This invention relates, inter alia, to methods of reducing oxidative stress in the placenta of a pregnant mammal, to methods of treating conditions involving such oxidative stress (e.g. preeclampsia), and to nanoparticulate compositions that can be used in such methods.

The listing or discussion of an apparently prior-published document in this specification should not necessarily be taken as an acknowledgement that the document is part of the state of the art or is common general knowledge.

Oxidative stress in the placenta, for example resulting from hypoxia / reoxygenation injury, is known to have a variety of damaging effects during pregnancy. Such effects include:

an increased risk of developing preeclampsia (see, for example, Hubel, C.A., Proc. Soc. Exp. Biol. Med. 1999, 222, 222-235);

- intrauterine growth restriction (see, for example, Scifres and Nelson, J. Physiol. 2009, 14, 3453-3458);

low birth weight (see, for example, Kim, Y-J. et ai , Reproductive Toxicology 2005, 19(4), 487-492); and

long-term effects upon the foetus, including an increased risk of disease in later life (see, for example, Curtis, D.J. et ai , Experimental Neurology 2014, 261, 386-395).

There are currently no authorised treatments available that have been proven to prevent or ameliorate such damaging effects. For example, despite a report of efficacy (in preventing preeclampsia) by oral administration of an oil-based formulation of coenzyme Qio (see Teran, E. et al. Int. J. Gynecol, and Obstet. 2009, 705, 43-45), various clinical trials involving administration of natural antioxidants such as vitamin C to pregnant women have failed to show any reduction in the risk of preeclampsia, or improvement in maternal or foetal outcomes (see, for example, Villar J. et al. BJOG 2009, 776, 780-788 and Rumbold A., et al. Cochrane Database Syst. Rev. 2008, 7).

Certain endogenous antioxidants, such as the quinone-based compound coenzyme Qio (CoQio, otherwise known as ubiquinone or ubidecarenone), are highly lipophilic molecules that have limited bioavailability when administered orally. In view of this, various formulations of CoQio have been made in an effort to improve the oral bioavailability of that substance. Such formulations include nanoparticulate formulations, either of CoQio on its own, or together with an excipient such as a lipid (see, for example: Siekmann B. , Westesen K. , Pharm. Res. 1995, 72, 201 -208; Kommuru, T.R. et ai , International Journal of Pharmaceutics 200 , 272(2), 233- 246; Bunjes H. et ai, Pharm. Res. 2001 , 78, 278-293; and Hsu, C-H. et ai, AAPS Pharm. Sci. Tech 2003, 4(3), article 32).

Polymeric nanoparticle formulations including CoQio are also known (see Thunemann A.F., General S., J. Control. Release 2001 , 75, 237-247 and Kwon, S.S. et ai , Colloids and Surfaces A: Physicochemical and Engineering Aspects 2002, 270(1 ), 95-104). It has been reported (see Sood et ai, Nat. Nanotechnol. 2011 , 6, 824-833) that, in an in vitro model involving a barrier formed from cells of the BeWo choriocarcinoma cell line (which has been widely used as an in vitro model of the placenta):

DNA damage in cells (fibroblasts) grown underneath a bi-/multi-layer barrier can be caused by exposure of the other (top) side of the barrier to agents (CoCr or ΤΊΟ2 nanoparticles) that do not pass through the barrier but that can cause oxidative injury (through the generation of reactive oxygen species) in cells that they contact; and

that DNA damage can be reduced by co-exposure of the top side of the barrier to Gap 27 (which blocks connexin 43 hemichannels and gap junctions), mitoquinone (MitoQ, which is a mitochondria-targeted antioxidant), Vitamin C (an antioxidant) or catalase (which breaks

However, that report does not demonstrate the existence of corresponding processes in vivo. Moreover, it neither discloses nor suggests means by which putative therapeutics can be targeted for delivery to the placenta of a mammal. Thus, there remains a need for a treatment that can prevent or ameliorate the damaging effects of oxidative stress in the placenta of a pregnant mammal, in particular a treatment that does not cross the placental barrier.

It has surprisingly been discovered that nanoparticulate formulations of antioxidants or ATP signalling inhibitors, in particular mitochondrially-targeted antioxidants or ATP signalling inhibitors, when administered to a pregnant mammal, can alleviate adverse consequences of oxidative stress in the placenta.

Thus, according to a first aspect of the invention, there is provided a method of treating or preventing oxidative stress in the placenta of a pregnant mammal, said method comprising administering a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor to said pregnant mammal.

In corresponding second and third aspects of the invention, there are provided, respectively: - a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor for use in treating or preventing oxidative stress in the placenta of a pregnant mammal; and

the use of a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor for the manufacture of a medicament for treating or preventing oxidative stress in the placenta of a pregnant mammal.

As discussed above, oxidative stress in the placenta of a pregnant mammal is associated with a variety of adverse outcomes for the mother and/or the foetus. Treating or preventing oxidative stress in the placenta can therefore ameliorate or prevent those adverse outcomes.

Thus, according to a fourth aspect of the invention there is provided a method of treating or preventing preeclampsia or extrinsic intrauterine growth restriction (IUGR), said method comprising administering a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor to a pregnant mammal suffering from preeclampsia or extrinsic IUGR, or at risk of developing preeclampsia or extrinsic IUGR.

In corresponding fifth and sixth aspects of the invention, there are provided, respectively: - a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor for use in treating or preventing preeclampsia or extrinsic IUGR in a pregnant mammal, wherein said pregnant mammal is suffering from preeclampsia or extrinsic IUGR, or at risk of developing preeclampsia or extrinsic IUGR; and

the use of a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor for the manufacture of a medicament for treating or preventing preeclampsia or extrinsic IUGR in a pregnant mammal, wherein said pregnant mammal is suffering from preeclampsia or extrinsic IUGR, or at risk of developing preeclampsia or extrinsic IUGR. Embodiments of the fourth to sixth aspects of the invention that may be mentioned include those in which the disorder to be treated (or prevented) is preeclampsia.

Similarly, administration of a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor can treat or ameliorate other adverse effects of oxidative stress in the placenta, such as preterm birth and low birth weight.

Thus, in a seventh aspect of the invention there is provided a method of

preventing preterm birth and/or of increasing birth weight, or

preventing, or reducing the severity of, neurodevelopmental disorders in the foetus, said method comprising administering a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor to a pregnant mammal.

The mammal When used herein, the term "mammal" includes references to human and non-human mammals. Non-human mammals that may be mentioned include simians, murines, rats, leporids, felines and canids.

Particular mammals that may be mentioned are humans. Thus, in particular embodiments of the first to seventh aspects of the invention, the pregnant mammal is a pregnant woman.

"Treating" and "preventing"

When used herein, and unless otherwise specified, the term "treating" includes references to abolishing, or reducing the severity of, symptoms of the disease, disorder or condition in question.

Similarly, when used herein, the term "preventing" includes references to: preventing the disease, disorder or condition in question (or symptoms of that disease, disorder or condition) from developing (or recurring); or

reducing the severity of the symptoms of that disease, disorder or condition that develop.

Diagnosis and treatment of oxidative stress in the placenta

Oxidative stress in a cell occurs when the cell is unable to readily detoxify reactive oxygen species (ROS, such as superoxide, singlet oxygen, peroxides and free radicals such as the hydroxyl radical) present in the cell, or to repair the damage caused by such ROS. Chemically, oxidative stress is associated with increased production of oxidising species and/or a significant decrease in the effectiveness of endogenous antioxidant defences (e.g. based upon molecules such as glutathione). Because of difficulties associated with directly measuring the oxidising molecules that are involved in oxidative stress (due to their inherent instability and/or high reactivity), the presence of oxidative stress is typically determined by reliance upon one or more biomarkers, such as the downstream products of ROSs. In this respect, methods of determining the existence of oxidative stress in the placenta of a pregnant mammal include biomarker-based methods known to those skilled in the art, such as measurement of 8-hydroxydeoxyguanosine (8-OH-dG) and/or malondialdehyde (MDA), for example in a urine sample of a pregnant mammal (see, for example, Kim, Y-J. et al., Reproductive Toxicology 2005, 19(4), 487-492, where higher concentrations of urinary 8-OH- dG and MDA are associated with pre-term births).

Thus, in certain embodiments of the first to third aspects of the invention, the pregnant mammal is a mammal exhibiting oxidative stress in the placenta, for example as determined by measurement of urinary concentrations of 8-OH-dG and/or MDA.

References herein to "treating oxidative stress in the placenta" include references to improving the ability of the cells in the placenta that are under stress to readily detoxify ROS present in those cells, or to repair the damage caused by such ROS. Such improvement may be determined, for example, by reference to the levels (e.g. as measured in a urine sample of a pregnant mammal) of relevant biomarkers of oxidative stress, such as 8-OH-dG and/or MDA, wherein a decrease in the level of such biomarkers is indicative of the treatment of oxidative stress in the placenta.

Similarly, references herein to "preventing oxidative stress in the placenta" include references to ensuring that, during the pregnancy of the mammal, cells in the placenta maintain the ability to readily detoxify ROS present in those cells, or to repair the damage caused by such ROS. Such maintenance may be determined, for example, by reference to the levels (e.g. as measured in a urine sample of a pregnant mammal) of relevant biomarkers of oxidative stress, such as 8-OH-dG and/or MDA, wherein preventing (e.g. for at least one week subsequent to administration of the nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor, such as for at least 2 weeks subsequent to that administration) a significant increase in the levels of one or more such biomarkers is indicative of the prevention of oxidative stress in the placenta. In this respect, references to "preventing a significant increase" include references to preventing an increase of greater than 50, 40, 30, 25, 20, 15 or, particularly, 10% in the levels of one or more biomarkers of oxidative stress (e.g. preventing such an increase in the levels of 8-OH-dG and/or MDA).

