Fetal uptake of intra-amniotically delivered dendrimers in a mouse model of intrauterine inflammation and preterm birth

Nanomedicine: Nanotechnology, Biology, and Medicine xx (2014) xxx – xxx nanomedjournal.com

Fetal uptake of intra-amniotically delivered dendrimers in a mouse model of intrauterine inflammation and preterm birth Irina Burd, MD, PhD a,⁎, 1 , Fan Zhang, BS b, 1 , Tahani Dada, MD a , Manoj K. Mishra, PhD b , Talaibek Borbiev, PhD a , Wojciech G. Lesniak, PhD b , Haitham Baghlaf, MD a , Sujatha Kannan, MD c , Rangaramanujam M. Kannan, PhD b,⁎⁎ a Integrated Research Center for Fetal Medicine, Gynecology and Obstetrics, Johns Hopkins School of Medicine, Baltimore, MD Center for Nanomedicine, Department of Ophthalmology, Wilmer Eye Institute, Johns Hopkins School of Medicine, Baltimore, MD c Anesthesiology and Critical Care Medicine, Johns Hopkins School of Medicine, Baltimore, MD Received 2 November 2013; accepted 9 March 2014

b

Abstract Intrauterine inflammation is associated with preterm birth and can lead to fetal neuroinflammation and neurobehavioral disorders in newborns. Dendrimers can intrinsically target and deliver drugs for the treatment of neuroinflammation. We explore whether hydroxyl polyamidoamine (PAMAM) dendrimer (G4-OH)-based nanomedicines can be delivered to the fetus by intra-amniotic administration, in a mouse model of intrauterine inflammation. The time-dependent accumulation of G4-OH-fluorophore conjugate was quantified by fluorescence. These studies suggest that, after intra-amniotic administration, there is significant accumulation of dendrimer in the fetus gut and brain. In addition, there is some fetal–maternal transport of the dendrimer. Confocal microscopy confirmed the presence of G4-OH in the fetal brain, with a large accumulation in the brain blood vessels and the brain parenchyma, and some microglial uptake. We believe that intraamniotic administration of G4-OH-drug nanomedicines may enable the treatment of diseases related to intrauterine inflammation and fetal neuroinflammation. © 2014 Elsevier Inc. All rights reserved. Key words: Intrauterine inflammation; PAMAM dendrimer; Intra-amniotic drug delivery; Biodistribution; Fetal brain

Background In the United States, approximately 12% of all live births are preterm. 1 Although mechanisms underlying spontaneous preterm birth are not fully understood, intrauterine inflammation is associated with most cases. The presence of intrauterine inflammation has been linked to a devastating spectrum of

This work was supported by Aramco Services Company fund (IB), NICHD K08HD073315 (IB), and NICHD R01 HD069562-01A1 (SK). ⁎Correspondence to: I. Burd, Integrated Research Center for Fetal Medicine, Department of Gynecology and Obstetrics, Johns Hopkins School of Medicine, Baltimore, MD. ⁎⁎Correspondence to: R.M. Kannan, Center for Nanomedicine, Wilmer Eye Institute, Johns Hopkins School of Medicine, Baltimore, MD. E-mail addresses: [email protected] (I. Burd), [email protected] (R.M. Kannan). 1 Contributed equally.

neurobehavioral disorders in these children, ranging from cognitive and learning disability to motor dysfunction such as cerebral palsy, 2 though not all cases of cerebral palsy are caused by inflammatory injury. Various models of systemic and local maternal inflammation have been used to simulate preterm birth and perinatal morbidity associated with prematurity and intrauterine inflammation. 3-10 Our team has utilized a mouse model of intrauterine inflammation (produced by localized intrauterine lipopolysaccharide [LPS] infusions) that recapitulates clinical features present in human disease to study mechanisms of fetal brain injury and to understand and develop therapeutic interventions. In this mouse model, we demonstrated that in addition to causing preterm birth, exposure to intrauterine inflammation leads to neuroinflammation and neurotoxicity in the fetal brain, making this model a useful tool for evaluation of therapeutic interventions for the mother and fetus. 5,6,11 Development of successful therapeutic interventions that address both preterm birth and prematurityrelated perinatal morbidity has been a challenge. Nanotechnology

http://dx.doi.org/10.1016/j.nano.2014.03.008 1549-9634/© 2014 Elsevier Inc. All rights reserved. Please cite this article as: Burd I., et al., Fetal uptake of intra-amniotically delivered dendrimers in a mouse model of intrauterine inflammation and preterm birth. Nanomedicine: NBM 2014;xx:1-9, http://dx.doi.org/10.1016/j.nano.2014.03.008

2

I. Burd et al / Nanomedicine: Nanotechnology, Biology, and Medicine xx (2014) xxx–xxx

Figure 1. (A-D) The process of intra-amniotic injection of dendrimer in one of the fetuses.

