Chemical Engineering Journal 425 (2021) 130498

Contents lists available at ScienceDirect

Chemical Engineering Journal

journal homepage: www.elsevier.com/locate/cej

Nano-in-Nano dendrimer gel particles for efficient topical delivery of

antiglaucoma drugs into the eye
Juan Wang a, Boxuan Li b, Da Huang c, Pedro Norat d, Marta Grannonico d, Remy C. Cooper e,
Qin Gui f, Woon Nam Chow g, h, Xiaorong Liu d, i, *, Hu Yang c, *
a
College of Biomedical Engineering, Sichuan University, Chengdu, Sichuan 610065, China
b
Tianjin Key Laboratory of Biomedical Materials, Key Laboratory of Biomaterials and Nanotechnology for Cancer Immunotherapy, Institute of Biomedical Engineering,
Chinese Academy of Medical Sciences & Peking Union Medical College, Tianjin 300192, China
c
Linda and Bipin Doshi Department of Chemical and Biochemical Engineering, Missouri University of Science and Technology, Rolla, MO 65409, United States
d
Department of Biology, University of Virginia, Charlottesville, VA 22904, United States
e
Department of Biomedical Engineering, Virginia Commonwealth University, Richmond, VA 23298, United States
f
Department of Chemical and Life Science Engineering, Virginia Commonwealth University, Richmond, VA 23219, United States
g
Department of Pathology, Virginia Commonwealth University, Richmond, VA 23298, United States
h
Department of Ophthalmology, Virginia Commonwealth University, Richmond, VA 23298, United States
i
Department of Psychology, University of Virginia, Charlottesville, VA 22904, United States

A R T I C L E I N F O A B S T R A C T

Keywords: Low bioavailability of topically applied drugs remains a significant challenge for long-term glaucoma therapy. To
Dendrimer enhance drug delivery efficiency, we developed dendrimer gel particles that collectively exhibit structural
Microgel benefits of dendrimer, hydrogel, and particles, using the inverse emulsion method coupled with the highly
Nanogel
efficient aza-Michael addition reaction (IEaMA). This hierarchical approach would maximize the utility of the
Brimonidine
structural features of existing ocular drug delivery systems. We have tested the delivery efficiency and efficacy of
Glaucoma
two first-line antiglaucoma drugs, brimonidine tartrate (BT) and timolol maleate (TM), which were loaded into
dendrimer gel particles of various sizes, i.e., nDHP (nano-in-nano dendrimer hydrogel particles, ~200 nm),
μDHP3 (3 μm), and μDHP10 (9 μm). We found that nDHP was superior to μDHP3 and μDHP10 in terms of
cytocompatibility, degradability, drug release kinetics, and corneal permeability. The nDHPs increased drug
corneal permeability by 17-fold compared to plain drug solution and enabled zero-order prolonged drug release
kinetics. The nDHP-based formulation demonstrated pronounced IOP-lowering effects in both single-dose test
and 7-day chronic daily dosing test in both Brown Norway rats and glaucoma mice. Taken together, we have
developed nano-in-nano dendrimer gel particles for precise dosing and enabling sustained and synergistic effi­
cacy of antiglaucoma drugs, which could be clinically impactful for improving glaucoma treatment.

1. Introduction hypotensive medication further raises the concerns of delivery effec­

tiveness and patient adherence to the therapy, a major obstacle for
Glaucoma remains a leading cause of blindness worldwide. The successful glaucoma management [7–10].
number of glaucoma patients will rise to 111.8 million in 2040 [1]. Using particles ranging in size from nanometers to microns to
Elevated intraocular pressure (IOP) is a characteristic risk factor of improve bioavailability and sustain the efficacy of antiglaucoma drugs is
glaucoma. All current treatments are to lower or control IOP and thereby on the horizon [11–17], but there are two main challenges. First, size,
slow down or reduce the subsequent vision loss [2]. Treatment typically structure, and physicochemical properties of particles, such as surface
begins with topical antiglaucoma medications. Topically administered charges, composition, and mucin adhesiveness [18], are essential to
drugs, however, have bioavailability of<5% due to high pre-corneal loss tune particles to overcome ocular barriers to gain enhanced preocular
and low cornea permeation [3–6]. Frequent daily dosing of ocular retention, corneal permeation, and absorption. The challenge is how to

* Corresponding authors at: Linda and Bipin Doshi Department of Chemical and Biochemical Engineering, Missouri University of Science and Technology, Rolla,
Missouri 65409, United States.
E-mail addresses: [email protected] (X. Liu), [email protected] (H. Yang).

https://doi.org/10.1016/j.cej.2021.130498
Received 2 February 2021; Received in revised form 18 May 2021; Accepted 21 May 2021
Available online 28 May 2021
1385-8947/© 2021 Elsevier B.V. All rights reserved.
J. Wang et al. Chemical Engineering Journal 425 (2021) 130498