Measurements of levels of 8-OH-dG or MDA (e.g. for measurements performed on a urine sample, levels adjusted for creatinine) may be performed by methods known to those skilled in the art (e.g. using competitive in vitro ELISA for determination of 8-OH-dG or, for the determination of MDA, HPLC for the adduct obtained with thiobarbituric acid), for example according to the procedures outlined in Kim, Y-J. et al. , Reproductive Toxicology 2005, 19(4), 487-492.

Diagnosis and treatment of preeclampsia

Preeclampsia is a progressive disorder that can develop some severe features. Severe features of preeclampsia can (as discussed in Hypertension in Pregnancy, a report by the Task Force of the American College of Obstetricians and Gynecologists - which report is discussed in Obstetrics & Gynecology 2013, 122(5), 1 122-1 131 ) be defined as any of the following.

(1 ) Hypertension: systolic≥160 or diastolic≥1 10 on two occasions at least 4 hours apart while the patient is on bed rest (unless antihypertensive therapy is initiated before this time).

(2) Thrombocytopenia (platelet count <100,000).

(3) Impaired liver function (elevated blood levels of liver transaminases to twice the normal concentration), severe persistent right upper quadrant (RUQ) or epigastric pain unresponsive to medication and not accounted for by alternative diagnoses, or both. (4) New development of renal insufficiency (elevated serum creatinine greater than 1 .1 mg/dL, or doubling of serum creatinine in the absence of other renal disease).

(5) Pulmonary oedema.

(6) New-onset cerebral or visual disturbances

In this respect, references herein to "treating preeclampsia" include references to abolishing, or reducing the severity of, any one or more of the above-mentioned severe features (e.g. to abolishing, or reducing the severity of, all of the above-mentioned severe features that manifest in the patient).

Similarly, references herein to "preventing preeclampsia" include references to preventing any one or more of the above-mentioned severe features (e.g. to preventing all of the above- mentioned severe features). A pregnant mammal may be determined as being at risk of developing preeclampsia by using methods known to those skilled in the art, such as methods discussed in Chapter 3 ("Prediction of preeclampsia") of the above-mentioned Hypertension in Pregnancy report. Such methods include those that can be used to determine, early in the term of a pregnancy (e.g. during the first trimester), individuals who are at risk of developing preeclampsia. Particular methods that may be mentioned include uterine artery Doppler velocitometry (i.e. the detection of abnormal blood flow in the uterine artery) and measurement of levels (e.g. in the serum of a pregnant mammal) of relevant biomarkers, such as:

anti-angiogenic factors, such as soluble fms-like tyrosine kinase-1 (sFlt-1 or sVEGFR- 1) or sEng (a soluble form of Endoglin);

pro-angiogenic factors, such placental growth factor (PIGF) or VEGF; and/or placental protein-13

(see, for example: (i) Herraiz, I. et al., Int. J. Mol. Sci. 2015, 76(8), 19009-19026, where high levels of sFlt-1 , or high values of the sFlt-1/ PIGF ratio are described as being indicative of an increased risk of developing preeclampsia; (ii) Levine R.J. et al., JAMA 2005, 293(1), 77-85, where decreased urinary PIGF at mid-gestation was found to be strongly associated with subsequent early development of preeclampsia; (iii) Kusanovic, J. P. et al., J. Matern. Fetal. Neonatal. Med. 2009, 22(1 1), 1021-1038, where the PIGF/sEng ratio and its delta and slope was found to be predictive of early-onset preeclampsia; and (iv) Kenny L.C. etal., Hypertension 2010, 56, 741-749, where analysis of the plasma levels of various metabolites representing a "signature" of preeclampsia were used to predict individuals at risk of developing preeclampsia).

Embodiments of the fourth to sixth aspects of the invention that may be mentioned connection with the treatment or prevention of preeclampsia include the following.

(1 a) The pregnant mammal is human and is identified as suffering from preeclampsia by exhibiting any one or more of symptoms (1) to (6) above. (2a) The pregnant mammal is human and is identified as being at risk of developing preeclampsia:

by determining that the pregnant mammal has oxidative stress in the placenta, for example, by measurement of indicative levels (e.g. in a urine sample of the pregnant mammal) of biomarkers such as 8-OH-dG and/or MDA;

- by the detection of abnormal blood flow in the uterine artery using Doppler velocitometry measurements;

by measurement of indicative levels (e.g. in the serum of the pregnant mammal) of biomarkers such as sFlt-1 , sVEGFR-1 , sEng, PIGF, VEGF and/or placental protein- 13; or, particularly

- if determined as having one or more risk factors associated with increased incidence of preeclampsia.

Risk factors associated with increased incidence of preeclampsia are known to those skilled in the art (see, for example, Duckitt, K. and Harrington, D. BMJ 2005, 330(7491), 565) and include, for example: a previous history of pre-eclampsia; pre-existing hypertension (e.g. chronic hypertension); pre-existing antiphospholipid syndrome (e.g. as determined by the presence of anticardiolipin antibodies or lupus anticoagulant or both), pre-existing renal disease; pre-existing chronic autoimmune disease (e.g. systemic lupus erythematosus); pre-existing diabetes (e.g. insulin-dependent diabetes); multiple (e.g. twin) pregnancy; nulliparity; family history of preeclampsia; a history of thrombophilia; raised body mass index before pregnancy (e.g. body mass index >35); maternal age ≥40; an interval of 10 years or more since a previous pregnancy; and suffering from asthma symptoms during pregnancy (e.g. where the patient has pre-existing asthma).

In a particular embodiment of the fourth aspect of the invention, there is provided a method of treating preeclampsia, said method comprising administering a nanoparticulate formulation of an ATP signalling inhibitor or, particularly, an organic antioxidant to a pregnant mammal suffering from preeclampsia.

In another particular embodiment of the fourth aspect of the invention, the method comprises the steps of:

(a) determining that a pregnant mammal is suffering from preeclampsia, or is at risk of developing preeclampsia; and

(b) administering a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor to said pregnant mammal.

Step (a) of this embodiment may be carried out by using any of the diagnostic (or risk-based) criteria or diagnostic methods (e.g. based upon biomarkers) described above, such as at (1a) and (2a) above.

In yet another particular embodiment of the fourth aspect of the invention, the method comprises administering a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor to a pregnant mammal that has been determined (e.g. in accordance with embodiment (1 a) above) to be suffering from preeclampsia, or that has been determined (e.g. in accordance with embodiment (2a) above) to be at risk of developing preeclampsia.

In a still further particular embodiment of the fourth aspect of the invention, there is provided a method of preventing preeclampsia, said method comprising administering a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor to a pregnant mammal at risk of developing preeclampsia (e.g. as determined by methods described at (2a) above). In this particular embodiment, when the mammal is human, the nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor may, for example, be administered to the pregnant mammal at a time up to 38 weeks gestation (e.g. up to 34 weeks, such as from 12 to 34 weeks gestation, or from 16 or 20 to 34 or 38 weeks gestation).

The features of the above-mentioned particular embodiments of the fourth aspect of the invention are paralleled in corresponding particular embodiments of the fifth and sixth aspects of the invention that relate to treating preeclampsia and/or to determining whether the pregnant mammal is suffering from (or is at risk of developing) preeclampsia before administering the nanoparticulate formulation. Further embodiments of the fourth to sixth aspects of the invention include those in which the nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor is administered to the pregnant mammal either once or more than once.

Diagnosis and treatment of extrinsic intrauterine growth restriction

Intrauterine growth restriction (lUGR, also known as intrauterine growth retardation) is a condition in which an unborn baby is smaller than it should be, i.e. small for gestational age (SGA), because it is not growing at a normal rate inside the womb. In this respect, lUGR is also known as "pathological SGA".

As defined herein, foetuses suffering from lUGR include human foetuses whose:

(a1) estimated weight is below the 10th percentile for their gestational age; and

(b1) abdominal circumference is below the 2.5th percentile for their gestational age. Determination of gestational age, as well as of estimated weight and abdominal circumference, may be made according to procedures known to those skilled in the art (e.g. procedures involving one or more ultrasound measurements (e.g. measurements at one or more of time points within the range of 12 to 40 weeks of pregnancy, such as at one or more of 12, 20, 28, 32, 36 and 40 weeks of pregnancy), such as those described in Loughna et al., Ultrasound 2009, 17(3), 161-167). Percentiles for estimated weight (by gestational age) may, for example, be those illustrated in Figure 1 of Peleg et al., Am. Fam. Physician 1998, 58(2), 453-460. Further, percentiles for abdominal circumference (by gestational age) may, for example, be those illustrated in Appendix 3 of Loughna et al., supra. When used herein, the term "extrinsic lUGR" refers to growth restriction of the foetus caused by factors extrinsic to the foetus. In this respect, references herein to "extrinsic lUGR" specifically exclude growth restriction caused solely by factors intrinsic to the foetus (e.g. chromosomal abnormalities in the foetus, or microbial or parasitic infections in foetus). The term "extrinsic lUGR" therefore includes references to growth restriction of the foetus caused, for example, by:

poor maternal health, for example due to low maternal pre-pregnancy weight (e.g. a pre-pregnancy weight of 45 kg or less, or a pre-pregnancy body mass index of less than 20), high maternal pre-pregnancy weight (e.g. a pre-pregnancy weight of 75 kg or more, or pre- pregnancy maternal body mass index of 40 or more), poor maternal nutrition, maternal anaemia, maternal alcohol abuse, maternal drug use, maternal smoking, recent pregnancy, pre-gestational diabetes, gestational diabetes, maternal pulmonary disease, maternal microbial or parasitic infection (e.g. maternal infection with Mycobacterium tuberculosis, Treponema pallidum, rubella, cytomegalovirus, herpes simplex virus, varicella zoster virus, Parvovirus B19, Toxoplasma gondii or Plasmodium, or maternal urinary tract infection or maternal bacterial vaginosis), maternal cardiovascular disease, maternal renal disease, maternal hypertension, maternal hypercoagulation disorders (e.g. antiphospholipid syndrome) and/or maternal coeliac disease; or

- uteroplacental factors, such as preeclampsia, multiple gestation, uterine malformations or placental insufficiency (e.g. uterine malformations giving rise to, or placental insufficiency caused by, abnormal or insufficient blood flow to the placenta, for example as determined by uterine artery Doppler velocitometry measurements).