may offer opportunities to address this challenge, provided it is safe and efficacious in the perinatal period. Nanomedicine-based delivery strategies are emerging in the treatment of inflammatory disorders. These approaches can improve drug bioavailability, targeting, and sustained efficacy. 12-15 Specifically, systemic administration of hydroxylterminated poly(amidoamine) (PAMAM)(G4-OH) dendrimers was reported to target activated microglia and astrocytes in the brain of newborn rabbits with cerebral palsy. N-acetylcysteine (NAC), an anti-inflammatory and antioxidant drug, was conjugated to the dendrimer, and the dendrimer–NAC conjugate was shown to significantly improve motor function and reduce neuronal injury and inflammation in the rabbit model of cerebral palsy. 15 In a guinea pig model of chorioamnionitis, cervical administration of hydroxyl-terminated PAMAM dendrimer, by itself, was shown to inhibit bacterial growth and prevent preterm birth. This finding indicates that hydroxyl-terminated PAMAM dendrimers can be used as a potential topical microbicide for the treatment of ascending uterine infections 16 and may be good candidates for intra-amniotic therapy. Thus, fetal uptake studies may offer valuable insights into their effectiveness as drug delivery agents. If neonatal attenuation of neuroinflammation with dendrimers can produce significant motor function improvement in cerebral palsy, fetal dendrimer therapies may have a significant impact on preventing cerebral palsy and preterm birth. A lack of studies on drug efficacies and trans-placental transport has limited the use of fetal nanotherapies in the perinatal period. This study seeks to ascertain whether dendrimers, when delivered intra-amniotically, can be taken up by the fetus.

Methods Animals We used a mouse model of intrauterine inflammation and preterm birth as described in an established protocol. 6,5,17 Briefly, we anesthetized pregnant CD1 mice (Charles River) with continuous isoflurane in oxygen and performed intrauterine injections of LPS (100 μg/dam in 100 μL phosphate-buffered saline solution; from Escherichia coli, 055:B5; (Sigma Chemical Co., St. Louis, MO) or phosphate-buffered saline solution (PBS) on day 17 of gestation (19 days is full-term gestation). A minilaparotomy was performed in the lower abdomen. Either LPS (n = 6 dams for 24 h; n = 3 dams for 6 h) or an equal volume of vehicle (normal saline (NS); control) (n = 4 dams for 24 h; n = 3 for 6 h) was infused into the uterus between the 1st and 2nd gestational sacs closest to the cervix. Using a Hamilton syringe, we injected dendrimer-Cy5 conjugates (DCy5; 10 mg/kg at a concentration of 2 μg/mL) intra-amniotically into each gestational sac on the right side immediately after LPS or vehicle injection (Figure 1). We chose this dose based on one study of intravenous dendrimer injection in newborn rabbits 15 and after taking into account that intra-amniotic injection is somewhat similar to oral administration in that the fetus swallows the amniotic fluid. The left side of the uterus was not injected and served as an internal control. Surgical incisions were closed, and the dams were recovered in individual cages. As depicted in Figure 1, the D-Cy5 solution spread quickly into the amniotic fluid within 1 to 3 sec. No leakage of D-Cy5 was seen from the picture during or after injection.

I. Burd et al / Nanomedicine: Nanotechnology, Biology, and Medicine xx (2014) xxx–xxx

Mice were euthanized according to a protocol approved by the Animal Care and Use Committee (Johns Hopkins University). The following maternal and fetal tissues were harvested and kept on ice: maternal kidney, maternal placenta, fetal brain, fetal lung, fetal stomach, fetal gut, fetal liver.

3

Recovery rate test To assess rate of recovery of D-Cy5 from the tissues following extraction, under the applied extraction procedure, we injected samples obtained from all analyzed organs with 100 ng of the conjugate and analyzed them using the method described above. Recovery rate of D-Cy5 from all organs was measured to be 90 ± 15% of D-Cy5.

Materials Data analysis Hydroxyl-terminated ethylenediamine-core poly(amidoamine) (PAMAM) generation-4 dendrimer [G4-OH] was purchased from Dendritech Inc. (Midland, MI, USA). HPLC grade methanol (SigmaAldrich), cyanine-5-N-hydroxysuccinimide ester (Cy5-NHS ester, GE Healthcare Life Science), stainless-steel beads (Fisher Scientific), DNA LoBind tube 2.0 mL (Eppendorf, Germany), TSK gel ODS-80 Ts column (250 × 4.6 mm, i.d., 5 μm, Tosoh Bioscience LLC, Japan), 4′,6-diamidino-2-phenylindole (DAPI, Molecular Probes), rabbit anti-Iba1 (Wako, Japan), and Alexa Fluor® 594 goat anti-rabbit IgG (H + L) antibody (Molecular Probes) were purchased. Preparation of the D-Cy5 The D-Cy5 was prepared according to previously published procedures (see also supplementary information). 18

For all samples, the emission intensities at 662 nm obtained from fluorescence spectra were corrected for background fluorescence and converted into D-Cy5 concentration (g/mL) using the appropriate calibration curve: Emission Intensity ¼ a  ½D−Cy5concentration  þ b where a and b are constants that depend on the concentration of D-Cy5 and the excitation/emission slit widths. D-Cy5 concentration (g/mL) was then converted to percentage of injected dose per gram of tissue [% of ID/g] using following equation: Percentage of injected  dose mg Concentration 0:1g  Organ weight ðgÞ  100% ¼ injected dose ðmgÞ

