. F. Baxter, I. A. Lázaro, A. Tao, A. K.
Cheetham, R. S. Forgan, D. Fairen-Jimenez, J. Mater. Chem. B 2016,
7697–7707.
[8] P. Horcajada, C. Serre, M. Vallet-Regí, M. Sebban, F. Taulelle, G. Férey,
Angew. Chem. Int. Ed. 2006, 45, 5974–5978; Angew. Chem. 2006, 118,
6120–6124.
[36] E. Ghasemi, A. Mirhabibi, M. Edrissi, R. Aghababazadeh, R. M. Brydson, J.
Nanosci. Nanotechnol. 2009, 9, 4273–4278.
[9] A. G. Márquez, T. Hidalgo, H. Lana, D. Cunha, M. J. Blanco-Prieto, C.
Álvarez-Lorenzo, C. Boissière, C. Sánchez, C. Serre, P. Horcajada, J. Mater.
Chem. B 2016, 4, 7031–7040.
[10] S. Rojas, P. S. Wheatley, E. Quartapelle-Procopio, B. Gil, B. Marszalek, R. E.
Morris, E. Barea, CrystEngComm 2013, 15, 9364–9367.
[11] A. C. Mckinlay, J. F. Eubank, S. Wuttke, B. Xiao, P. S. Wheatley, P. Bazin,
M. Daturi, A. Vimont, G. De Weireld, P. Horcajada, C. Serre, R. E. Morris,
Chem. Mater. 2013, 88, 1592–1599.
[12] F. J. Carmona, S. Rojas, C. C. Romão, J. A. R. Navarro, E. Barea, C. R.
Maldonado, Chem. Commun. 2017, 53, 6581–6584.
[13] Markets&Markets, Topical Drug Delivery Market, 2020.
[14] N. Kanikkannan, S. Kandimalla, S. S. Lamba, M. Singh, Curr. Med. Chem.
2000, 7, 593–608.
[37] S. Rojas, I. Colinet, D. Cunha, T. Hidalgo, F. Salles, C. Serre, N. Guillou, P.
Horcajada, ACS Omega 2018, 3, 2994–3003.
[38] C. Hansch, A. Leo, D. Hoekman, J. Med. Chem. 1996, 39, 1189–1190.
[39] L. H. Xie, M. M. Xu, X. M. Liu, M. J. Zhao, J. R. Li, Adv. Sci. 2020, 7,
1901758–1901802.
[40] P. G. M. Mileo, Computational Exploration of Water Adsorption and
Proton Conduction in Porous Materials, Université Montpellier, 2019.
[41] J. Canivet, A. Fateeva, Y. Guo, B. Coasne, D. Farrusseng, Chem. Soc. Rev.
2014, 43, 5594–5617.
[42] J. Huang, X. Jin, J. Mao, B. Yuan, R. Deng, S. Deng, J. Hazard. Mater.
2012, 217–218, 406–415.
[43] “Anti-Acne Treatment Cream with 2% Salicylic Acid (Formulation #O-
0057),” can be found under https://www.ulprospector.com/es/eu/
PersonalCare/Detail/77/238019/Anti-Acne-Treatment-Cream-with-2-Sali-
cylic-Acid-Formulation-O-0057, 2020.
[15] J. R. Vane, R. M. Botting, Am. J. Med. 1998, 104, 2S–8S.
[16] J. R. Paterson, J. R. Lawrence, QJM – Mon. J. Assoc. Physicians 2001, 94,
445–448.
[44] A. Klimowicz, S. Farfał, S. Bielecka-Grzela, J. Clin. Pharm. Ther. 2007, 32,
[17] C. L. Campbell, Jama 2007, 297, 1–7.
143–148.
[18] R. Amann, B. A. Peskar, Eur. J. Pharmacol. 2002, 447, 1–9.
[19] S. D. Taherzade, S. Rojas, J. Soleimannejad, P. Horcajada, Nanomaterials
2020, 10, 2296–2316.
[20] S. Rojas, T. Baati, L. Njim, L. Manchego, F. Neffati, N. Abdejelil, S.
Saguem, C. Serre, M. F. M. F. Najjar, A. Zakhama, P. Horcajada, N.
Abdeljelil, S. Saguem, C. Serre, M. F. M. F. Najjar, A. Zakhama, P.
Horcajada, J. Am. Chem. Soc. 2018, 140, 9581–9586.
[45] T. N. da Silva, F. Reynaud, P. H. de S. Picciani, K. G. de Holanda, E. Silva,
T. N. Barradas, Int. J. Biol. Macromol. 2020, 164, 2558–2568.
[46] J. H. Mottaz, P. J. Mckeever, J. L. Neveaux, A. S. Zelickson, Int. J. Dermatol.
1988, 27, 596–600.
[47] J. C. Tsai, S. A. Chuang, M. Y. Hsu, H. M. Sheu, Int. J. Pharm. 1999, 188,
145–153.
[48] B. C. Sekhar, Topical Composition for Delivery of Salicylates, 2007.
[49] S. Ueda, K. Mitsugi, K. Ichige, K. Yoshida, T. Sakuma, S. I. Ninomiya, T.