The existence (and cause) of poor maternal health, as well as the existence of the above- mentioned uteroplacental factors, may be determined by those skilled in the art using known techniques. However, a feature common to all such causes of extrinsic IUGR is a capability of inducing or increasing oxidative stress in the placenta. For example, it is known that: maternal smoking during pregnancy increases oxidative stress (see de Almeida Olympio Rua et ai, J. Biomed. Sci. 2014, 21(1), 105);

immunological responses to invading pathogens include the generation of reactive oxygen species (e.g. to kill those pathogens) which, whilst protective, can lead to oxidative stress that initiates adverse effects (see M. Pohanka, Folia Microbiol. 2013, 58, 503-513); urinary tract infections during pregnancy may aggravate oxidative stress (see Ciragil et ai, Mediators of Inflammation 2005, 5, 309-311);

increased oxidative stress has been observed in pregnant women with pre-gestational type 1 diabetes, or with gestational diabetes (see Peuchant et ai, Clin. Biochem. 2004, 37, 293-298); and

maternal malaria infections are known to induce both anaemia and oxidative stress during pregnancy (see Akanbi et ai, Asian Pacific Journal of Tropical Medicine 2010, 21 1- 214).

Pregnant mammals at risk of developing extrinsic IUGR therefore include those suffering from, or at risk of developing, poor maternal health (e.g. as caused by any one or more of the factors listed above) and/or one or more of preeclampsia, multiple gestation, uterine malformations and placental insufficiency. In this respect, pregnant mammals at risk of developing extrinsic IUGR include those identified as having one or more of the above-defined risk factors associated with increased incidence of preeclampsia. In a particular embodiment of the fourth aspect of the invention, there is provided a method of treating extrinsic IUGR, said method comprising administering a nanoparticulate formulation of an ATP signalling inhibitor or, particularly, an organic antioxidant to a pregnant mammal suffering from extrinsic IUGR. In another particular embodiment of the fourth aspect of the invention, the method comprises the steps of:

(a) determining that a pregnant mammal is suffering from extrinsic IUGR, or is at risk of developing extrinsic IUGR; and (b) administering a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor to said pregnant mammal.

In this embodiment of the fourth aspect of the invention, determining that a pregnant mammal is suffering from extrinsic IUGR comprises, when the mammal is human, the step of determining whether the foetus of the pregnant mammal meets criteria (a1) and (b1) above.

In yet another particular embodiment of the fourth aspect of the invention, the method comprises administering a nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor to a pregnant mammal that has been determined (e.g. for a human mammal, by assessment of whether the foetus meets criteria (a1) and (b1) above) to be suffering from extrinsic IUGR, or to be at risk of developing extrinsic IUGR.

The features of the above-mentioned particular embodiments of the fourth aspect of the invention are paralleled in corresponding particular embodiments of the fifth and sixth aspects of the invention that relate to treating extrinsic IUGR and/or to determining whether the pregnant mammal is suffering from (or is at risk of developing) extrinsic IUGR before administering the nanoparticulate formulation. Further embodiments of the fourth to sixth aspects of the invention include those in which the nanoparticulate formulation of an organic antioxidant or an ATP signalling inhibitor is administered to the pregnant mammal either once or more than once.

Preventing preterm birth and/or increasing birth weight

References herein to "preventing preterm birth" include references to reducing the number of days that a birth is preterm. In this respect, references herein to "preterm birth" include references to birth in humans before 37 weeks. Further, references to "increasing birth weight" include reference to preventing low birth weight. In this respect, the term "low birth weight" includes references to:

a birth weight of less than 2,500 g (e.g. less than 1 ,500 g), regardless of gestational age; or

70% or less (e.g. 65, 60, 55 or 50% or less) of the average birth weight for the same gestational age.

Embodiments of the seventh aspect of the invention that may be mentioned include those in which the pregnant mammal is human and is identified as being at risk, or has been determined to be at risk, of developing preeclampsia (e.g. as determined by methods described at (2a) above).

Neurodevelopmental disorders When used herein, the term "neurodevelopmental disorder" includes references to disorders involving an impairment of the growth and development of the brain or central nervous system. Particular neurodevelopmental disorders that may be mentioned include schizophrenia, autism, attention deficit/hyperactivity disorder and/or disorders associated with reduced dendritic length, loss of dendrites, loss of glutamate receptors, loss of N-methyl-D-aspartate receptors and/or increase of dopamine receptors.

The nanoparticulate formulation The formulation employed in the methods according to the first to seventh aspects of the invention is a pharmaceutical formulation (i.e. a formulation suitable for pharmaceutical use). That formulation comprises nanoparticles, which nanoparticles comprise or consist of the antioxidant or the ATP signalling inhibitor. Nanoparticles comprising the antioxidant or ATP signalling inhibitor include, for example, nanoparticles formed from polymers, wherein the antioxidant or ATP signalling inhibitor is encapsulated by and/or adsorbed to the polymeric nanoparticles. In particular embodiments of the invention that may be mentioned, the nanoparticulate formulation of the antioxidant or ATP signalling inhibitor comprises a formulation in which the antioxidant or ATP signalling inhibitor is adsorbed to polymeric nanoparticles.

When used herein, the term "nanoparticles", refers to particles less than 1000 nm in size, e.g. particles having a Z-average size of less than 1000 nm. Where Z-average particle sizes are defined herein, those sizes may conveniently be determined by determined by dynamic light scattering, for example using a Zetasizer Nano ZS from Malvern Instruments.

In particular embodiments of the first to seventh aspects of the invention, the nanoparticulate formulation comprises or consists of nanoparticles (which nanoparticles consist of or, particularly, comprise the organic antioxidant or ATP signalling inhibitor) having a Z-average size, as determined by dynamic light scattering, of from 1 to 900 nm, such as from 100 to 500 nm or, particularly, from 150 to 300 nm. Where the nanoparticulate formulation comprises other particulate materials, those particulate materials are prepared separately from the nanoparticles comprising or consisting of the organic antioxidant or ATP signalling inhibitor. For example, those other particulate materials may be materials or mixtures (e.g. pharmaceutical excipients) that do not comprise or consist of the organic antioxidant or ATP signalling inhibitor. In such embodiments, the nanoparticulate formulation may, for example, be a pharmaceutical formulation according to the tenth aspect of the invention (as described below), in which component (b) of the formulation includes one or more particulate materials. Nanoparticles comprising or consisting of the antioxidant or ATP signalling inhibitor may be prepared according to or by analogy with techniques known to those skilled in the art, for example: for nanoparticles consisting of the antioxidant or ATP signalling inhibitor, by particle size reduction (e.g. grinding or milling) techniques, or by techniques involving emulsification followed by solidification or precipitation (see, for example, Pharm. Res. 1995, 72, 201-208); for compositions in which the antioxidant or ATP signalling inhibitor is encapsulated by polymeric nanoparticles, formation of the polymeric nanoparticles in the presence of the antioxidant or ATP signalling inhibitor, wherein the nanoparticles are formed using known techniques, such as nanoprecipitation, emulsification-diffusion, emulsification-coacervation, multiple emulsions and supercritical fluid technology (see, for example, Reis, CP. et al., Nanomedicine: Nanotechnology, Biology and Medicine 2006, 2(1), 8-21); and

- for compositions in which the antioxidant or ATP signalling inhibitor is adsorbed to polymeric nanoparticles (e.g. by hydrophobic and/or electrostatic interactions), by admixing pre-formed polymeric nanoparticles with the antioxidant or ATP signalling inhibitor, optionally in the presence of a solvent (e.g. an organic solvent such as dimethylsulfoxide, or an aqueous solvent system such as water or an aqueous solution of inorganic salts (e.g. a NaCI or phosphate-buffered saline solution)).

The polymeric nanoparticles may be formed from any suitable polymer, such as a biodegradable or bioerodible polymer (e.g. a polymer based upon lactic acid, glycolic acid and/or other hydroxyalkanoic acids) or, in particular, a poly(amino acid).

Poly(amino acid) polymers that may be mentioned include those based upon polymers of glutamic acid, for example (block) copolymers of poly(y-glutamic acid) and a hydrophobic amino acid such as Leu, Trp or, particularly Phe, or a Ci _-4 alkyl ester thereof (e.g. L-leucine methyl ester, L-tryptophan methyl ester or, particularly, L-phenylalanine ethyl ester). Nanoparticles of such copolymers may conveniently be prepared according to known techniques (see, for example, Kim H. et al., Macromol. Biosci. 2009, 9, 842-848).