Sample preparation The method for extraction and quantification of D-Cy5 conjugate is described elsewhere. 18 Briefly, all tissue samples were defrosted before further processing. Each sample was weighed and homogenized in 1 mL of methanol with a TissueLyser LT, Qiagen homogenizer (30 min), followed by sonication in an FS110D Fisher Scientific sonicator for 30 min. To obtain clear supernatants, we centrifuged samples for 15 min at 15,000 rpm and 4 °C in a 5424R Eppendorf centrifuge. Supernatants were then diluted with methanol to a concentration of 100 mg of tissue per 1 mL of methanol and analyzed by fluorescence spectroscopy. Fluorescence spectroscopy Fluorescence spectra of D-Cy5 conjugate and extracts were recorded using a Shimadzu RF-5301 spectrofluorometer. Calibration curves for D-Cy5 conjugates (λmax = 662 nm) obtained from spectra recorded from 650 nm to 720 nm with an excitation wavelength of 645 nm for its methanol solutions, at concentration ranging from 1 ng/mL to 10 μL under different slit widths. To maximize intensity of emission, we applied different sets of excitation and emission slit widths for different D-Cy5 concentrations. All calibration curves exhibited excellent linearity, with R ~ 0.999 (Figure 2, A). Spectra for each extract were recorded using different sets of excitation and emission slit widths. We selected the one with the highest emission intensity for quantification of D-Cy5, applying calibration curve obtained at the same condition. To compensate for autofluorescence originating from the matrix, we subtracted the intensities registered for samples from animals without D-Cy5 injections.

For each organ, we analyzed at least three samples. Results are expressed as average standard error, n = 3. 3.6. Immunohistochemistry study of D-Cy5 brain accumulation and biodistribution Fetal brains were fixed with 4% paraformaldehyde and then processed with gradient sucrose solution (10%, 20%, 30%). Processed brains were cryosectioned on a Leica CM1850 cryostat (Leica Biosystems, USA). Tissue sections were stained with DAPI, which stains nuclei, and anti-ionized calciumbinding adapter molecule 1 (anti-Iba1, primary antibody). Goat anti-rabbit 594 (secondary antibody) was then used to stain the microglia/macrophages. Sections were viewed on a Zeiss LSM 710 Meta Confocal Microscope (Carl Zeiss, USA). We used a uniform parameter (gain, offset) for D-Cy5 channel to facilitate comparison of images with the same magnification. Results We first analyzed the fetal biodistribution and the maternal kidney uptake of the dendrimer using D-Cy5 fluorescence quantification and immunohistochemistry with the appropriate calibration curves. We examined the role of fetal inflammation on the biodistribution of dendrimer by comparing D-Cy5 uptake in LPS-treated and healthy controls. Fluorescence spectroscopy The presence of D-Cy5 in the tissue can be demonstrated by comparing the obtained fluorescence emission spectra after tissue extraction (with D-Cy5 treatment) with those of control

4

I. Burd et al / Nanomedicine: Nanotechnology, Biology, and Medicine xx (2014) xxx–xxx

Figure 2. (A) Representative calibration curve for D-Cy5. The curve was based on spectra recorded at emission slit width 10 and excitation slit width 10. (B) The emission spectra of extracts obtained from maternal kidney of a control mouse (blue, dashed line) and a mouse treated with lipopolysaccharide and D-Cy5 (red, solid line). Recordings were made at excitation slit 5 and emission slit 10, with excitation at 645 nm. Inset: emission spectrum of D-Cy5 conjugate recorded using the same conditions at concentration of 10 −7 g/mL. High intensity peaks in emission spectra at around 662 nm indicate presence of D-Cy5 in maternal kidney after administration of the conjugate into the amniotic sac.

(without D-Cy5 treatment, or background) and free D-Cy5 solution (calibration). Figure 2, B illustrates this for maternal kidney. Cy5 is a near-infrared imaging dye with emission λmax = 670 nm. The spectrum of kidney extraction from LPSand D-Cy5-treated animals had a distinguishable peak around 657-700 nm. This wavelength corresponded to the peak present in the spectrum of free D-Cy5 solution, indicating the presence of D-Cy5 in the kidney of LPS- and D-Cy5-treated animals. However, at the same wavelength range, the spectrum of the kidney extraction from control animals (with no D-Cy5) was flat, suggesting no measurable emission at this wavelength. Biodistribution of D-Cy5 in fetal organs, placenta, and maternal kidney Studying biodistribution in the fetus is a challenge because of the extremely small sizes of the organs (brain ~ 60 mg), and the difficulties in isolating them clearly. The error bars in the measurements reflect this, in addition to the intrinsic animal-toanimal variations. To produce statistically significant analysis, we used more than 20 dams to produce n = 3-6 dams for each study group (given the high preterm birth rate in this animal model). Every analyzed organ sample is a combination of organs from all fetuses from one dam (~5 fetuses/dam). Therefore, the error bars reflect variation in the uptake from many combined fetuses from different mothers. Fluorescence quantification with D-Cy5 enabled dendrimer uptake to be quantified up to ng/g sensitivities, without the need for radiolabeling. Figure 3 and Table 1 shows the biodistribution of D-Cy5 as a percentage of the administered dose in major fetal organs, maternal kidney, and placenta for both LPS-treated and NS-treated animals at 6 and 24 h after D-Cy5 injection. The total amount of D-Cy5 present in major fetal organs, maternal kidney, and placenta, taken together, was less than 8% at 6 h, and did not change appreciably at 24 h. Because the fetus