build a particulate system that synchronizes particle transport to the Freeze Biologicals (Rogers, AR). Cell culture medium for primary
anterior chamber of the eye with drug delivery and release kinetics. human corneal epithelial cells (HCECs) was purchased from American
Secondly, when first-line monotherapy for glaucoma becomes ineffec­ Type Culture Collection (ATCC, Manassas, VA).
tive, two or three drugs are often combined to produce additional IOP
reduction. Those drugs in the free drug form in ophthalmic solution are 2.2. Characterization
unlikely to reach the eye at the ratio prescribed due to the variance in
physicochemical properties and corneal permeation, bringing about A field emission SEM microscope (Hitachi FE-SEM Su-70, Japan)
only suboptimal therapeutic outcomes. Features such as high loading with an acceleration voltage of 5 kV was used to image µDHPs, and JEM-
capacity and versatility that enables accommodating drugs of distinct 1400 Plus TEM for nDHPs. The hydrodynamic sizes and zeta potentials
pharmacological and physicochemical properties are thus desired. of DHPs were measured on Zetasizer Nano-ZS90 (Malvern Instruments,
We attempted to address the above two issues by using polyamido­ U.K.). Antiglaucoma drug released from DHPs was quantified by using a
amine (PAMAM) dendrimers as building blocks [19,20]. PAMAM den­ Waters HPLC system equipped with a dual absorbance UV detector. The
drimers are highly branched nanostructured synthetic polymers [21], mobile phase was the mixture of acetonitrile, water, and trifluoroacetic
ideal for drug and gene delivery, primarily due to their multivalency, acid (1/1/0.05 v/v/v). Eluents were monitored at both 309 and 220 nm.
enabling multi-functionalization and electrostatic interactions [22].
Furthermore, PAMAM dendrimers exhibit unimolecular micelle encap­ 2.3. Preparation of DHPs
sulation behavior [23]. Hydrophobic drugs can be encapsulated into
dendrimer interior cavities via host–guest interactions. However, such DHPs were prepared by using the IEaMA method developed by us
host–guest interactions may be disrupted easily under physiological [24]. The oil phase was a mixture of hexane and surfactant (Table S1).
conditions, causing premature drug leakage and uncontrolled release, a The aqueous solution was prepared by mixing G5 (10 wt%) and PEG-DA
major limitation for dendrimer-based drug encapsulation and delivery. (the molar ratio of amine/acrylate = 1/1) in deionized water followed
To expand the capacity of dendrimers for drug encapsulation and by vortex for 10 s. The oil-to-water volume ratio (O/W) was tuned to
delivery, we developed a new method, i.e., the inverse emulsion aza- obtain different micro-nano particle sizes (Table S1). For µDHP prepa­
Michael addition reaction (IEaMA) method, to make dendrimer hydro­ ration, the aqueous solution was poured into the oil phase and dispersed
gel particles (DHPs) [24]. DHPs integrate the advantages of dendrimer, at 30,000 rpm for 60 s by using an IKA disperser. The emulsion system
hydrogel and nano/micro-particle. Therefore, DHPs hold great promise was then stirred at 500 rpm for 2 h, allowing the cross-linking within the
to be used to deliver multiple antiglaucoma drugs for development of the stabilized emulsion droplets to form DHPs. For nDHP preparation, the
next generation of dendrimer-based antiglaucoma drug delivery sys­ aqueous solution was added to the oil phase and directly stirred at 500
tems. DHPs are highly swellable in physiological solution and appear as rpm for 2 h by using an IKA stirrer. The mixture was then centrifuged
transparent colloidal dispersions, eliminating themselves as a source for (4,900 rpm for µDHPs and 10,000 rpm for nDHPs) for 3 min, and the oil
blurred vision or eye irritation [25]. In this work, we demonstrated that layer was discarded. The residue was washed with deionized water 3
DHPs in the nanometer to micrometer range can be made by tuning times to obtain surfactant-free DHPs.
parameters of the IEaMA-based process: oil-to-water ratio, surfactant To prepare drug-loaded µDHPs and nDHPs, drugs were mixed with
concentration, and stirring speed. The two first-line IOP-lowering drugs, G5 aqueous solution (10 wt%) prior to mixing with PEG-DA. Specif­
brimonidine tartrate (BT) and timolol maleate (TM), were used as model ically, 1 mg of BT or TM was added to 53 μL of 10 wt% of G5 aqueous
drugs and packaged into DHPs. The delivery efficiency and safety of solution, followed by vortex and sonication to make a clear solution.
DHPs with three discrete sizes, i.e., nano-in-nano DHP (nDHP) on a scale Drug loading was measured by HPLC. On the basis of 1 mg particle,
of hundreds nanometers (~200 nm) and two micronized DHPs—μDHP3 µDHP10 carries 272.2 µg of BT, µDHP3 carries 215.1 µg of BT, and nDHP
(3 μm), and μDHP10 (9 μm)—were assessed. We found that nDHP was contains 8.6 µg of BT.
superior to μDHP3 and μDHP10 in terms of cytocompatibility, degrad­
ability, drug release kinetics, and corneal permeability. In vivo tests 2.4. Degradation studies
using both normotensive rats and glaucoma mice showed that nDHP-
based formulation induced more pronounced IOP-lowering effects in DHPs in dry form (w0) were dispersed in contrived tears (0.5 mL) and
both single-dose test and 7-day once-daily dosing test. incubated at 37 ℃ for 6 h, 24 h, and 48 h, respectively. At the end of
incubation, dispersions were subjected to centrifugation (10,000 rpm, 3
2. Materials and methods min). The pellet was washed once with deionized water and then freeze-
dried and weighed (wt). Mass remaining (%) was determined as wt/w0 ×
2.1. Materials 100. The degraded µDHPs were imaged by using SEM [24].

DAB-core PAMAM dendrimer generation 5 (G5) was purchased from 2.5. Drug release studies
NanoSynthons (Mt Pleasant, MI). Poly(ethylene glycol) diacrylate (PEG-
DA, Mn = 575 g/mol), span80, tween80, fluorescein isothiocyanate In general, BT-loaded DHPs were dispersed in 1 mL of pH 7.4 PBS
isomer I (FITC), trifluoroacetic acid (TFA), and cell proliferation reagent and kept at 37 ℃. At predetermined time points, the suspension was
WST-1 were purchased from Sigma-Aldrich (St. Louis, MO). Brimoni­ subjected to centrifugation (10,000 rpm, 3 min). The supernatant was
dine tartrate (BT) was purchased from AvaChem Scientific (San Antonio, collected for BT concentration analysis. The pellet was re-suspended in
TX). Timolol maleate (TM), hexane, acetonitrile (ACN), and phosphate 1 mL of fresh PBS to continue the release study. Each formulation was
buffered saline (PBS, 10×) were purchased from Fisher Scientific tested three times under identical conditions. Cumulative BT release as a
(Pittsburgh, PA). Triton X-100 was purchased from Solarbio Life Sci­ function of time was reported and release kinetics models were used to
ences (BJ, CN). Dulbecco’s modified Eagle’s medium (DMEM), trypsin- study release mechanisms [26].
EDTA (0.25%), and penicillin–streptomycin (10,000 U/mL) were pur­
chased from Life Technologies (Carlsbad, CA). ELISA kits of human TNF- 2.6. Ex vivo permeation studies
α and IL6 were purchased from Thermo Fisher Scientific (Waltham, MA).
Contrived tears developed from greater than 15 components known to Cornea extracted from fresh rabbit eye was mounted in a Franz
be present in tears, balancing protein, salt and pH alike, in line with diffusion cell system [27]. The receiver chamber was filled with 5 mL of
accepted tear formulations, was purchased from Ursa BioScience pH 7.4 PBS. BT-loaded DHPs suspended in 250 µL of PBS was loaded to
(Maryland, USA). Fresh rabbit whole eyes were purchased from Pel- the donor chamber. At pre-determined time points up to 4 h, an aliquot