Particular poly(amino acid) polymers that may be mentioned include those in which one or more (e.g. all) of the following apply:

(1 b) the polymer is a copolymer formed by reaction of poly(y-glutamic acid) with a Ci _-4 alkyl ester of a hydrophobic amino acid and a peptide coupling reagent, such as 1 or more molar equivalents (e.g. 1 molar equivalent, relative to the poly(y-glutamic acid)) of a carbodiimide (e.g. 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide);

(2b) the poly(y-glutamic acid) used to form the copolymer comprises both D- and L-y- glutamic acid residues (e.g. wherein the D/L ratio of the polymer starting material is from 1 :5 to 5: 1 , such as about 6:4 or, particularly, about 7:3);

(3b) the poly(y-glutamic acid) used to form the copolymer has a weight average molecular weight of from 150 to 700 kDa, such as 200 to 500 kDa or, particularly, 350 to 400 kDa (or the average number of y-glutamic acid monomers in the poly(y-glutamic acid) used to form the copolymer is from 1 ,000 to 5,000, such as from 2,000 to 3,000);

(4b) the Ci _-4 alkyl ester of a hydrophobic amino acid is a C1-2 alkyl ester of Leu, Trp or, particularly Phe (e.g. a C1-2 alkyl ester of L-Leu, L-Trp or, particularly L-Phe, such as L- phenylalanine ethyl ester); (5b) the polymer comprises a mixture of poly(y-glutamic acid) and poly(y-glutamic acid) coupled to the hydrophobic amino acid or C1-4 alkyl ester thereof, wherein the percentage of poly(y-glutamic acid) molecules that are coupled to the hydrophobic amino acid or C1-4 alkyl ester thereof is from 5 to 70%, such as from 40 to 60% or, particularly, about 50%);

(6b) the polymer is formed by reaction of 1 molar equivalent of the poly(y-glutamic acid) with 0.8 to 1.2 molar equivalents (e.g. 1.0 molar equivalent) of the hydrophobic amino acid (or a C1-4 alkyl ester thereof);

(7b) the polymer is formed by reaction at sub-ambient temperature (e.g. 1 to 10°C, such as 4°C);

(8b) the polymer is formed by reaction in the presence of an aqueous solvent system (e.g. an aqueous solvent system having a pH of greater than 7.0, such as a 0.3 N solution of NaHCOs);

(9b) formation of the polymer is followed by dialysis of the reaction mixture (or of an aqueous mixture containing the reaction products) using a membrane designed to permit transit of low molecular weight (e.g. 50,000 Da or less) molecules.

Aqueous suspensions of nanoparticles may be prepared using the above-mentioned poly(y- glutamic acid) / hydrophobic amino acid copolymers, for example, by a method involving the steps of:

(1 c) dissolving the copolymer in a polar aprotic solvent (e.g. dimethylsulfoxide (DMSO)), for example to form a solution concentration of 10 mg copolymer per ml_;

(2c) adding an equivalent volume of a NaCI solution (e.g. a 0.15 M NaCI solution); and (3c) removing the polar aprotic solvent from the resulting mixture (e.g. by dialysing against distilled water).

The size of the nanoparticles prepared according to such methods may be controlled, for example, by adjusting the ionic strength of the solution added in step (2c). Further, the ionic strength required to achieve a particular (average) nanoparticle size may vary, depending upon the identity (and percentage incorporation of) the hydrophobic amino acid present in the copolymer. Typically, a higher degree of incorporation of hydrophobic amino acid into the copolymer translates into a lower ionic strength (e.g. lower concentration of NaCI) required to reach a selected (average) nanoparticle size.

The aqueous suspensions of nanoparticles so prepared may, if desired, be freeze-dried (e.g. to be later reconstituted in an aqueous solvent system, such as phosphate-buffered saline (PBS, e.g. 10 mg/ml_ PBS)).

The nanoparticulate formulation may be administered to the mammal by any suitable route, including via parenteral, oral or mucosal routes.

A particular route of administration that may be mentioned is parenteral (e.g. intravenous) administration. Thus, in specific embodiments of the first to seventh aspects of the invention, the nanoparticulate formulation is administered intravenously (and/or is formulated for intravenous administration).

The medical practitioner, or other skilled person, will be able to determine a suitable dosage for the nanoparticulate formulation, and hence the amount of the organic antioxidant or ATP signalling inhibitor that should be included in any particular pharmaceutical formulation (whether in unit dosage form or otherwise).

The organic antioxidant

As used herein, the term "antioxidant" refers to an organic molecule that inhibits the oxidation of other molecules. Further, the term "organic" includes references to molecules containing one or more carbon-hydrogen bonds.

Particular antioxidants that may be mentioned include molecules that:

have a molar mass of less than or equal to 1 ,500 Da, such as less than 1 ,000 Da or, particularly, less than 750 Da; and/or

comprise a phenol, 1 ,4-benzoquinone, 1 ,4-dihydroxybenzene, N-oxide or thiol moiety. Antioxidants that may be mentioned in connection with the first to seventh aspects of the invention include:

(1 d) endogenous antioxidants, such as Vitamin C (ascorbic acid), Vitamin E (a-tocopherol), creatinine, CoQio, urate, apocynin and diapocynin;

(2d) neutral synthetic compounds, such as compounds based upon phenol (e.g. butylated hydroxytoluene), 1 ,4-benzoquinone (e.g. idebenone), or N-oxide (e.g. 2,2,6,6- tetramethylpiperidin-1-yl)oxyl (TEMPO), 4-hydroxy-TEMPO or phenyl-a-te/f-butyl nitrone); and

(3d) mitochondrially-targeted antioxidants, for example synthetic compounds comprising a phenol, 1 ,4-benzoquinone, 1 ,4-dihydroxybenzene, N-oxide or thiol moiety that is covalently bonded to a lipophilic cation.

In particular embodiments of the first to seventh aspects of the invention, the organic antioxidant is a compound comprising a phenol, 1 ,4-benzoquinone, 1 ,4-dihydroxybenzene, N- oxide or thiol moiety that is covalently bonded to a lipophilic cation, wherein:

(1 e) the lipophilic cation is an ammonium or, particularly, a phosphonium cation (for example, a trialkylphosphonium, tribenzylphosphonium or, particularly, a triarylphosphonium cation such as a triphenylphosphonium cation);

(2e) the phenol, 1 ,4-benzoquinone, 1 ,4-dihydroxybenzene, N-oxide or thiol moiety is covalently bonded to the lipophilic cation via a linker (for example, an alkylene linker, such as a CMS alkylene linker or, particularly, a -(CH2)i-is- linker such as -(CH2)io-);

(3e) the 1 ,4-benzoquinone moiety is a 2,3,5-trisubstituted benzoquinone that is connected to the lipophilic cation via the 6-position of the ring, wherein the substituents at the 2-, 3- and 5-positions of the ring are selected from alkyl (e.g. methyl) and alkoxy (e.g. methoxy), for example to form a 2,3-dimethoxy-5-methyl-1 ,4-benzoquinon-6-yl moiety; and/or

the counterion to the lipophilic cation is a non-nucleophilic or weakly nucleophilic anion, such as tetrafluoroborate, hexafluorophosphate or, particularly, an alkanesulfonate such as trifluoromethanesulfonate or, particularly, methanesulfonate.

In more particular embodiments of the first to seventh aspects of the invention, the organic antioxidant is mitoquinone (MitoQ),

which has the chemical name [10-(4,5-dimethoxy-2-methyl-3,6-dioxo-1 ,4-cyclohexadien- 1- yl)decyl]triphenylphosphonium, methanesulfonate.

The above-mentioned antioxidants are either commercially available or may be prepared by methods known to those skilled in the art (e.g. for certain mitochondrially-targeted antioxidants, by methods as described in any of WO 99/26582, WO 99/26954, WO 2005/019232 and WO 2005/019233, the disclosures of which documents are hereby incorporated by reference).

Thus, particular embodiments of the first to seventh aspects of the invention that may be mentioned include those in which the organic antioxidant or an ATP signalling inhibitor is an organic antioxidant such as mitoquinone. In this respect, particular embodiments that may be mentioned include the following.

A method of treating or preventing oxidative stress in the placenta of a pregnant mammal, said method comprising administering a nanoparticulate formulation of mitoquinone to said pregnant mammal.

A nanoparticulate formulation of mitoquinone for use in treating or preventing oxidative stress in the placenta of a pregnant mammal.

The use of a nanoparticulate formulation of mitoquinone for the manufacture of a medicament for treating or preventing oxidative stress in the placenta of a pregnant mammal.

A method of treating or preventing preeclampsia or extrinsic IUGR, said method comprising administering a nanoparticulate formulation of mitoquinone to a pregnant mammal suffering from preeclampsia or extrinsic IUGR, or at risk of developing preeclampsia or extrinsic IUGR.

A nanoparticulate formulation of mitoquinone for use in treating or preventing preeclampsia or extrinsic IUGR in a pregnant mammal, wherein said pregnant mammal is suffering from preeclampsia or extrinsic IUGR, or at risk of developing preeclampsia or extrinsic IUGR.

The use of a nanoparticulate formulation of mitoquinone for the manufacture of a medicament for treating or preventing preeclampsia or extrinsic IUGR in a pregnant mammal, wherein said pregnant mammal is suffering from preeclampsia or extrinsic IUGR, or at risk of developing preeclampsia or extrinsic IUGR.

A method of

preventing preterm birth and/or of increasing birth weight, or

preventing, or reducing the severity of, neurodevelopmental disorders in the foetus,

said method comprising administering a nanoparticulate formulation of mitoquinone to a pregnant mammal.

In embodiments (1f) to (7f) above, the nanoparticulate formulation of mitoquinone may, for example, comprise mitoquinone adsorbed to and/or encapsulated by polymeric nanoparticles. In such embodiments, the mitoquinone may, in particular, be adsorbed to nanoparticles formed from a polymer of glutamic acid (e.g. a (block) copolymer of poly(y-glutamic acid) and a hydrophobic amino acid such as Leu, Trp or, particularly Phe, or a C1 -4 alkyl ester thereof (e.g. L-leucine methyl ester, L-tryptophan methyl ester or, particularly, L-phenylalanine ethyl ester), such as a poly(amino acid) polymer for which one or more (e.g. all) of (1 b) to (9b) above apply), for example using a process having steps (1d) to (3d) above.