swallows the dendrimers that are administered in the amniotic sac, excretes them via urine into the amniotic fluid and swallows them again, it is somewhat understandable that the dendrimer uptake in the fetal organs was not significantly different between 6 and 24 h. This also suggests that the dendrimer may be retained in the amniotic sac to an appreciable extent over the 24 h, even though it is present in the maternal kidney and placenta. It would have been valuable to quantify the dendrimer in the amniotic fluid to account for where all the administered dendrimer went. However, in mouse models, collection of amniotic fluid is technically difficult because of the small size of the fetus (~ 2 cm), the relatively small space between the fetus and the amniotic membrane, and the small amount of amniotic fluid (less than 20 μL per sac). 19 Studies in large animals may enable amniotic fluid quantification. The total amount of D-Cy5 present in all the collected fetal organs was ~ 1 to 2% of injected dose. Fetal brain was estimated to uptake D-Cy5 at ~ 0.2-0.3% of total injected dose. This value is relatively higher than the amount reported in other studies, where D-Cy5 was administered intravenously to newborn rabbits and quantified by the same method we used here. 18 D-Cy5 uptake was also seen in fetal lung (~ 0.1-0.2%), stomach (0.2%), liver (0.2%), and gut (~ 0.2-0.8%). A higher presence of D-Cy5 in the stomach and gut is understandable because fetuses were ingesting the amniotic fluid. Compared to fetal organs, maternal organs (maternal kidney and placenta) showed higher accumulation of D-Cy5 (4-7%). Specifically, 3-6% of injected dose was seen in placenta and 0.61.4% was seen in maternal kidney. Two factors probably contribute to the high level of D-Cy5 in the placenta: (i) the amniotic sac into which D-Cy5 is administered is in contact with the placenta, and thus the dendrimer could partition to the placenta; (ii) the dendrimer from fetal circulation could reach the maternal circulation and be present in the large amount of blood in the placenta (we did not perfuse the mother).

I. Burd et al / Nanomedicine: Nanotechnology, Biology, and Medicine xx (2014) xxx–xxx

5

Figure 3. Biodistribution of D-Cy5 in major fetal organs (A) and placenta and maternal kidneys (B) at 6 h (light blue, light orange bars) and 24 h (dark blue and dark orange bars) after injection into amniotic sac for normal saline-treated (light blue, dark blue bars) and lipopolysaccharide-treated (light orange and dark orange bars) animals. For normal saline and D-Cy5 treated group, sacrificed at 6 h post injection, n = 3 dams (~ 5 fetuses/dam); for normal saline and D-Cy5 treated group, sacrificed at 24 h post injection, n = 4 dams (~ 5 fetuses/dam); for LPS and D-Cy5 treated group, sacrificed at 6 h post injection, n = 3 dams (~ 5 fetuses/dam); for LPS and D-Cy5 treated group, sacrificed at 24 h post injection, n = 6 dams (~ 5 fetuses/dam).

Table 1 Biodistribution of D-Cy5 in the major fetal organs, placenta, and maternal kidneys 6 and 24 h post-injection. Organ

Fetus brain Fetus lung Fetus stomach Fetus liver Fetus gut Placenta Maternal kidney

NS + D-Cy5 6 h [% of ID/g]

NS + D-Cy5 24 h [% of ID/g]

LPS + D-Cy5 6 h [% of ID/g]

LPS + D-Cy5 24 h [% of ID/g]

Average

SEM

Average

SEM

Average

SEM

Average

SEM

0.235 0.095 0.185 0.184 0.144 2.810 0.617

0.013 0.028 0.032 0.037 0.008 0.789 0.177

0.188 0.124 0.184 0.353 0.776 4.048 0.947

0.093 0.024 0.068 0.059 0.088 1.020 0.239

0.305 0.215 0.158 0.125 0.388 6.223 1.409

0.014 0.005 0.081 0.032 0.010 0.950 0.177

0.135 0.086 0.178 0.216 0.451 6.161 1.047

0.012 0.010 0.048 0.053 0.039 0.444 0.148

Data are expressed as percent of injected dose per gram of tissue. ID, injected dose; LPS, lipopolysaccharide; NS, normal saline; SEM, standard error of the mean.