2
J. Wang et al. Chemical Engineering Journal 425 (2021) 130498

of 500 µL was withdrawn from the receiver chamber and analyzed with lyophilized followed by ultraviolet sterilization for 0.5 h and then
HPLC. Fresh PBS (500 µL) was added to the receiver chamber following dispersed in sterile PBS before administrated. The right eye of each rat
each sampling. The permeability coefficient, P, was then determined was dosed while the left eye was left untreated. The IOPs of both eyes
[28]. Each experiment was repeated three times. were measured immediately before treatment (0 h) and at 0.75, 1.6,
4.25, and 24 h postdosing using ICare TONOLAB tonometer.
2.7. Cytotoxicity assessment
2.9.2. One week of daily dosing assessment
NIH3T3 fibroblast cells or HCECs were seeded at a density of 1 × 104 Ten rats were randomly grouped to two (n = 5) and their right eyes
cells/well in a 96-well plate and cultured overnight. The ATCC medium received BT/PBS or BT/nDHPFITC (2 × 5 μL, 0.1% w/v BT equivalent)
for HCEC supplemented with Corneal Epithelial Cell Growth Kit com­ topically at 8:00 am each day for 7 days. The IOPs of both the right eye
ponents (apo-transferrin, epinephrine, extract P, hydrocortisone hemi­ and the left eye were measured immediately before and after adminis­
succinate, L-glutamine, rh insulin, and CE growth factor) was prepared tration and 4 h later. At the end of experiment, rats were euthanized, and
following the manufacturer’s instructions. The cells were then incubated eyeballs were removed and processed to obtain tissue slices and H&E
with DHPs at various concentrations (0–1.25 mg/mL) for 48 h. DHPs stained slices. Tissue images were taken under Nikon inverted micro­
were freshly prepared, lyophilized followed by ultraviolet sterilization scope ECLIPSE Ti-U and analyzed using ImageJ.
for 0.5 h and then dispersed in sterile PBS before incubation. Given that
(1) filtration sterilization is recommended for future clinical usage and 2.10. In vivo IOP evaluation in a genetic glaucoma model
(2) nDHP is highly squeezable and its hydrodynamic diameter is 209
nm, filtration through 0.2-µm pore size is another possible sterilization Angiopoietin 1 gene knockout mice (A1 cKO mice, seven months old,
method. Cell viability relative to untreated cells was then determined by n = 4) with elevated IOP were used as a glaucoma disease model [31]. In
using the WST-1 proliferation assay [29]. The IC50 of nDHP on HCEC each group, the right eyes were topically treated with BT/nDHP (4 μL,
was calculated following a four-parameter logistic regression (4PL) 0.15% w/v BT equivalent) and TM/nDHP (4 μL, 0.5% w/v TM equiva­
fitting. lent), while the left eyes were treated with a control solution composed
Pro-inflammatory cytokines TNF-α and IL6 released from HCECs of free non-formulated drugs (BT and TM) at the same dose. The IOP was
were quantified. Briefly, HCECs were seeded in 24-well plate (1 × 105 monitored at pre-determined time intervals. The animal protocol 4173
cells/well) and cultured 24 h. The medium of each well was replaced was approved by the institutional animal care and use committees of
with fresh medium containing nDHP at different concentrations (50, University of Virginia.
87.5, or 125 μg/mL). Following 4 h incubation, the culture media were
collected and centrifuged. The supernatants were subjected to ELISA 2.11. Statistical analysis
analysis following the manufacturer’s instructions.
The data were expressed as means ± standard deviation (SD) and
2.8. Hemolysis toxicity were analyzed using one-way analysis of variance (ANOVA). A value of
p < 0.05 was considered statistically significant. The adjusted means of
Hemolysis assay was performed to evaluate the membrane damaging IOP reduction (△IOP) at 12 PM each day following one week of daily
effect of DHPs. Whole rabbit red blood (2 mL) were centrifuged at 3500 dosing on rats is derived using an analysis of covariance adjusted for
rpm for 10 min. The top two layers (plasma, leukocytes, and platelets) baseline IOP and based on observed data for all randomized subjects.
were discarded, and the pellet (erythrocytes) was collected and washed
with sterile PBS (pH 7.4) to ensure no cell lysis. Erythrocytes were then 3. Results
re-suspended with 10 mL of sterile PBS (pH 7.4). To 750 µL of eryth­
rocyte suspension, 50 µL of µDHP10, µDHP3, nDHP, G5, PEG-DA (0.5% 3.1. Size-tunable dendrimer gel particles
w/v), PBS (negative control) and Triton X-100 (1% w/v, positive con­
trol) were added respectively and incubated for 1 h, 2 h, and 3 h under The inverse emulsion method was adopted for producing the micro/
mild shaking at 37 ℃. For µDHP10, µDHP3, nDHP, and G5, a concen­ nanodroplets (Scheme 1). Multiple parameters such as surfactant
tration of 0.5% w/v G5 equivalent was used. At pre-determined time­ composition (tween80/span80 ratio), surfactant-to-hexane ratio, O/W
points, samples were centrifuged and 100 µL of supernatant was ratio, and dispersion intensity were used to manipulate gel particle size
withdrawn. The absorbance of each supernatant sample was read by (Table S1). We prepared dendrimer gel particles smaller than 10 µm as
using a microplate reader (VERSAmax, Molecular Devices) at 450 nm larger particles tend to induce foreign body sensation and reflex tearing
wavelength. Hemolysis percentage was then calculated using the [6]. μDHP10 (Fig. 1a) and μDHP3 (Fig. 1b) are spherical. SEM image
equation below [30]. analysis revealed that their mean diameters are 8.7 ± 0.2 μm and 3.3 ±
0.4 μm, respectively. This method also generates uniform nanogels (78
Hemolysis(%) =
Absorbance of sample − Absorbance of PBS ± 19 nm) (Fig. 1c). Dendrimer micro/nanogels are micro/nanosized
(Absorbance of TritonX − 100) − Absorbance of PBS swellable networks in solution and the gel particles can be dispersed into
PBS to form colloidal suspension. The swollen dendrimer gel particles
2.9. In vivo IOP measurements in normotensive glaucoma model experienced significant size expansion. In particular, the diameter of
μDHP3 increased to 4.0 μm (PDI = 0.299), and nDHP has a diameter of
Normotensive adult Brown Norway female rats (Charles River, Wil­ 209.6 nm (PDI = 0.284) (Fig. 1d). The hydrodynamic size of swollen
mington, MA), 5 months old, were used and housed under proper con­ μDHP10 exceeded the detection limit (10 μm diameter) of the DLS in­
ditions at Virginia Commonwealth University (VCU). The rats were kept strument. Our SEM/TEM images and DLS results showed the typical
under a cycle of 12-h light and 12-h dark. The animal protocol micro/nanogel structures, morphologies and size features, consistent
AD10001426 was approved by the institutional animal care and use with the characterization of micro/nanogels reported in literatures. The
committees of VCU. positively charged surfaces of all the three particles were confirmed by
their ζ potential values of 29.0 ± 3.1 mV, 22.1 ± 2.0 mV, and 12.9 ± 0.3
2.9.1. Single-dose assessment mV for μDHP10, μDHP3, and nDHP, respectively, which are signifi­
Eight rats were randomly divided into two groups (n = 4), and they cantly different with the ζ potential value of unmodified G5 (50.4 ± 0.1
received topical treatment of BT/PBS or BT/nDHP (2 × 5 μL, 0.1% w/v mV, the P values of μDHP10, μDHP3, nDHP vs G5 are 1.3 × 10-4, 7.5 ×
BT equivalent). nDHPs for in vivo studies are freshly prepared, 10-6, 2.2 × 10-9, respectively, Fig. 1e).