The ATP signalling inhibitor

When used herein, the term "ATP signalling inhibitor" includes references to compounds that:

(I) block pannexin or gap junction hemichannel signalling; or

(II) block an ATP receptor.

Compounds under category (I) above include ¹⁰ Panx1 (Trp-Arg-Gln-Ala-Ala-Phe-Val-Asp-Ser- Tyr), scrambled ¹⁰ Panx1 (Phe-Ser-Val-Tyr-Trp-Ala-Gln-Ala-Asp-Arg), carbenoxolone disodium ((3β,20β)-3-(3-carboxy-1-oxopropoxy)-1 1-oxoolean-12-en-29-oic acid disodium), Gap 19 (Lys-Gln-lle-Glu-lle-Lys-Lys-Phe-Lys) or, particularly, Gap 26 (Val-Cys-Tyr-Asp-Lys-Ser-Phe- Pro-lle-Ser-His-Val-Arg) or Gap 27 (Ser-Arg-Pro-Thr-Glu-Lys-Thr-lle-Phe-lle-lle), which compounds are commercially available (e.g. from Tocris Bioscience) but may also be prepared by methods known to those skilled in the art (see also, for example: Basic Res Cardiol. 2013, 108(1), 309; Am. J. Physiol. - Cell Physiol. 2007, 293(3), C11 12-C1 119; J. Biol. Chem. 2007, 282, 2386; and Biochem. Soc. Trans. 2001 , 29(4), 606-612).

Compounds under category (II) above include compounds that block (e.g. inhibit or antagonise) purinergic receptors (e.g. P2 receptors, such as P2X receptors). Such compounds include, for example, pyridoxalphosphate-6-azophenyl-2',4'-disulfonic acid (otherwise known as PPADS) and compounds disclosed in WO 03/080579, the disclosures of which are hereby incorporated by reference. Exemplary compounds disclosed in WO 03/080579 include Examples 17 and 158 of that document, namely 2-(1-adamantyl)-N-[2-({2- [(2-hydroxyethyl)amino]ethyl}amino)quinolin-5-yl]acetarnide and 2-(1-adamantyl)-N-[2-({2- [bis(2-hydroxyethyl)amino]ethyl}amino)quinolin-5-yl]acetarTi ide, respectively (otherwise known as Compound 17 and GSK1271360, respectively). Nanoparticulate Formulations

The invention also relates to novel nanoparticulate formulations described herein.

In this respect, in an eighth aspect of the invention there is provided a nanoparticulate formulation comprising a mitochondrially-targeted antioxidant or ATP signalling inhibitor.

Similarly, in a ninth aspect of the invention, there is provided a nanoparticulate formulation comprising a mitochondrially-targeted antioxidant or ATP signalling inhibitor for use in medicine.

Further, in a tenth aspect of the invention, there is provided a pharmaceutical formulation comprising:

(a) a nanoparticulate formulation comprising a mitochondrially-targeted antioxidant or ATP signalling inhibitor; and

(b) a pharmaceutically acceptable adjuvant, diluent or carrier.

In an eleventh aspect of the invention there is provided a combination product comprising: (A) a nanoparticulate formulation comprising a mitochondrially-targeted antioxidant or ATP signalling inhibitor; and

(B) another therapeutic agent,

wherein each of components (A) and (B) is optionally formulated in admixture with a pharmaceutically-acceptable adjuvant, diluent or carrier.

Embodiments of the eighth to eleventh aspects of the invention that may be mentioned include those in which the nanoparticulate formulation is a nanoparticulate formulation as described above, wherein the organic antioxidant or ATP signalling inhibitor is a mitochondrially-targeted antioxidant or ATP signalling inhibitor.

The mitochondrially-targeted antioxidant

Embodiments of the eighth to eleventh aspects of the invention that may be mentioned include those in which the mitochondrially-targeted antioxidant comprises a phenol, 1 ,4- benzoquinone, 1 ,4-dihydroxybenzene, N-oxide or thiol moiety that is covalently bonded to a lipophilic cation.

Such mitochondrially-targeted antioxidants include the molecules described and claimed in any of WO 99/26582, WO 99/26954, WO 2005/019232 and WO 2005/019233, the disclosures of which documents are hereby incorporated by reference. In this respect, embodiments of such mitochondrially-targeted antioxidants in connection with the eighth to eleventh aspects of the invention include those described in the section above relating to the organic antioxidant (see, in particular, points (1 e), (2e), (3e) and/or (4e) of that section).

In a particular embodiment, the mitochondrially-targeted antioxidant is mitoquinone. Thus, particular embodiments of the eighth to eleventh aspects of the invention that may be mentioned include those in which the organic antioxidant or an ATP signalling inhibitor is a mitochondrially-targeted antioxidant, such as mitoquinone.

The nanoparticulate formulation

The formulation of the eighth to eleventh aspect of the invention comprises nanoparticles, which nanoparticles comprise or consist of the mitochondrially-targeted antioxidant or ATP signalling inhibitor as described above.

In particular embodiments, the formulation of the eighth to eleventh aspect of the invention is a pharmaceutical formulation (i.e. a formulation suitable for pharmaceutical use). Nanoparticles comprising the mitochondrially-targeted antioxidant or ATP signalling inhibitor include, for example, nanoparticles formed from polymers, wherein the antioxidant or ATP signalling inhibitor is encapsulated by and/or adsorbed to the polymeric nanoparticles. In particular embodiments of the invention that may be mentioned, the nanoparticulate formulation of the mitochondrially-targeted antioxidant or ATP signalling inhibitor comprises a formulation in which the mitochondrially-targeted antioxidant or ATP signalling inhibitor is adsorbed to polymeric nanoparticles.

Particular nanoparticulate formulations that may be mentioned include those in which the antioxidant or ATP signalling inhibitor is encapsulated by and/or adsorbed to polymeric nanoparticles and the resulting, drug-loaded nanoparticles, when suspended at a concentration of 10 mg/mL in phosphate-buffered saline (or, alternatively, in water) maintained at room temperature (e.g. 25°C), release from 5 to 40% (e.g. from 8 to 35%, such as from 10 to 20%) of the drug (e.g. Gap26 or, particularly, mitoquinone) from the nanoparticles over a period of 24 hours. In such embodiments, the amount of drug released may be quantified, for example, using either the Lowry protein assay (for peptide-based drugs) and/or uv absorption.

If present, the polymer forming nanoparticles that encapsulate the antioxidant or ATP signalling inhibitor, and/or to which the antioxidant or ATP signalling inhibitor is adsorbed (e.g. using a process having steps (1 d) to (3d) above), may be any of the polymers described above in connection with the first to seventh aspects of the invention (e.g. a (block) copolymer of poly(y- glutamic acid) and a hydrophobic amino acid such as Leu, Trp or, particularly Phe, or a Ci _-4 alkyl ester thereof (e.g. L-leucine methyl ester, L-tryptophan methyl ester or, particularly, L- phenylalanine ethyl ester), such as a poly(amino acid) polymer for which one or more (e.g. all) of (1 b) to (9b) above apply). In particular embodiments of the eighth to eleventh aspect of the invention, the nanoparticulate formulation comprises nanoparticles having a Z-average size, as determined by dynamic light scattering, of from 1 to 900 nm, such as from 100 to 500 nm or, particularly, from 150 to 300 nm (e.g. the nanoparticulate formulation comprises particles that comprise or consist of the mitochondrially-targeted antioxidant or ATP signalling inhibitor and that have a Z-average size, as determined by dynamic light scattering, of from 1 to 900 nm, such as from 100 to 500 nm or, particularly, from 150 to 300 nm).

The invention also provides for a method of making a nanoparticulate formulation comprising a mitochondrially-targeted antioxidant or ATP signalling inhibitor and a polymer, said method comprising contacting (e.g. mixing in an aqueous (e.g. saline and/or phosphate-buffered saline) suspension) the mitochondrially-targeted antioxidant or ATP signalling inhibitor with polymeric nanoparticles (e.g. nanoparticles formed from a poly(amino acid) polymer to which any one or more of (1 b) to (9b) above apply, such as nanoparticles formed from such a polymer and obtained (or obtainable) by a method having steps (1 c) to (3c) above). The invention also relates to nanoparticulate formulations comprising a mitochondrially-targeted antioxidant or ATP signalling inhibitor and a polymer, which formulations are obtainable by such a method.

The nanoparticulate formulation of the eighth to eleventh aspects of the invention may be formulated for any form of administration (e.g. parenteral, oral or mucosal administration).

A particular route of administration that may be mentioned is parenteral (e.g. intravenous) administration. Thus, in specific embodiments of the eighth to eleventh aspects of the invention, the nanoparticulate formulation is formulated for intravenous administration (e.g. as a pyrogen-free aqueous suspension in a solvent system based upon sterile water, such as saline, or as a pyrogen-free dry powder for reconstitution as a suspension in such a solvent system).

Pharmaceutical formulations comprising the nanoparticulate formulation of the eighth aspect of the invention may be prepared by any of the methods well-known in the pharmaceutical art, for example as described in Remington's Pharmaceutical Sciences, 17th ed., Mack Publishing Company, Easton, PA., (1985).

The combination product

The combination product of the eleventh aspect of the invention may be either a single (combination) pharmaceutical formulation or a kit-of-parts.