D-Cy5 accumulation in fetal brain To understand the higher uptake in fetal brain, we studied DCy5 accumulation and distribution in fetal brain using confocal microscopy at 24 h after administration. Initially, a low magnification (10 ×) was used to study the overall distribution of D-Cy5 in the whole brain slice (Figure 4, A and B). Samples from D-Cy5-treated fetus showed a significantly higher Cy5 signal than did samples from non-D-Cy5-treated fetuses. This result indicates that through intra-amniotic administration, D-Cy5 can reach the fetal brain. Once D-Cy5 is administered into amniotic fluid, it gets swallowed by the fetus, traverses the GI tract, is absorbed into circulation, and penetrates the immature blood–brain barrier, reaching the brain parenchyma. Therefore, we examined the tissue slices under higher magnification (20 ×) (Figure 4, C and D). Significant D-Cy5 accumulation could be seen in the blood vessels at 24 h, with appreciable D-Cy5 in the brain parenchyma (well above background levels). The Cy5 signal in the brain parenchyma, but not in the blood vessels or

cells, was much higher at 24 h (Figure 4, D) than that in the nonD-Cy5 administered background (Figure 4, C), indicating that DCy5 may have diffused into brain parenchyma from the blood vessels in the brain. In the NS-injected healthy brains, microglia had a relatively ramified morphology (Figure 5, A). In LPS-treated animals with fetal neuroinflammation, by 6 h, there was a hint of activation (Figure 5, B). By 24 h, microglia showed an amoeboid morphology, suggestive of significant activation (Figure 5, C). Although significant amounts of the D-Cy5 fluorescence were present in the brain blood vessels and parenchyma, some microglia uptake was seen (Figure 6, inset). Because the D-Cy5 may be present in the amniotic sac for an extended period of time, more and more of it would be consumed and transferred to the circulation, explaining the large presence in the blood vessels at longer times, compared to intravenous administration, which showed near complete blood clearance by 24 h. Because the fetuses are small, it is not possible to perfuse them to remove blood for imaging.

6

I. Burd et al / Nanomedicine: Nanotechnology, Biology, and Medicine xx (2014) xxx–xxx

Figure 4. Row 1: Biodistribution of D-Cy5 in lipopolysaccharide (LPS)-treated fetus brain at 24 h post-D-Cy5 administration. Blue: DAPI staining of nucleus; red: D-Cy5. (A) Without D-Cy5 administration, coronal section; (B) with D-Cy5 administration, axial section. Images were taken under 10 × magnification with tilescan. Row 2: Higher magnification study of D-Cy5 biodistribution in LPS-treated fetus brain at 24 h post-D-Cy5 administration. Green: anti-Iba1 staining of microglia/macrophages; red: D-Cy5. (C) Without D-Cy5 administration. (D) With D-Cy5 administration. Images were taken under 20× magnification with tilescan. In all images, gain and offset for Cy5 channel were chosen to be the same.

Figure 5. Microglia/macrophage morphology at different time point after lipopolysaccharide (LPS) treatment. Green: anti-Iba1 staining of microglia/ macrophages; images were taken under 20 × magnification with tilescan. (A) Six hours after normal saline (NS) treatment. (B) Six hours after LPS treatment. (C) Twenty-four hours after LPS treatment.

Discussion Comparison of oral and intra-amniotic delivery Oral delivery requires that nanocarriers be relatively small and have appropriate surface attributes to facilitate their transepithelial transport. On one hand, several in vitro studies

on the ability of PAMAM dendrimer to improve drug transport across epithelial cells have shown that PAMAM dendrimer could pass through the epithelial barrier by paracellular and transcellular transport mechanism and help increase the uptake of small drug molecules. 20 On the other hand, the in vivo studies, although very few, also suggested

I. Burd et al / Nanomedicine: Nanotechnology, Biology, and Medicine xx (2014) xxx–xxx

Figure 6. Confocal microscopy image from brain at 24 h after D-Cy5 administration. Image was taken under 20× magnification with tilescan. The magnified image in the lower left corner shows cellular uptake of D-Cy5 by microglia/macrophages in the brain (white arrows).

that dendrimer can enhance intestinal penetration and increase the absorption of camptothecin. 21 In our study, we observed substantial D-Cy5 accumulation in the fetal gut. This finding indicates the similarity between oral and intra-amniotic delivery from the fetal point of view. However, because DCy5 is continuously present in the amniotic fluid and exposed to the fetal GI tract, the absorption is expected to be higher than that of oral delivery. This attribute can help in achieving a better efficacy if dendrimer is loaded with drug molecules. Biodistribution of hydroxyl-terminated PAMAM dendrimer These are the first studies on intra-amniotically delivered nanoparticles. Our findings suggest that the swallowed hydroxyl PAMAM dendrimers are taken up into circulation and into the major organs. The dendrimer appears to be taken up by the fetal brain within 6-24 h. The level of dendrimer in fetal brain as a fraction of the injected dose was 10-fold higher than that attained by postnatal intravenous therapy. In terms of dendrimer pharmacokinetics, we have previously shown that dendrimer accumulation reaches a peak value earlier than 24 h after an intravenous injection. 18 These dendrimers get cleared from blood circulation within 24 h when administered intravenously. The current studies suggest that, after intraamniotic administration, the dendrimer is retained in the amniotic fluid, providing repeated opportunities for the fetus to take them up orally and excrete them. Transport of D-Cy5 into the brain parenchyma may be facilitated by an immature blood–brain-barrier. 22,23 Microglial activation in the LPSadministered fetuses appears to take ~ 6-24 h. By 24 h, the dendrimer is present mostly in the blood vessels in the brain,