3
J. Wang et al. Chemical Engineering Journal 425 (2021) 130498

Scheme 1. Schematic illustration of the preparation of dendrimer gel particles of various sizes by using the invert emulsion method coupled with aza-Michael
addition reaction and tuning three parameters oil-to-water ratio, surfactant concentration, and stirring speed.

3.2. In vitro degradation and safety assessment kinetics, a release mechanism that could control antiglaucoma drug
release for sustained efficacy.
The prepared dendrimer gel particles (μDHP10, μDHP3, and nDHP) nDHP brings more BT across the cornea. Within 2 h, nDHP enabled
were further assessed in terms of cytotoxicity, degradability, drug 4.1 µg of BT to permeate through the cornea while μDHP10, μDHP3
release and corneal permeability. nDHP maintains much higher cyto­ resulted in cumulative amounts of 2.4 µg and 3.5 µg of BT across the
compatibility than μDHP10 and μDHP3. When particle concentration is cornea. Furthermore, permeability coefficients for BT mediated by
at or lower than 50 μg/mL (equivalent to 20 μg/mL of G5), none of them μDHP10, μDHP3, and nDHP are 3.5 × 10− 7 cm/s, 1.5 × 10− 6 cm/s, and
cause apparent cytotoxicity to NIH3T3 cells (Fig. 2a). As the concen­ 7.6 × 10− 6 cm/s, respectively (Fig. 2d).
tration increases to 1.25 mg/mL (equivalent to 500 μg/mL of G5), The permeability was further tested by using FITC-labeled microgel
μDHP10 and μDHP3 cause 33% of cell death, 2.8-fold higher than cell and DiR-labeled microgel, respectively. As shown in Fig. S3a, no FITC
death induced by nDHP. was detected permeated across the cornea in the first 2 h. At 4 h, there
We studied the in vitro degradation kinetics of dendrimer gel parti­ are 62% of FITC crossed the cornea. The DiR signals (Fig. S3b) in the eye
cles incubated in contrived tears. During the first 6 h-incubation, indicate gel particles permeated through the cornea and transported in
µDHP10, µDHP3, and nDHP were relatively stable and experienced Schlemm’s canal.
negligible mass loss (Fig. 2b). When incubation was extended to 24 h,
more than 70% of µDHP10 degraded. <17% of µDHP10 remained after
3.4. nDHP sustains in vivo IOP lowering effects
48-h incubation. µDHP3 degraded more slowly. Approximately 19% of
its mass remained after 48-h incubation. Moreover, more than 40% of
The IC50 of nDHP on HCEC is 0.42 mg/mL. A dose of 0.25 mg/mL of
nDHPs remained intact at 48 h post-incubation. µDHP10 and µDHP3
nDHP or lower is deemed safe (Fig. 3a). We assessed the production
maintained their particle morphology but became porous and flattened
levels of proinflammatory cytokines, TNF-α and IL6 in HCECs in
over the observation period of 48 h (Fig. S1). The hemolytic toxicity of
response to different doses of nDHPs (Fig. 3b). Unmodified PAMAM
DHPs was assessed along with G5 and PEG-DA. As shown in Fig. S2,
dendrimers show proinflammatory activities in vivo [32]. It is reported
none of the tested groups has shown substantial hemolytic behavior,
that 0.8 μM of G5 could induce 200 pg/mL of TNF-α secretion, and 400
indicating their good blood compatibility.
pg/mL of IL6 secretion in J774A.1 macrophage cells [33]. At the tested
doses, nDHPs did not stimulate TNF-α production. nDHP at doses as high
3.3. nDHP enables zero-order drug release kinetics and superb corneal as 125 μg/mL (equivalent to 3.5 μM of G5) caused 279 pg/mL of IL6
permeability production, 35% higher compared to the baseline (the P values of
μDHP10, μDHP3, nDHP vs Untreated are 0.0022, 0.0022, 0.0019,
The release of BT from µDHP10 and µDHP3 was quick. Nearly 100% respectively). The TNF-α and IL-6 levels are 90% and 15% higher in
of the encapsulated BT molecules were released within 2 h via Fickian patients receiving preserved latanoprost than in normal controls [34].
diffusion (Fig. 2c). Within the first 6 h, 74% of BT was released from The IOP-lowering effect of a single dose of BT/nDHP was examined
nDHP. Furthermore, BT release from nDHP followed zero-order release in normotensive rats and compared with the BT/PBS control group. Both

4
J. Wang et al. Chemical Engineering Journal 425 (2021) 130498

Fig. 1. Morphology, size and zeta potential characterizations of micro/nano dendrimer hydrogel particles. a) SEM image of μDHP10, b) SEM image of μDHP3, c)
TEM image of nDHP, d) size distributions, and e) zeta potentials, *** P < 0.001 vs G5.