Thus, this aspect of the invention encompasses a pharmaceutical formulation including a nanoparticulate formulation comprising a mitochondrially-targeted antioxidant or ATP signalling inhibitor, as hereinbefore defined, and another therapeutic agent, in admixture with a pharmaceutically acceptable adjuvant, diluent or carrier (which formulation is hereinafter referred to as a "combined preparation"). It also encompasses a kit of parts comprising components:

(i) a pharmaceutical formulation including a nanoparticulate formulation comprising a mitochondrially-targeted antioxidant or ATP signalling inhibitor, as hereinbefore defined, in admixture with a pharmaceutically acceptable adjuvant, diluent or carrier; and

(ii) a pharmaceutical formulation including another therapeutic agent, in admixture with a pharmaceutically-acceptable adjuvant, diluent or carrier,

which components (i) and (ii) are each provided in a form that is suitable for administration in conjunction with the other.

Component (i) of the kit of parts is thus component (A) above in admixture with a pharmaceutically acceptable adjuvant, diluent or carrier. Similarly, component (ii) is component (B) above in admixture with a pharmaceutically acceptable adjuvant, diluent or carrier.

Combination products that may be particularly useful (e.g. for the treatment or prevention of extrinsic IUGR) include those in which component (B) (or component (ii) of the kit of parts) comprises one or more therapeutic agents selected from agents known to those skilled in the art to improve maternal health (e.g. that combat poor maternal health having any one or more of the causes listed above in connection with extrinsic IUGR). In particular embodiments of the invention that may be mentioned, component (B) (or component (ii)) comprises another therapeutic agent that is an antihypertensive agent, such as:

an angiotensin II receptor antagonist (a sartan, such as azilsartan, candesartan, eprosartan, fimasartan, irbesartan, losartan, olmesartan, telmisartan, valsartan or 2-butyl-5- chloro-3-[[4-[2-(2H-tetrazol-5-yl)phenyl]phenyl]methyl]imida zole-4-carboxylic acid (otherwise known as EXP 3174));

an ACE inhibitor (such as benazepril, captopril, cilazapril, enalapril, fosinopril, imidapril, lisinopril, perindopril, quinapril, ramipril, trandolapril, zofenopril or arfalasin);

a calcium channel blocker (such as amlodipine, aranidipine, azelnidipine, barnidipine, benidipine, cilnidipine, clevidipine, isradipine, efonidipine, felodipine, lacidipine, lercanidipine, manidipine, nicardipine, nifedipine, nilvadipine, nimodipine, nisoldipine, nitrendipine, pranidipine, diltiazem, bepridil, gallopamil, mibefradil, verapamil, flunarizine, fluspirilene or fendiline);

an adrenergic receptor antagonist (e.g. an alpha and/or beta blocker such as atenolol, carvedilol, labetalol, metoprolol, nadolol, nebivolol, oxprenolol, pindolol, propranolol, timolol, phentolamine, indoramin, phenoxybenzamine, prazosin, terazosin or tolazoline);

an alpha-2 adrenergic receptor agonist (such as clonidine, guanabenz, guanfacine, methyldopa or moxonidine);

a renin inhibitor (such as aliskiren);

- an aldosterone receptor antagonist (such as eplerenone or spironolactone); and/or a diuretic (e.g. a thiazide diuretic, such as bendroflumethiazide, benzothiadiazine, chlorothiazide, epitizide or hydrochlorothiazide). The nanoparticulate formulations of the invention may have the advantage that they: deliver the mitochondrially-targeted antioxidant or ATP signalling inhibitor to the cells of the placenta; do not cross the placental membrane; and/or compared to known formulations of the same antioxidant or ATP signalling inhibitor, generate low plasma levels of the antioxidant or ATP signalling inhibitor. As such, the use of nanoparticulate formulations of antioxidants or ATP signalling inhibitors may provide the advantage of an increased margin of safety for the foetus (an hence an increased therapeutic window for the antioxidant or ATP signalling inhibitor).

The aspects of the invention described herein (e.g. the above-mentioned formulations, combinations, methods and uses) may have the advantage that, in the treatment of the conditions described herein, they may be more convenient for the physician and/or patient than, be more efficacious than, be less toxic than, have better selectivity over, have a broader range of activity than, be more potent than, produce more sustained therapeutic effects than, produce fewer side effects than, have a better pharmacokinetic and/or pharmacodynamic profile than, have more suitable solid state morphology than, have better long term stability than, or may have other useful pharmacological properties over similar formulations, combinations, methods (treatments) or uses known in the prior art for use in the treatment of those conditions or otherwise. Brief description of the drawings

Figs. 1a to 1f, as well as Figs. 2a to 2c relate to characterisation of drug delivery nanoparticles.

Fig. 1 a shows adsorption efficiency of drug to particle (total adsorbed drug weight to NPs/initial feeding amount of drug weight) ^χ 100

Fig. 1 b shows amount of drug loaded to nanoparticle (drug weight ^g)/NP weight (1 mg))

Figs. 1 c and 1d show rate of absorption of Gap 26 (c) and MitoQ (d) with time. Figs. 1e and 1f show release with time of Gap 26 (e) and MitoQ (f) from nanoparticles into PBS.

Figs. 2a and 2b show confocal images of bilayered BeWo barriers after 4 hour

(a) and 24 hour (b) exposure (above the barrier) to the highest dose of NPs tested in vitro (2 mgs/mL).

Fig 2c shows amount of flouresence detected in the tissue culture medium below the bilayered BeWo barriers 24 hours after exposure above the barrier to different doses of nanoparticles.

Figs. 3a to 3d and 4a to 4f relate to effects of drug delivery nanoparticles on the secretions from model trophoblast barriers in tissue culture.

Figs. 3a to 3d show levels of DNA damage in fibroblasts, as recorded by the alkaline comet assay, after 24 hour exposure to conditioned media below bilayered

BeWo barriers (Figs 3a, 3d) or bilayered primary human trophoblast barriers (Fig. 3c) or around explants of human placenta (Fig. 3b) after 24 hour exposure to altered oxygen (Figs. 3a to 3c, 2-12%; Fig. 3d, 2-21 %) with or without nanoparticle exposure with or without drug and compared to values for normal tissue culture medium (C) or that conditioned at 21 % oxygen without change (21 %).

Figs. 4a to 4d show length of dendrites in dissociated cortical neurones 6 days after comparable exposure to conditioned media below or around bilayered BeWo barriers (Figs. 4a, 4d) bilayered primary human cytotrophobalst barriers (Fig. 4b) or explants of human placenta (Fig. 4c) which were exposed for 24 hours to altered oxygen (2-12%) (e-g) or to 30 μΜ hydroquinone 30 μΜ benzoquinone with and without nanoparticles with and without drug.

Fig. 4e shows DNA damage in fibroblasts 24 hours after direct exposure to 30 μΜ hydroquinone 30 μΜ benzoquinone with and without nanoparticles with and without drug.

Fig. 4f shows dendrite lengths in dissociated cortical neurones 6 days after direct exposure to nanoparticles with and without drug. Figs. 5a to 5c, 6a to 6c, 7a to 7d, 8a to 8g relate to the effect of oxidative stress and mitoquinone nanoparticle injection on the rat placenta.

Fig. 5a shows a confocal image of rat placenta after maternal injection with saline, counterstained with DAPI (the brighter patches). A noteworthy feature of that image is the low level of staining overall.

Fig. 5b shows a confocal images of rat placenta after maternal injection with fluorescent nanoparticles. A noteworthy feature of that image is the considerable increase (relative to the image of Fig. 5a) in staining due to the presence of the fluorescent nanoparticles inside the cells of the placenta.

Fig. 5c is an expanded view of a section of the image shown in Fig. 5b, showing the (bright) nanoparticles inside the cells.

Fig. 6a shows a confocal images of brain (cortex) cells of rat foetus after maternal injection with saline.

Fig. 6b shows a confocal images of brain (cortex) cells of rat foetus after maternal injection with fluorescent nanoparticles (not seen), counterstained with DAPI (the brighter patches; a noteworthy feature being the difference in staining compared to the images of Figs. 5b and 5c, wherein the intensity in Fig. 6b is much lower due to the lack of the fluorescent nanoparticles).

Fig. 6c is an expanded view of a section of the image shown in Fig. 6b. A noteworthy feature of that image is the fact that no fluorescent nanoparticles are seen inside the cells (in contrast with the image of Fig. 5c).

Figs. 6d and 6e show, respectively, confocal images of foetal liver cells at GD16 after maternal injection with saline (Fig. 6d) orfluorescent NPs (Fig. 6e), counterstained with DAPI.

Figs. 7a to 7d show light micrographs showing the gross histology of the placenta (H&E stain) at GD20 under in vivo normoxia (Figs. 7a, 7c) or hypoxia (Figs.

7b, 7d) with maternal saline (Figs. 7a, 7b) or MitoQ-NP injection (Figs. 7c, 7d).

Fig. 8a shows total area of blood vessels per field of view.

Fig. 8b shows mean diameter of vessels in the GD20 rat placenta. Figs. 8c to 8g show levels of fluorescent dichlorofluorescein (DCF) in foetal brain (Fig. 8c) and liver (Fig. 8d), maternal brain (Fig. 8e) and liver (Fig. 8f) and placenta (Fig. 8g) after exposure to altered oxygen in vivo, with or without maternal MitoQ-NP injection (biological replicates: foetal and placental samples, n = 6; maternal samples, n = 3). *p < 0.05, ***p < 0.001.

Figs. 9a to 9h and 10a to 10c relate to analysis of secreted molecules from the placental barrier.