7

with some presence in the parenchyma and in activated microglia. This finding may explain why the quantified data did not reflect a significant difference in D-Cy5 brain uptake between LPS-treated and healthy controls. Most likely, a significant amount of D-Cy5 in brain tissue was present in blood vessels in both sets of animals, reducing the difference caused by cellular uptake. The continuous presence of the D-Cy5 in amniotic fluid and uptake from fetal gut into the vascular compartment may lead to a slow and continuous delivery of the dendrimer. The in vivo model that we used results in preterm birth 70-95% of the time, 5 which makes the analysis of time points beyond 24 h difficult. In future studies, we will investigate the kinetics of dendrimer uptake and distribution in tissue of offspring after birth. We found that uptake of D-Cy5 was lower in fetal lung than in fetal gut. In the gestation period, fetal lung vasculature is not yet well developed; therefore not much dendrimer accumulates in the lung. Additionally, the flow exchange created by fetal breathing movements is less than that caused by swallowing movement. However, the amniotic fluid absorption by fetal gut is part of the primary route for the removal of amniotic fluid, which will greatly contribute to the absorption of amniotic fluid/D-Cy5 in the gut. Although we did not analyze the skin uptake owing to the small size of the mouse fetus, we believe that it will show D-Cy5 uptake. Because fetal skin is a part of the transmembranous pathway (the others are fetal-surface of the placenta and umbilical cord), the transfer between amniotic fluid and fetal blood will favor the uptake of D-Cy5 in these systems. In studies related to the safety of dendrimers in maternal–fetal applications, we have previously shown that partition of dendrimers to, and diffusion through, the fetal membrane is slow. 28 Also, intravenously administered hydroxyl-terminated G4-OH dendrimer was shown to be safe even at a dose of 550 mg/kg in a rabbit model of cerebral palsy. 24 Here, we found that accumulation of D-Cy5 in major fetal organs accounted for 1-2% of total injected dose at both analyzed time points. Eventually, the accumulated D-Cy5 is expected to clear out from the newborn. Therefore, although substantial work is still needed to investigate toxicity and efficacy, these studies suggest that intra-amniotic delivery is a feasible administration method for the treatment of fetal brain injury. Transport of hydroxyl-terminated PAMAM dendrimer across blood–placenta barrier One of the interesting aspect of our results is the fact that the dendrimer is transported from the amniotic fluid to the placenta and also to the maternal kidney. The presence of D-Cy5 in the maternal kidney suggests that fetal dendrimer is absorbed into the fetal circulation and thereafter through the placenta to the maternal side. This fetal–maternal transport is interesting considering that similar studies with other nanoparticles reported a lack of transport between maternal and fetal circulation in in vitro and in vivo models. 25-31 Although murine (haemotrichorial) and human (haemonochorial) placental developments have considerable differences, at E17-19 of gestation (time periods used in these experiments), murine placenta exhibits a chorioallontoic placentation, which,

8

I. Burd et al / Nanomedicine: Nanotechnology, Biology, and Medicine xx (2014) xxx–xxx

despite its differences in endocrine function, is thought to have similar transport mechanisms. 32,33 The two species have similar invasive pathways for trophoblasts and similar exchange pathways involving the syncytiotrophoblasts. 32 The full-term gestation for mice is 19 days, whereas it is 40 weeks for humans. In this experiment, the survival surgery and intrauterine injection of LPS/D-Cy5 were performed on gestational day 17, which corresponds to approximately 90% of the gestation length, or week 35-36 of a human pregnancy. Therefore, a larger therapeutic window will be available in humans for preventive treatments before birth. The transport mechanism of various types of nanoparticles across placenta has been investigated previously. 28-31 In those studies, nanoparticles administered from the maternal side were reported to be detected in the placenta. However, the amount present in placenta can vary widely depending on the physicochemical properties of the nanoparticles. It has been suggested that size and surface chemical properties are the two major factors that influence the transport mechanism of nanoparticles. 34 The transport mechanism of nanoparticles is highly sizedependent. Small nanoparticles (~ 5 nm) are more likely to cross the layers of tissues in placenta by paracellular diffusion. However, larger molecules (~ 50-240 nm) may cross by a transcellular mechanism. For example, ex vivo studies by Wick et al suggested that fluorescently labeled polystyrene beads with sizes of approximately 50 nm and 80 nm can transport through the blood–placenta barrier without affecting the viability of tissue. The presence of polystyrene beads in fetal perfusion medium can be as high as 29%-36% of initial perfused dose after 180 min of perfusion. However, when the size increased beyond 240 nm, less than 8% of the initial perfused dose of polystyrene beads was detected in the fetal perfusion medium. 26 Generation-4 PAMAM dendrimers are much smaller (~ 4 nm) than other nanoparticles that have been studied, a factor that could influence the transport properties. Ex vivo studies using a human placental perfusion system showed that the G4-OH dendrimers do not penetrate the blood–placenta barrier to an appreciable extent from the maternal side to the fetal side. Because the blood circulation time of these dendrimers is low (~ 2-3 h), appreciable transfer from the maternal side to fetal side would not be expected. 28 We have previously shown that when free drug (antipyrine) and fluorescently labeled hydroxylterminated PAMAM dendrimer were perfused across the human placenta, the maternal–fetal transplacental rate of PAMAM was b 0.1% and significantly less than that of the positive control (antipyrine). 28 This finding suggested that during pregnancy, hydroxyl-terminated PAMAM dendrimer can selectively deliver therapeutics to the maternal compartment without significantly affecting the fetus. Although the previous ex vivo study showed that the blood–placenta barrier could impede the transport of hydroxyl-terminated PAMAM dendrimer from the maternal side to the fetal side, our data indicated that when the same dendrimer was delivered from the fetal side, there was appreciable fetal– maternal transport. Three to six percent of the intra-amniotically administered dendrimer was detected in the placenta, and ~ 1% was found in the maternal kidney. These values remained stable even 24 h after injection in both the LPS-treated and NS-treated