formulations started to reduce IOP immediately after administration and drugs. IOP reduction is observed in both eyes treated with free drugs and
exhibited peak effects at approximately 4 h. But BT/nDHP causes an formulated drugs at 1 h and 24 h post-treatment (n = 6). The nDHP
average of 4.5 mmHg IOP reduction, 2.6-fold stronger than the IOP treated groups exhibited an IOP value of 19.65 ± 1.94 mmHg at 1 h post
reduction by BT/PBS eye drops (P = 0.011, Fig. S4). The IOP-lowering treatment, which was significantly lower than that of the free group
effects of both formulations diminished afterward over the course of (22.28 ± 0.96 mmHg), with a P value of 0.026. The IOP lowering effect
24 h. Once-daily dosing of BT/nDHP resulted in cumulative IOP- at 24 h of the formulated group (18.68 ± 1.35 mmHg) was also signif­
lowering effects (Fig. 4). BT/nDHP exhibits a sustained prominent icant than the free drug group (21.15 ± 2.10 mmHg, P = 0.047, Fig. 6b).
IOP-lowering effect following 7-day once-daily dosing regimen. The IOP Together, the formulated drug exhibits potent and long-lasting effects on
lowering efficacy treated by BT/nDHP at day 7 was 4-fold higher than controlling IOP.
that treated by BT/PBS (P = 0.014, Fig. 4a). According to the covariance
analysis that was performed to eliminate the effect of initial IOP 4. Discussion
(baseline) on IOP reduction, the adjusted means of IOP reduction
following BT/nDHP treatment are higher than that of BT/PBS every day For a macroemulsion, an emulsion usually having micron-sized
(Fig. 4b), over 6-fold higher potency than BT/PBS in terms of IOP droplets, high energy input is required [35,36]. The energy input is
reduction on day 1-3. Sustained retention of FITC-labeled nDHP in the generally from mechanical devices, including disperser, homogenizer,
cornea was confirmed at the end of treatment (Fig. S5). BT/nDHP does and ultrasound generator [35]. We prepared μDHPs by using high-
not cause any irritation, including abnormal tearing and blinking shearing dispersion (30,000 rpm) generated by IKA disperser (T10
(Fig. 5a). Whole-mount histologic globe preparations show that BT/ basic ULTRA-TURRAX). We continued our previous formula of
nDHP does not cause any discernable architectural or pathologic tween80/span80 1/5 (w/w) and surfactant-to-hexane 1/70 (v/v) [24],
changes to ocular tissue elements (Fig. 5b). Furthermore, the structural but only adjusted the O/W ratio from 35/1 (v/v) to 10/1 (v/v) to finally
integrity of the various intraocular compartments, including the anterior get microgels with different sizes: 3 µm and 9 µm on average. Unlike the
and posterior chambers, remains intact (Fig. 5c). macroemulsion, a nanoemulsion with size ranges from 20 to 200 nm
The IOP-lowering effects of nDHP-based antiglaucoma fixed combi­ could form spontaneously and does not require energy input [36]. Since
nation therapy, i.e., BT/nDHP&TM/nDHP, were examined in Angio­ tween80 has a higher HLB value (Hydrophile-Lipophile Balance) of 15.0
poietin 1 gene knockout (A1 cKO) mice, a genetic model of open-angle than that of span80 (4.3), an increase of tween80 component helps
glaucoma. Before drug application, both eyes of A1 cKO mice had an obtain smaller droplets in this water-in-oil emulsion system. Conditions
average of IOP of (23 ± 3) mmHg, significantly higher than the wild type of tween80/span80 4.5/5 (w/w) and a mild stirring (500 rpm by mag­
baseline IOP, which was around 15 mmHg (Fig. 6a) [31]. The IOP netic stirrer) are used for nanogel formation. A much higher proportion
elevation was stable, and it showed no fluctuation daily either (Fig. 6a). of surfactant with 3.7/70 (v/v) of surfactant-to-hexane is used for
Both eyes of A1 cKO mice received a dose of fixed combination topically, nanogel. The O/W ratio to form a nanogel is increased dramatically to
with one eye treated with free drugs and the other eye with formulated 762/1 (v/v).

5
J. Wang et al. Chemical Engineering Journal 425 (2021) 130498

Fig. 2. Assessment on dendrimer micro/nano gel particles. a) Cytotoxicity on NIH3T3 cells following 48-h incubation, n = 8. b) Degradation in contrived tears
incubated at 37 ℃, n = 3. c) Drug release, n = 3. d) BT permeation across the cornea, n = 3.

Fig. 3. a) Dose-dependent cytotoxicity of nDHP to HCEC cells after 48 h incubation, n = 8. b) Cytokines of TNF-α and IL6 expression on HCEC after 4 h incubation, n
= 4, ** P < 0.01 vs Untreated.

The cross-linking reaction of G5 with PEG-DA consumes most, but reactant concentration (10 wt% of G5), the degrees of un-reacted amines
not all, of the surface amines, resulting in the reduction of surface in the gel particles of different sizes vary. Our work suggests that cross-
charges of the formed gel particles. The cross-linking reaction of G5 with linked dendrimer network on the basis of aza-Michael addition reaction
PEG-DA is different with PEGylation of dendrimer. PEGylation of den­ degrade via self-triggered aminolysis, in which unreacted amine groups
drimer leads to a modified dendrimer molecule but still a single mole­ attack ester bonds in the network [37]. When more amines participate in
cule. However, the formation of dendrimer hydrogel particle is not a the cross-linking reaction, the formed network will become more
simple process of “PEGylation of dendrimers”, but rather a process to structurally stable because of reduced aminolysis. The degradation ki­
build a cross-linked structure. The 1:1 amine-to-acrylate ratio and the netics of μDHPs and nDHPs was found to be size-dependent. The smaller
steric hindrance when a dendritic molecule reacts with a linear polymer the particle size is, the more stable dendrimer gel particles are. nDHP is
chain lead to the generation of a cross-linked network with some amines more stable than µDHP10 and µDHP3. The lack of hemolysis toxicity of
remaining unreacted. What is interesting is that, under the same DHPs as well as G5 and PEG-DA indicates the blood compatibility of

6
J. Wang et al. Chemical Engineering Journal 425 (2021) 130498

DHPs.
Although the BT to G5 feed ratio is the same in the water phase, the
washing process results in different drug loss and finally leads to
different drug loading efficiency for different particles. Compared to the
quick release of BT from μDHPs, BT release was extended by nDHP and
showed zero-order release kinetics. By comparing the permeability co­
efficient of free BT, i.e., 4.5 × 10− 7 cm/s, μDHP10 does not improve the
corneal permeability of BT at all, and μDHP3 only slightly improves BT’s
corneal permeability. BT from nDHP has obtained the highest perme­
ability coefficient. In addition, for BT/nDHP formulation, the transport
of BT across the cornea almost matches its drug release kinetics. It is
meaningful for a topical ocular delivery system to synchronize drug
transport to the anterior chamber of the eye with drug delivery and
release kinetics. The prolonged retention of μDHPs is ascertained by the
permeation of FITC-labeled μDHP and the presence of NiR dye-labeled
μDHP on the ocular surface. Since there are no cleavable bonds be­
tween FITC and dendrimer, and the degradation of microgel particles in
4 h is negligible, it is reasonable to believe that microgel particles have
permeated through the cornea. These preliminary studies confirm the
μDHP’s capacities of both prolonged retention on the cornea and
permeation across the cornea.
Since nDHP possesses high cytocompatibility and enables BT to have
significantly improved permeability, zero-order release kinetics, we
selected nDHP for further in vivo examination. Prior to in vivo studies,
we decided a safe dose of nDHP by assessing cytotoxicity of nDHP to
human cornea epithelium cells (HCECs), the foremost ocular barrier for
Fig. 4. a) In vivo IOP-lowering effect of BT/nDHP and BT/PBS in normotensive
topically instilled drugs or formulations. Results show that nDHP is
rats after 7- day daily dosing topical administration. * P < 0.05 BT/nDHP vs acceptable to be used as topical antiglaucoma drug carrier at appro­
BT/PBS. b) The adjusted means of △IOP at 12 PM each day is derived using an priate doses. The sustained potent IOP-lowering effect of BT/nDHP from
analysis of covariance adjusted for baseline IOP and based on observed data for repeat once-daily dosing of BT/nDHP is attributed to the increased
all randomized subjects. The dose of each formulation is 2 × 5 μL, 0.1% w/v BT corneal permeation, zero-order drug release, and retention in the cornea
equivalent for 7 days. stroma. This cumulative decline on IOP indicates the nDHP formulation
has the great potential to reduce dosing frequency. Angiopoietin/TEK
(ANGPT/TEK) signaling is critical for Schlemm’s canal (SC) formation
and function in mice and humans [31,38]. Mice lacking ANGPT1