Figs. 9a to 9c show Bioanalyzer concentrations of small RNAs and microRNAs measured in conditioned media from below BeWo barriers (Fig. 9a), from rat placentae (Fig. 9b) and from human placental explants (Fig. 9c). Cells and tissues were exposed to varying oxygen conditions in vivo (M(21 %), M(11 %)) and in vitro (21 %, 2%, 2-12%).

Fig. 9d shows concentrations of total amino acids conditioned media from rat placenta tissue under varying oxygen conditions in vivo and ex vivo.

Figs. 10a to 10c show levels of BMPs in conditioned media that had been exposed to varying oxygen levels (21 %, 2%, 2-12%) in vitro either below BeWo barriers (Figs. 10a, 10b) or surrounding placentae collected from rats exposed to varying levels of maternal oxygen in vivo (Fig. 10c) is shown. Total BMP levels and levels of BMP4, BMP6 or BMP9 were estimated by inhibiting specific BMPs through antibodies and measuring relative luciferase activity using a BMP bioassay (Fig. 10a). BMP2 levels were measured via ELISA (Figs. 10b, 10c).

Figs. 1 1a, 1 1 b, 1 1c, 12 and 13a to 13j relate to the effect of indirect hypoxia and hypoxia- reoxygenation on foetal growth and the foetal brain.

Figs. 11 a and 1 1 b show birth weight (Fig. 11 a) and the body weight at P30 (Fig. 1 1 b) of rats that in utero had been exposed to either normoxia or maternal hypoxia.

Fig. 1 1 c shows placenta weight (n = 6 litters) from rats exposed to maternal normoxia, M(21 %), or hypoxia, M(1 1 %), preceded by maternal administration of saline or MitoQ-NPs.

Fig. 12 shows concentration of amino acids transported by transport system L measured in conditioned media from rat placenta tissue that had been exposed to varying oxygen conditions both in vivo (normoxia, hypoxia) and ex vivo (21 %, 2%).

Figs. 13a to 13j show the effect of indirect hypoxia on neuronal dendrite length, which was measured in vitro by applying BeWo conditioned media that had been exposed to varying oxygen levels to rat cortical cultures (Fig. 13a). Dendrite lengths were also measured in vivo in the brains of foetuses that had been exposed to maternal hypoxia with or without MitoQ-NP injection and are shown for the reticular nucleus (Fig. 13b), the cortex (Fig. 13c) and the somatosensory cortex (Fig. 13j). Representative images of neurons in the thalamic reticular nucleus (Figs. 13d to 13f) and the somatosensory cortex (Figs. 13g to 13i) from foetuses that either had (Figs. 13e, 13f, 13h, 13i) or had not (Figs. 13d, 13g) been exposed to maternal hypoxia, either with (Figs. 13f, 13i) or without (Figs. 13d, 13e, 13g, 13h) MitoQ-NP injection. Figs. 14a to 14f, 15a to 15f, 16a to 16f, 17a to 17f and 18a to 18c relate to the effects of maternal MitoQ-NP injection analysed in vitro, ex vivo and in vivo after 6 d maternal normoxia, M(21 %), or hypoxia, M(11 %).

Figs. 14a to 14c show the effects on dendrite lengths in vitro of applying media conditioned by offspring placenta (Fig. 14b; n=15, 10,8,24) or foetal plasma (Fig. 14c; n=21 , 18,23,18), collected from the in vivo experiment, or media conditioned by BeWo bilayers exposed to altered oxygen in vitro (Fig. 14a; number of biological replicates from left to right: n=25,25, 15, 15), to cortical cultures.

Figs. 14d to 14f show the effects in vivo in offspring brains in somatosensory cortex (Fig. 14e; SSC; n=8 different brains), combined somatosensory, auditory and retrosplenial cortex (Fig. 14d; CTX; n=21 ,23,21 ,23) or thalamic reticular nucleus (Fig. 14f; TRN; n=8)

Figs. 15a to 15d show process lengths of tyrosine hydroxylase (TH)-positive neurons in vitro following exposure to BeWo conditioned medium (Fig. 15a; n=25,25, 15, 15), placenta conditioned medium (Fig. 15b; n=12,10,8, 10), plasma (Fig.

15c; n=21 ,27, 19, 19) or in neuron only-cultures following exposure to placenta conditioned medium (Fig. 15d; n=5,5,5,5).

Figs. 15e and 15f show TH ⁺ process length in vivo in CTX (Fig. 15e; n=121 , 108, 123,98) and TRN (Fig. 15f; n=63,56,59,43).

Figs. 16a to 16f show GluN1 receptor subunit staining intensity in vitro following exposure to BeWo conditioned medium (Fig. 16a; n=15, 15,8,9), placenta conditioned medium (Fig. 16b; n=10, 10, 15, 10) or plasma (Fig. 16c; n=28, 10, 17, 18), and in vivo in SSC (Fig. 16d; n=3), CTX (Fig. 16e; n=3) and retrosplenial cortex (Fig. 16f; RSC; n=3).

Figs. 17a to 17f show the ratio of astrocytes to neurons in vitro following exposure to BeWo conditioned medium (Fig. 17a; n=15, 15, 18, 12), placenta conditioned medium (Fig. 17b; n=10) or plasma (Fig. 17c; n=5) and in vivo in hippocampus (Fig. 17e; HPC, n=23, 18, 18,24), CTX (Fig. 17d; n=70,66,72,77) and TRN (Fig. 17f; n=32,40,44,28).

Figs. 18a to 18c show the density of parvalbumin (PV)-positive neurons in vivo in SSC (Fig. 18a; n=21 , 42,42,33), CTX (Fig. 18b; n=63, 126, 126,98) and TRN (Fig. 18c; n=35,51 , 55,43). *p<0.05, **p<0.01 , ***p<0.001 , using ANOVA with Tukey's test for multiple comparisons.

Examples

Material and Methods

Nanoparticle preparation

An amphiphilic copolymer of poly(Y-glutamic acid) and L-phenylalanine ethylester (γ-PGA- Phe) was synthesized as previously described in Kim H. et al., Macromol Biosci. 2009 9(9), 842-848, using a 50% Phe grafting degree.

Specifically, 303.5 mg of a poly(Y-glutamic acid) having a D:L ratio of 7:3 and an average molecular weight of 200 to 500 kDa (Wako Pure Chemical Industries Ltd., catalog no. 165- 21364) was mixed at room temperature with 50 mL of a 50 mM NaHCC>3 solution and then stirred until the polymer dissolved (which takes approximately 30 to 60 minutes). The resulting mixture was cooled (on ice) for approximately 10 minutes before 450.5 mg of 1-ethyl-3-(3- dimethylaminopropyl)carbodiimide, hydrochloride (EDC; mixed with 1 mL of 50 mM NaHCC>3 solution) was added, followed approximately 4½ minutes later by 540 mg of L-phenylalanine ethyl ester hydrochloride (Phe; mixed with 1 mL of 50 mM NaHCC>3 solution) and 488.2 mg of Alexa Fluor 488 hydrazide, sodium salt (Molecular Probes, catalog no. A-10436; again mixed with 1 mL of 50 mM NaHCC>3 solution). After the reaction mixture was covered with foil (to protect from light) and stirred with cooling (on ice) for a further 60 minutes, the resulting mixture was removed from the ice and allowed to warm to room temperature, before being stirred overnight at room temperature.

Dialysis of the resulting mixture was then performed (to remove EDC, NaHCC>3, and unreacted Phe). The membrane used should have a molecular weight cut-off designed to prevent the synthesised polymer from escaping (the molecular weight of the polymer being in the range of 200 to 500 kDa). The particular membrane selected (Spectra/Por 7 Dialysis Membrane, Spectrum Laboratories) had a molecular weight cut-off of 15 kDa. The procedure used for the dialysis was as follows. 1) Fill a large beaker with 4L dH ₂ 0, or MilliQ H ₂ 0.

2) Cut a piece of membrane off, 30-40 cm in length.

3) Clean the membrane (it has sodium azide on it to prevent growth of any bacteria).

Place in a beaker of H2O. Using a tap, wash the inside of the membrane with H2O. Shake up and down. Repeat 2-3 times.

4) Repeat the cleaning of the membrane with dhbO

5) While cleaning, check that there are no holes in the membrane

6) Place the membrane back in the beaker of H2O, fold over one end of the membrane (about 3 cm) and clamp it closed. Secure the clamp with a rubber band.

7) Use a stripette to add the reaction solution into the membrane. Avoid making

bubbles. Rinse the conical flask with a few mL H2O and add to the membrane.

Membrane should be -half full. If too full, it can fill with H2O due to osmotic pressure and burst.

8) Squeeze air out of membrane, fold and clamp the open end of the membrane. Secure the clamp with a rubber band.

9) Put the entire sealed membrane into the 4L flask of H2O. Add stirrer bar and place on magnetic stirrer. Cover with cling-film. Protect from light by covering.

10) Change the H2O to fresh -every hour.

1 1) Leave dialysis for 2-3 days.

The dialyzed solutions were freeze-dried. The freeze-dried material was then washed with pure ethanol according to the following procedure.

1) To a beaker containing the freeze-dried material, add 20 mL 100% Ethanol (do not pipette up and down as the polymer will stick to the pipette). 2) Using a metal spatula, mix the polymer into the ethanol, removing polymer stuck to the sides.

3) Pour the 20 ml_ into a pre-weighed 50 ml_ falcon tube.

4) Add another 20 ml_ 100% Ethanol to the beaker and again use the spatula to remove any polymer from the sides of the glass.

5) Pour the 20 ml_ into the falcon tube.

6) Use a vortex to fully disperse the polymer in the ethanol (Note: the polymer does not dissolve in Ethanol).

7) Centrifuge the falcon (ensure balanced by weighing) at 3500 rpm.

8) Pour off supernatant.