groups. Because the extent of dendrimer transport from the fetal to the maternal compartment was not significant, most of the dendrimer could have remained in the amniotic fluid. Thus, intraamniotic treatment of fetal inflammation with dendrimer-drug nanodevices may be possible with tailored release of the drug into the fetal compartment.

Conclusion In this study, we have demonstrated that in a mouse pre-term labor model, delivery of neutral, generation-4 PAMAM dendrimers into the amniotic fluid results in measureable uptake and accumulation in the fetal organs, with some uptake into the placenta, and transport into the maternal compartments. The accumulation of D-Cy5 in fetal brain after intra-amniotic administration was relatively higher (as a fraction of the injected dose) than that achieved after postnatal intravenous administration. Therefore intra-amniotic administration may be a viable way to deliver therapeutics for the treatment of fetal brain injuries and other fetal disorders related to inflammation. However, the potential application of this technology depends on identifying those neonates at highest risk of developing brain injury. Interestingly, intra-amniotic administration route is akin to oral delivery, suggesting that hydroxyl-PAMAM dendrimers may be orally bioavailable. Building on the recent large animal studies with hydroxy dendrimers for cardiac arrest, these studies suggest that opportunities may also exist in maternal fetal medicine for PAMAM dendrimers. 35 Acknowledgment The authors thank Wilmer Imaging and Microscopy Core Facility for the use of the facilities.

Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.nano.2014.03.008.

References 1. CDC health disparities and inequalities report — United States 2011. Centers for Disease Control and Prevention, 2011 January 14. 2. Goldenberg RL, Culhane JF, Iams JD, Romero R. Epidemiology and causes of preterm birth. Lancet 2008;371:75-84. 3. Elovitz MA, Wang Z, Chien EK, Rychlik DF, Phillippe M. A new model for inflammation-induced preterm birth: the role of platelet-activating factor and Toll-like receptor-4. Am J Pathol 2003;163:2103-11. 4. Burd I, Chai J, Gonzalez J, Ofori E, Monnerie H, Le Roux PD, et al. Beyond white matter damage: fetal neuronal injury in a mouse model of preterm birth. Am J Obstet Gynecol 2009;201:279 e1-8. 5. Burd I, Bentz AI, Chai J, Gonzalez J, Monnerie H, Le Roux PD, et al. Inflammation-induced preterm birth alters neuronal morphology in the mouse fetal brain. J Neurosci Res 2010;88:1872-81. 6. Elovitz MA, Brown AG, Breen K, Anton L, Maubert M, Burd I. Intrauterine inflammation, insufficient to induce parturition, still evokes fetal and neonatal brain injury. Int J Dev Neurosci 2011;29:663-71.