Fig. 5. a) Images of rat eyes immediately after instillation of BT/nDHPFITC. b) Ocular tissues of rat eyes following daily instillation of BT/PBS eye drops and BT/
nDHPFITC PBS suspensions for 7 days, magnification 200 × . c) The whole globe tissue images of rat eye after 7 days’ instillation with BT/nDHPFITC PBS suspensions.

7
J. Wang et al. Chemical Engineering Journal 425 (2021) 130498

biodistribution and clearance of nDHPs will also be studied.

5. Conclusions

In this work, we successfully miniaturized dendrimer hydrogel to

form nanostructured dendrimer hydrogel particles (nDHPs) using the
IEaMA method developed by us. We identified parameters that are
essential for fabricating nanostructured dendrimer gel particles. This
method enables the scalable production of dendrimer gel particles with
tunable size and surface charge, which are closely related to cyto­
compatibility, corneal permeability, degradation, and drug release.
Compared to micron-sized dendrimer gel particles, nDHP promotes
corneal permeability and realizes zero-order release kinetics of anti­
glaucoma drugs. nDHP has a longer residence time in the cornea stroma
and makes the encapsulated antiglaucoma drugs exert sustained and
stronger IOP-lowering effects as confirmed in normotensive and genetic
glaucoma models. Based on its intrinsically new properties and ability to
enable antiglaucoma drugs to exert prominent IOP lowering effects, we
believe nDHP represents a new generation of advanced platform. Not
only do nDHPs retain the DH properties as we demonstrated before, but
they also have additional compelling features for antiglaucoma drug
delivery as follows: (i) it facilitates corneal permeation of combined
drugs at the ratio as prescribed by confining the drugs into particulate
structures; (ii) it has high drug encapsulation capacity for both hydro­
philic drugs (or drug salt form) and hydrophobic drugs as well as it
enables programmable synchronized release of the delivered drugs, a
key feature for precise dosing of ocular drugs in combinations; and (iii)
nDHPs have the potential for developing more efficient antiglaucoma
formulations and deliver emerging classes of new drugs such as Rho
kinase inhibitor for therapeutic interventions via multiple mechanisms.

Declaration of Competing Interest

The authors declare that they have no known competing financial

Fig. 6. a) The IOP elevation of A1 cKO mice compared to wildtype littermate interests or personal relationships that could have appeared to influence
controls before formulation treatment. Data present as mean ± SE; n = 6 (left the work reported in this paper.
and right eyes of 3 mice); *** P < 0.001. b) IOP changes in A1 cKO mice
following topical treatment of one dose of BT/PBS&TM/PBS or BT/nDHP&TM/
Acknowledgments
nDHP (4 μL, 0.15% w/v BT equivalent; 4 μL, 0.5% w/v TM equivalent). * P <
0.05 vs BT/PBS&TM/PBS.
This work was supported, in part, by the National Institutes of Health
(R01EY024072 and R01EY029121) and the Fundamental Research
develop a hypomorphic SC, which is insufficient for normal aqueous
Funds for the Central Universities (Grant No. 20822041D4094,
humor outflow, resulting in sustained IOP elevation [38–40]. Angio­
1082204112665). Services and products in support of the research
poietin conditional knockout mice (A1 cKO) exhibit consistent and
project were generated by the Virginia Commonwealth University
chronic IOP elevation from about 20-days of age, and the IOP elevation
Cancer Mouse Models Core Laboratory, supported, in part, with funding
is stable with age for individual mice, which makes it an ideal model
to the Massey Cancer Center from NIH-NCI Cancer Center Support Grant
system to study the short- and long-term drug effects on IOP regulation.
P30 CA016059.
The baseline IOP of wild type mice is around 15 mmHg, and the A1 cKO
mice develop ocular hypertension bilaterally, reaching ~ 20 mmHg by
6 weeks of age, and maintain an elevated IOP in adulthood [31]. The Appendix A. Supplementary data
reduced IOP in A1 cKO provides further evidence of the delivery effi­
ciency of new formulation. Future studies will be centered on the Supplementary data to this article can be found online at https://doi.
establishment of the pharmacokinetics of antiglaucoma drugs delivered org/10.1016/j.cej.2021.130498.
by nDHP and long-term protection on vision.
In view of the efficient IOP-lowering effect, the flexibility of multi- References
drug loading and general applicability of this nDHP formulation, a
[1] Y.-C. Tham, X. Li, T.Y. Wong, H.A. Quigley, T. Aung, C.-Y. Cheng, Global
comprehensive mechanistic study including preocular retention, the prevalence of glaucoma and projections of glaucoma burden through 2040: a
transport pathway and distribution after topical administration of nDHP systematic review and meta-analysis, Ophthalmology 121 (2014) 2081–2090,
will be conducted in future work. The “nano-in-nano” dendrimer gel https://doi.org/10.1016/j.ophtha.2014.05.013.
[2] H.A. Quigley, Glaucoma, Lancet 377 (2011) 1367–1377, https://doi.org/10.1016/
formulation represents an advanced structure and innovative use of
S0140-6736(10)61423-7.
dendrimer building blocks and it holds great potential in the application [3] R.D. Bachu, P. Chowdhury, Z.H.F. Al-Saedi, P.K. Karla, S.H.S. Boddu, Ocular drug
of glaucoma therapy. The preocular retention study will be carried out delivery barriers—role of nanocarriers in the treatment of anterior segment ocular
diseases, Pharmaceutics 10 (2018) 28, https://doi.org/10.3390/
using fluorescent dye labeled nDHPs characterized by in vitro immu­
pharmaceutics10010028.
nocytochemistry and in vivo optical coherence tomography (OCT). The [4] C.J.F. Bertens, M. Gijs, F.J.H.M. van den Biggelaar, R.M.M.A. Nuijts, Topical drug
pharmacokinetics of antiglaucoma drugs delivered by nDHPs, delivery devices: a review, Exp. Eye Res. 168 (2018) 149–160, https://doi.org/
10.1016/j.exer.2018.01.010.