Sudou, J. Dermatol. Sci. 2002, 28, 211–218.
[21] S. Rojas, N. Guillou, P. Horcajada, ACS Appl. Mater. Interfaces 2019, 11,
22188–22193.
[22] “http://jp-minerals.org/vesta/en/,” 2020.
[50] K. A. Mocniak, I. Kubajewska, D. E. M. Spillane, G. R. Williams, R. E. Morris,
RSC Adv. 2015, 5, 83648–83656.
[51] S. Tai, W. Zhang, J. Zhang, G. Luo, Y. Jia, M. Deng, Y. Ling, Microporous
Mesoporous Mater. 2016, 220, 148–154.
[23] T. Baati, L. Njim, F. Neffati, A. Kerkeni, M. Bouttemi, R. Gref, M. F. Najjar,
A. Zakhama, P. Couvreur, C. Serre, P. Horcajada, Chem. Sci. 2013, 4,
1597–1607.
[24] À. Ruyra, A. Yazdi, J. Espín, A. Carné-Sánchez, N. Roher, J. Lorenzo, I.
Imaz, D. Maspoch, Chem. A Eur. J. 2015, 21, 2508–2518.
[25] A. Ruyra, A. Yazdi, J. Espín, A. Carné-Sánchez, N. Roher, J. Lorenzo, I.
Imaz, D. Maspoch, Chem. A Eur. J. 2015, 21, 2508–2518.
[26] M. Hoop, C. F. Walde, R. Riccò, F. Mushtaq, A. Terzopoulou, X. Z. Chen,
A. J. de Mello, C. J. Doonan, P. Falcaro, B. J. Nelson, J. Puigmartí-Luis, S.
Pané, Appl. Mater. Res. 2018, 11, 13–21.
[27] G. Chen, X. Leng, J. Luo, L. You, C. Qu, X. Dong, H. Huang, X. Yin, J. Ni,
Molecules 2019, 24, DOI 10.3390/molecules24071211.
[28] P. Ghosh, Y. J. Colón, R. Q. Snurr, Chem. Commun. 2014, 50, 11329–
11331.
[52] R. Zhang, Z. Wang, Z. Zhou, D. Li, T. Wang, P. Su, Y. Yang, Ind. Eng.
Chem. Res. 2019, 58, 3876–3884.
[53] L. Ghezzi, A. Spepi, M. Agnolucci, C. Cristani, M. Giovannetti, M. R. Tiné,
C. Duce, Appl. Clay Sci. 2018, 160, 88–94.
[54] N. Vilaça, F. Morais-Santos, A. F. Machado, A. SirkecioWlu, M. F. R. Pereira,
M. Sardo, J. Rocha, P. Parpot, A. M. Fonseca, F. Baltazar, I. C. Neves, J.
Phys. Chem. C 2015, 119, 3589–3595.
[55] K. S. Paudel, M. Milewski, C. L. Swadley, N. K. Brogden, P. Ghosh, A. L.
Stinchcomb, Ther. Delivery 2010, 1, 109–131.
[56] A. Dhakshinamoorthy, M. Alvaro, H. Chevreau, P. Horcajada, T. Devic, H.
Garcia, Catal. Sci. Technol. 2012, 3, 324–330.
[29] J. H. Cavka, S. Jakobsen, U. Olsbye, N. Guillou, C. Lamberti, S. Bordiga,
K. P. Lillerud, J. Am. Chem. Soc. 2008, 130, 13850–13851.
[30] G. Mouchaham, L. Cooper, N. Guillou, C. Martineau, E. Elkaïm, S.
Bourrelly, P. L. Llewellyn, C. Allain, G. Clavier, C. Serre, T. Devic, Angew.
Chem. Int. Ed. Engl. 2015, 54, 13297–13301.
[57] M. Kandiah, M. H. Nilsen, S. Usseglio, S. Jakobsen, U. Olsbye, M. Tilset, C.
Larabi, E. A. Quadrelli, F. Bonino, K. P. Lillerud, D. Lyon, Chem. Mater.
2010, 10, 6632–6640.
[31] Y. Liu, J. F. Eubank, A. J. Cairns, J. Eckert, V. C. Kravtsov, R. Luebke, M.
Eddaoudi, Angew. Chem. Int. Ed. Engl. 2007, 46, 3278–3283.
[32] H. Chevreau, A. Permyakova, F. Nouar, P. Fabry, C. Livage, F. Ragon, A.
Garcia-Marquez, T. Devic, N. Steunou, C. Serre, P. Horcajada, CrystEng-
Comm 2016, 18, 4094–4101.
[33] P. Horcajada, S. Surblé, C. Serre, D.-Y. Hong, Y.-K. Seo, J.-S. Chang, J.-M.
Grenèche, I. Margiolaki, G. Férey, Chem. Commun. 2007, 100, 2820–
2822.
Manuscript received: December 18, 2020
Revised manuscript received: February 25, 2021
Accepted manuscript online: March 1, 2021
Eur. J. Inorg. Chem. 2021, 1325–1331
www.eurjic.org
1331
© 2021 Wiley-VCH GmbH