I. Burd et al / Nanomedicine: Nanotechnology, Biology, and Medicine xx (2014) xxx–xxx 7. Zuckerman L, Weiner I. Maternal immune activation leads to behavioral and pharmacological changes in the adult offspring. J Psychiatr Res 2005;39:311-23. 8. Debillon T, Gras-Leguen C, Verielle V, Winer N, Caillon J, Roze JC, et al. Intrauterine infection induces programmed cell death in rabbit periventricular white matter. Pediatr Res 2000;47:736-42. 9. Kannan S, Saadani-Makki F, Muzik O, Chakraborty P, Mangner TJ, Janisse J, et al. Microglial activation in perinatal rabbit brain induced by intrauterine inflammation: detection with 11C-(R)-PK11195 and smallanimal PET. J Nucl Med 2007;48:946-54. 10. Saadani-Makki F, Kannan S, Lu X, Janisse J, Dawe E, Edwin S, et al. Intrauterine administration of endotoxin leads to motor deficits in a rabbit model: a link between prenatal infection and cerebral palsy. Am J Obstet Gynecol 2008;199. 11. Burd I, Chai J, Gonzalez J, Ofori E, Monnerie H, Le Roux PD, et al. Beyond white matter damage: fetal neuronal injury in a mouse model of preterm birth. Am J Obstet Gynecol 2009;201:279.e1-8. 12. Menjoge AR, Kannan RM, Tomalia DA. Dendrimer-based drug and imaging conjugates: design considerations for nanomedical applications. Drug Discov Today 2010;15:171-85. 13. Hayder M, Poupot M, Baron M, Nigon D, Turrin CO, Caminade AM, et al. A phosphorus-based dendrimer targets inflammation and osteoclastogenesis in experimental arthritis. Sci Transl Med 2011;3:81ra35. 14. Shaunak S, Thomas S, Gianasi E, Godwin A, Jones E, Teo I, et al. Polyvalent dendrimer glucosamine conjugates prevent scar tissue formation. Nat Biotechnol 2004;22:977-84. 15. Kannan S, Dai H, Navath RS, Balakrishnan B, Jyoti A, Janisse J, et al. Dendrimer-based postnatal therapy for neuroinflammation and cerebral palsy in a rabbit model. Sci Transl Med 2012;4:130ra46. 16. Wang B, Navath RS, Menjoge AR, Balakrishnan B, Bellair R, Dai H, et al. Inhibition of bacterial growth and intramniotic infection in a guinea pig model of chorioamnionitis using PAMAM dendrimers. Int J Pharm 2010;395:298-308. 17. Breen K, Brown A, Burd I, Chai J, Friedman A, Elovitz MA. TLR-4dependent and -independent mechanisms of fetal brain injury in the setting of preterm birth. Reprod Sci 2012;19:839-50. 18. Lesniak W, Mishra MK, Jyoti A, Balakrishnan B, Zhang F, Nance E, et al. Biodistribution of fluorescently-labeled PAMAM dendrimers in neonatal rabbits: effect of neuroinflammation. Mol Pharm 2013;10:4560-71. 19. Cheung CY, Brace RA. Amniotic fluid volume and composition in mouse pregnancy. J Soc Gynecol Investig 2005;12:558-62. 20. Sadekar S, Ghandehari H. Transepithelial transport and toxicity of PAMAM dendrimers: implications for oral drug delivery. Adv Drug Deliv Rev 2012;64:571-88.

9

21. Sadekar S, Thiagarajan G, Bartlett K, Hubbard D, Ray A, McGill LD, et al. Poly(amido amine) dendrimers as absorption enhancers for oral delivery of camptothecin. Int J Pharm 2013;456:175-85. 22. Saunders NR, Liddelow SA, Dziegielewska KM. Barrier mechanisms in the developing brain. Front Pharmacol 2012;3:46. 23. Abbott NJ, Patabendige AA, Dolman DE, Yusof SR, Begley DJ. Structure and function of the blood–brain barrier. Neurobiol Dis 2010; 37:13-25. 24. Balakrishnan B, Nance E, Johnston MV, Kannan R, Kannan S. Nanomedicine in cerebral palsy. Int J Nanomedicine 2013;8:4183-95. 25. Myllynen PK, Loughran MJ, Howard CV, Sormunen R, Walsh AA, Vahakangas KH. Kinetics of gold nanoparticles in the human placenta. Reprod Toxicol 2008;26:130-7. 26. Wick P, Malek A, Manser P, Meili D, Maeder-Althaus X, Diener L, et al. Barrier capacity of human placenta for nanosized materials. Environ Health Perspect 2010;118:432-6. 27. Cartwright L, Poulsen MS, Nielsen HM, Pojana G, Knudsen LE, Saunders M, et al. In vitro placental model optimization for nanoparticle transport studies. Int J Nanomedicine 2012;7:497-510. 28. Menjoge AR, Rinderknecht AL, Navath RS, Faridnia M, Kim CJ, Romero R, et al. Transfer of PAMAM dendrimers across human placenta: prospects of its use as drug carrier during pregnancy. J Control Release 2011;150:326-38. 29. Yamashita K, Yoshioka Y, Higashisaka K, Mimura K, Morishita Y, Nozaki M, et al. Silica and titanium dioxide nanoparticles cause pregnancy complications in mice. Nat Nanotechnol 2011;6:321-8. 30. Refuerzo JS, Godin B, Bishop K, Srinivasan S, Shah SK, Amra S, et al. Size of the nanovectors determines the transplacental passage in pregnancy: study in rats. Am J Obstet Gynecol 2011;204:546 e5-9. 31. Ali H, Kalashnikova I, White MA, Sherman M, Rytting E. Preparation, characterization, and transport of dexamethasone-loaded polymeric nanoparticles across a human placental in vitro model. Int J Pharm 2013;454:149-57. 32. Malassine A, Frendo JL, Evain-Brion D. A comparison of placental development and endocrine functions between the human and mouse model. Hum Reprod Update 2003;9:531-9. 33. Dilworth MR, Sibley CP. Review: transport across the placenta of mice and women. Placenta 2013;34:S34-9 Suppl. 34. Kulvietis V, Zalgeviciene V, Didziapetriene J, Rotomskis R. Transport of nanoparticles through the placental barrier. Tohoku J Exp Med 2011;225:225-34. 35. Mishra MK, Beaty CA, Lesniak WG, Kambhampati SP, Zhang F, Wilson MA, et al. Dendrimer brain uptake and targeted therapy for brain injury in a large animal model of hypothermic circulatory arrest. ACS Nano 2014;8:2134-47.