8
J. Wang et al. Chemical Engineering Journal 425 (2021) 130498

[5] D. Huang, Y.-S. Chen, I.D. Rupenthal, Overcoming ocular drug delivery barriers [25] J.C. Cuggino, L.I. Tartara, L.M. Gugliotta, S.D. Palma, C.I. Alvarez Igarzabal,
through the use of physical forces, Adv. Drug Deliv. Rev. 126 (2018) 96–112, Mucoadhesive and responsive nanogels as carriers for sustainable delivery of
https://doi.org/10.1016/j.addr.2017.09.008. timolol for glaucoma therapy, Mat. Sci. Eng. C 118 (2021), https://doi.org/
[6] A. Subrizi, E.M. del Amo, V. Korzhikov-Vlakh, T. Tennikova, M. Ruponen, A. Urtti, 10.1016/j.msec.2020.111383.
Design principles of ocular drug delivery systems: importance of drug payload, [26] L. Serra, J. Doménech, N.A. Peppas, Drug transport mechanisms and release
release rate, and material properties, Drug Discov. Today 24 (2019) 1446–1457, kinetics from molecularly designed poly(acrylic acid-g-ethylene glycol) hydrogels,
https://doi.org/10.1016/j.drudis.2019.02.001. Biomaterials 27 (2006) 5440–5451, https://doi.org/10.1016/j.
[7] H.A. Quigley, 21st century glaucoma care, Eye 33 (2019) 254–260, https://doi. biomaterials.2006.06.011.
org/10.1038/s41433-018-0227-8. [27] Q. Yuan, Y. Fu, W.J. Kao, D. Janigro, H. Yang, Transbuccal delivery of CNS
[8] L. Golas, C. Marando, L. Seibold, M.B. Pantcheva, P.Y. Ramulu, M.Y. Kahook, J. therapeutic nanoparticles: synthesis, characterization, and in vitro permeation
R. SooHoo, Glaucoma treatments, patient perspectives and preferences, Investig. studies, ACS Chem. Neurosci. 2 (2011) 676–683, https://doi.org/10.1021/
Ophthalmol. Vis. Sci. 56 (2015) 3703. cn200078m.
[9] V.K. Yellepeddi, S. Palakurthi, Recent advances in topical ocular drug delivery, [28] L. Xu, N. Sheybani, S. Ren, G.L. Bowlin, W.A. Yeudall, H. Yang, Semi-
J. Ocul. Pharmacol. Ther. 32 (2016) 67–82, https://doi.org/10.1089/ interpenetrating network (sIPN) co-electrospun gelatin/insulin fiber formulation
jop.2015.0047. for transbuccal insulin delivery, Pharm. Res. 32 (2015) 275–285, https://doi.org/
[10] L. Dickmann, Ocular therapeutics: drug delivery and pharmacology, Mol. Pharm. 10.1007/s11095-014-1461-9.
13 (2016) 2875–2876, https://doi.org/10.1021/acs.molpharmaceut.6b00703. [29] L. Xu, O.Y. Zolotarskaya, W.A. Yeudall, H. Yang, Click hybridization of immune
[11] Y. Ikuta, S. Aoyagi, Y. Tanaka, K. Sato, S. Inada, Y. Koseki, T. Onodera, H. Oikawa, cells and polyamidoamine dendrimers, Adv. Healthc. Mater. 3 (2014) 1430–1438,
H. Kasai, Creation of nano eye-drops and effective drug delivery to the interior of https://doi.org/10.1002/adhm.201300515.
the eye, Sci. Rep. 7 (2017) 44229, https://doi.org/10.1038/srep44229. [30] M.H. Asim, M. Ijaz, A. Mahmood, P. Knoll, A. Jalil, S. Arshad, A. Bernkop-
[12] D.R. Janagam, L. Wu, T.L. Lowe, Nanoparticles for drug delivery to the anterior Schnürch, Thiolated cyclodextrins: mucoadhesive and permeation enhancing
segment of the eye, Adv. Drug Deliv. Rev. 122 (2017) 31–64, https://doi.org/ excipients for ocular drug delivery, Int. J. Pharm. 599 (2021), 120451, https://doi.
10.1016/j.addr.2017.04.001. org/10.1016/j.ijpharm.2021.120451.
[13] H.S.B. Sai, Polymeric nanoparticles for ophthalmic drug delivery: an update on [31] B.R. Thomson, M. Grannonico, F. Liu, M. Liu, P. Mendapara, Y. Xu, X. Liu, S.
research and patenting activity, Recent Pat. Nanomed. 2 (2012) 96–112, https:// E. Quaggin, Angiopoietin-1 knockout mice as a genetic model of open-angle
doi.org/10.2174/1877912311202020096. glaucoma, Transl. Vis. Sci. Technol. 9 (2020) 16, https://doi.org/10.1167/
[14] E. Sánchez-López, M.A. Egea, B.M. Davis, L. Guo, M. Espina, A.M. Silva, A. tvst.9.4.16.
C. Calpena, E.M.B. Souto, N. Ravindran, M. Ettcheto, A. Camins, M.L. García, M. [32] I. Durocher, D. Girard, In vivo proinflammatory activity of generations 0–3
F. Cordeiro, Memantine-loaded PEGylated biodegradable nanoparticles for the (G0–G3) polyamidoamine (PAMAM) nanoparticles, Inflamm. Res. 65 (2016)
treatment of glaucoma, Small 14 (2018) 1701808, https://doi.org/10.1002/ 745–755, https://doi.org/10.1007/s00011-016-0959-5.
smll.201701808. [33] P.C. Naha, M. Davoren, F.M. Lyng, H.J. Byrne, Reactive oxygen species (ROS)
[15] L.J. Luo, D.D. Nguyen, J.Y. Lai, Dually functional hollow ceria nanoparticle induced cytokine production and cytotoxicity of PAMAM dendrimers in J774A.1
platform for intraocular drug delivery: a push beyond the limits of static and cells, Toxicol. Appl. Pharm. 246 (2010) 91–99, https://doi.org/10.1016/j.
dynamic ocular barriers toward glaucoma therapy, Biomaterials 243 (2020) 16, taap.2010.04.014.
https://doi.org/10.1016/j.biomaterials.2020.119961. [34] J.M. Martinez-de-la-Casa, F. Perez-Bartolome, E. Urcelay, J.L. Santiago, J. Moreno-
[16] C. Hu, J. Sun, Y. Zhang, J. Chen, Y. Lei, X. Sun, Y. Deng, Local delivery and Montañes, P. Arriola-Villalobos, J.M. Benitez-del-Castillo, J. Garcia-Feijoo, Tear
sustained-release of nitric oxide donor loaded in mesoporous silica particles for cytokine profile of glaucoma patients treated with preservative-free or preserved
efficient treatment of primary open-angle glaucoma, Adv. Healthc. Mater. 7 (2018) latanoprost, Ocul. Surf. 15 (2017) 723–729, https://doi.org/10.1016/j.
1801047, https://doi.org/10.1002/adhm.201801047. jtos.2017.03.004.
[17] S. Swetledge, J.P. Jung, R. Carter, C. Sabliov, Distribution of polymeric [35] C. Solans, P. Izquierdo, J. Nolla, N. Azemar, M.J. Garcia-Celma, Nano-emulsions,
nanoparticles in the eye: implications in ocular disease therapy, J. Nanobiotechnol. Curr. Opin. Colloid Interface Sci. 10 (2005) 102–110, https://doi.org/10.1016/j.
19 (2021) 10, https://doi.org/10.1186/s12951-020-00745-9. cocis.2005.06.004.
[18] S.-N. Kim, S.A. Ko, C.G. Park, S.H. Lee, B.K. Huh, Y.H. Park, Y.K. Kim, A. Ha, K. [36] T.F. Tadros, Emulsion formation, stability, and rheology, in: T.F. Tadros (Ed.),
H. Park, Y.B. Choy, Amino-functionalized mesoporous silica particles for ocular Emulsion Formation and Stability, Wiley-VCH Verlag GmbH & Co. KGaA, 2013,
delivery of brimonidine, Mol. Pharm. 15 (2018) 3143–3152, https://doi.org/ pp. 1–75.
10.1021/acs.molpharmaceut.8b00215. [37] J. Wang, H. He, R.C. Cooper, Q. Gui, H. Yang, Drug-conjugated dendrimer
[19] J. Wang, H. He, R.C. Cooper, H. Yang, In situ-forming polyamidoamine dendrimer hydrogel enables sustained drug release via a self-cleaving mechanism, Mol.
hydrogels with tunable properties prepared via aza-Michael addition reaction, ACS Pharm. 16 (2019) 1874–1880, https://doi.org/10.1021/acs.
Appl. Mater. Interfaces 9 (2017) 10494–10503, https://doi.org/10.1021/ molpharmaceut.8b01207.
acsami.7b00221. [38] T. Souma, S.W. Tompson, B.R. Thomson, O.M. Siggs, K. Kizhatil, S. Yamaguchi,
[20] J. Wang, G.S. Williamson, M.G. Lancina III, H. Yang, Mildly cross-linked dendrimer L. Feng, V. Limviphuvadh, K.N. Whisenhunt, S. Maurer-Stroh, T.L. Yanovitch,
hydrogel prepared via aza-Michael addition reaction for topical brimonidine L. Kalaydjieva, D.N. Azmanov, S. Finzi, L. Mauri, S. Javadiyan, E. Souzeau,
delivery, J. Biomed. Nanotechnol. 13 (2017) 1089–1096, https://doi.org/ T. Zhou, A.W. Hewitt, B. Kloss, K.P. Burdon, D.A. Mackey, K.F. Allen, J.B. Ruddle,
10.1166/jbn.2017.2436. S.H. Lim, S. Rozen, K.N. Tran-Viet, X. Liu, S. John, J.L. Wiggs, F. Pasutto, J.
[21] M.G. Lancina III, H. Yang, Dendrimers for ocular drug delivery, Can. J. Chem. 95 E. Craig, J. Jin, S.E. Quaggin, T.L. Young, Angiopoietin receptor TEK mutations
(2017) 897–902, https://doi.org/10.1139/cjc-2017-0193. underlie primary congenital glaucoma with variable expressivity, J. Clin. Invest.
[22] S. Kannan, H. Dai, R.S. Navath, B. Balakrishnan, A. Jyoti, J. Janisse, R. Romero, R. 126 (2016) 2575–2587, https://doi.org/10.1172/JCI85830.
M. Kannan, Dendrimer-based postnatal therapy for neuroinflammation and [39] B.R. Thomson, S. Heinen, M. Jeansson, A.K. Ghosh, A. Fatima, H.K. Sung, T. Onay,
cerebral palsy in a rabbit model, Sci. Transl. Med. 4 (2012) 130ra46, https://doi. H. Chen, S. Yamaguchi, A.N. Economides, A. Flenniken, N.W. Gale, Y.K. Hong,
org/10.1126/scitranslmed.3003162. A. Fawzi, X. Liu, T. Kume, S.E. Quaggin, A lymphatic defect causes ocular
[23] D.A. Tomalia, L.S. Nixon, D.M. Hedstrand, The role of branch cell symmetry and hypertension and glaucoma in mice, J. Clin. Invest. 124 (2014) 4320–4324,
other critical nanoscale design parameters in the determination of dendrimer https://doi.org/10.1172/JCI77162.
encapsulation properties, Biomolecules 10 (2020) 642, https://doi.org/10.3390/ [40] B.R. Thomson, T. Souma, S.W. Tompson, T. Onay, K. Kizhatil, O.M. Siggs, L. Feng,
biom10040642. K.N. Whisenhunt, T.L. Yanovitch, L. Kalaydjieva, D.N. Azmanov, S. Finzi, C.
[24] J. Wang, R.C. Cooper, H. He, B. Li, H. Yang, Polyamidoamine dendrimer microgels: E. Tanna, A.W. Hewitt, D.A. Mackey, Y.S. Bradfield, E. Souzeau, S. Javadiyan, J.
Hierarchical arrangement of dendrimers into micrometer domains with expanded L. Wiggs, F. Pasutto, X. Liu, S.W. John, J.E. Craig, J. Jin, T.L. Young, S.E. Quaggin,
structural features for programmable drug delivery and release, Macromolecules Angiopoietin-1 is required for Schlemm’s canal development in mice and humans,
51 (2018) 6111–6118, https://doi.org/10.1021/acs.macromol.8b01006. J. Clin. Invest. 127 (2017) 4421–4436, https://doi.org/10.1172/JCI95545.