9) Repeat the ethanol washing until the colour of the supernatant has become clear (~3 times).

Excess solvent was removed from the washed material by drying under vacuum to produce particulate material.

Washed and dried γ-PGA-Phe particulate material prepared as described above was dissolved in DMSO to afford a 10 mg/mL solution. An equivalent volume of 0.15 M NaCI was added to that solution to produce a clear suspension. The solution obtained in this manner was dialyzed against distilled water to remove the DMSO. The resulting aqueous suspension of nanoparticles was then freeze-dried.

Suspensions of nanoparticles in phosphate-buffered saline (PBS) were prepared by adding at least 200 of PBS to the freeze-dried nanoparticles. After a few seconds of sonication of the mixture, a clear suspension is obtained.

A suspension of nanoparticles (10 mg/mL) in PBS, diluted 1 :1 in water, was analysed by dynamic light scattering (Zetasizer Nano ZS (Malvern Instruments), and the nanoparticles were determined to have a Z-average diameter of 180 nm and a polydispersity index of 0.12. Using the same sample and the same measurement device (Zetasizer Nano ZS (Malvern Instruments)), the nanoparticles were also determined (by laser Doppler microelectrophoresis) to have a Zeta potential of -20 mV.

The γ-PGA-Phe nanoparticles prepared as described above (10 mg/mL) were mixed with Gap 26 (2 mg/mL), MitoQ (2 mg/mL; see, for example, WO 2005/019232) or Compound 17 (0.5 mg/mL) at equivalent volume in 0.2 M NaCI, and incubated at 4°C for 12 h. The resulting, drug-loaded nanoparticles were isolated by centrifugation, washed with PBS, and resuspended in PBS (10 mg/mL). Both in vitro and in vivo experiments used nanoparticles that carried a final dose of 0.5μΜ MitoQ relative to their surrounding culture media or maternal blood. Nanoparticles with and without MitoQ were conjugated (by amide coupling, as described above) with Alexa Fluor 488 (Molecular Probes, Inc.) for localization studies. Nanoparticle measurements

The amount of MitoQ which was adsorbed to nanoparticles was evaluated by the Lowry method for Gap 26, and by UV absorption measurement for MitoQ (at 278 nm) and Compound 17 (at 331 nm). The adsorption efficiency (%) was calculated as (adsorbed drug weight /initial feeding weight) ^χ 100. The drug loading was calculated as loaded drug weight ^g)/nanoparticle weight (1 mg). The release of drug from nanoparticles in vitro was calculated by immersing them in PBS or culture medium with 10% foetal calf serum (FCS) and measuring drug (e.g. MitoQ) in the supernatant. Cell Cultures

Bilayered cell barriers were prepared using BeWo cells, a choriocarcinoma cell line (gift from Dr Margaret Saunders, University of Bristol, UK; see Bhabra G. et ai, Nat. Nanotechnol. 2009, 4, 876-83 and Sood et ai, Nat. Nanotechnol. 2011 , 6, 824-833), or primary human trophoblast cells, grown on Transwell 0.4 μΜ pore polyester inserts (Corning, UK). Villous cytotrophoblasts were collected from placentas delivered by elective caesarean section at term and prepared as described previously (Curtis D.J. et ai, Exp. Neurol. 2014, 261, 386-95). Culture media for BeWo barriers consisted of Dulbecco's modified eagle medium (DMEM), with F-12 (Sigma-Aldrich) containing 10% FBS (Thermo Fisher Scientific) and 2 mM L- glutamine (Sigma-Aldrich). Primary trophoblast cells were incubated in DMEM/F-12 with 10% FBS, 1.2 mM L-glutamine, 100 μg/mL gentamicin and 40 U/mL penicillin-streptomycin (Thermo Fisher Scientific). (Trophoblast culture media was made using 44% Dulbecco's modified eagle medium (DMEM), 44% F12, 1 % Gentamicin, 0.4% Pen/Strep, 0.6% L- Glutamine and 10% foetal bovine serum (FBS).) Cell lines were routinely tested for mycoplasma contamination. Integrity of cell barriers was tested by applying fluorescein isothiocyanate-conjugated BSA (FITC-BSA; Sigma-Aldrich) at a final concentration of 100 μg/mL to the top of cell barriers in Transwell inserts and an equal concentration of non- conjugated BSA to the medium below barriers. Fluorescence intensity was measured 1 h later in the media below the barriers and in the insert. Primary human BJ skin fibroblasts (ATCC CRL-2522) were grown in Minimal Essential Medium (Sigma-Aldrich), supplemented with 10% FBS (Invitrogen), 2% HEPES buffer (Sigma-Aldrich), 1 mM sodium pyruvate solution (Thermo Fisher Scientific), 100 U/mL penicillin-streptomycin (Sigma) and 2 mM L-glutamine (Sigma). (Fibroblast culture media was made using Minimal Essential Medium, supplemented with 10% FBS, 2% HEPES buffer, 1 % sodium pyruvate solution, 1 % Pen/Strep solution and 1 % L-glutamine.) Trypan Blue assay was used to assess cell viability, in three independent experiments.

Cortical cultures were prepared from dissociated rat E18 cortical tissue and grown on glass coverslips as described previously (Curtis D.J. et ai, Exp. Neurol. 2014, 261, 386-95), in Gibco Neurobasal media with 1x Gibco B-27 Supplement, 1x antibiotic-antimycotic (all Thermo Fisher Scientific) and 2 mM L-glutamine. (Neurobasal culture media was made using 96% Gibco Neurobasal media (stored at 4°C, protected from light), 2% B-27 Supplement (10 mL of 50X stock; Gibco), 1 % L-glutamine (250 μΜ) and 1 % Pen/Strep (50 Mg/mL).) To produce neuron-only cultures 40 μΜ 5-fluoro-2-deoxyuridine was added to each well at 5 days in vitro. Astrocyte-only cultures were prepared by growing cortical cultures in DM EM with 10% FBS, 2 mM L-glutamine and 1x antibiotic-antimycotic. Cultures were accepted as astrocyte-only if neuron number was zero in three representative coverslips. The effect of glutamate was investigated by adding MK-801 , a NMDA receptor antagonist, at a concentration of 10 μΜ to each well and incubating for 4 h before exposure of the cells to conditioned media. Exposures of cortical cultures were performed at a minimum in triplicate for condition medium from each of 3 placenta (one from each dam), creating a minimum of 9 data points for each condition. At least three different sets of cortical cultures were tested. Conditioned media

Conditioned media were prepared using bilayered barriers described above, using explants of first trimester human placenta were obtained with patients consent and ethical approval and prepared according to previous protocols (see Curtis D.J. et al., Exp. Neurol. 2014, 261, 386- 95). Barriers were grown and tissues were incubated at 21 %, 8% or 2% oxygen in a SCI-tive hypoxia chamber (Baker Ruskinn, USA). This was followed by a 24 h exposure period, at the start of which the media under the barriers or around the tissue was replaced with the appropriate culture media (e.g. neurobasal or fibroblast media, prepared as described above), depending on the cell type to be subsequently exposed. Some wells were additionally exposed to NPs above the barrier or around the tissue (approximately 1.1 μg/mL combined nanoparticle). The media was then conditioned by the explant or barrier for 24 h at the same or altered oxygenation (e.g. the neurobasal or fibroblast media were conditioned by the explant or barrier for 24 hours at altered or similar oxygenation at 2%, 8%, 12% or 21 % oxygen in a Ruskin 5ive oxygen chamber). Additional neurobasal media were conditioned by rat placenta (see below) at 21 % or 2% oxygen for 24 hours or below and beneath BeWo barriers, exposed above the barrier to 30 μΜ hydroquinone and 30 μΜ benzoquinone, with and without nanoparticles for 24 hours.

Exposure of cells to conditioned media in vitro

Cells were dissociated from embryonic (E18) rat cortex and grown on glass coverslips for 12 days. The conditioned media was applied to them for 6 days or to fibroblasts for 1 day at 21 % oxygen in vitro, to test their effects on dendrite length, astrocyte process length, cell number, tyrosine hydroxylase positive cell processes, glutamate and GABA receptors, and DNA damage as previously described (see Sood et al., Nat. Nanotechnol. 2011 , 6, 824-833 and Curtis D.J. et al., Exp. Neurol. 2014, 261, 386-95). Cortical cultures were also exposed to conditioned medium collected from astrocyte-only cultures that had been exposed for 24 h to rat placenta conditioned medium. Potential leakage of MitoQ from barriers or placenta was measured in the conditioned medium using liquid chromatography-tandem mass spectrometry (LC MS/MS; see Rodriguez-Cuenca, S. et al. Free Radical Biology & Medicine 2010, 48, 161- 172).

Immunocytochemistry

Cortical cultures were fixed in supercold methanol (-20°C), washed with PBS and blocked with 5% BSA, 5% NGS in PBS for 30 min. They were incubated with primary antibodies against MAP2 (1 :2000, #188004; Synaptic Systems, Germany), GFAP (1 : 1000, #3670; Cell Signalling Technology), tyrosine hydroxylase (1 :500, ab112; abeam) and GluNI (1 :500, ab9864; Merck Millipore), overnight at 4°C. Sections were stained with secondary antibodies Alexa Fluor 488 anti-rabbit IgG, Alexa Fluor 488 anti-mouse IgG or Alexa Fluor 488 anti guinea pig IgG (all Thermo Fisher Scientific, diluted 1 :500) for 2 h at room temperature under minimal light conditions, washed with PBS and mounted in DAPI mounting media. Five images per coverslip were taken on a confocal microscope (SP2-AOBS, Leica). Analysis of the slides was performed with the experimenter blind to the experimental group. Dendrite lengths were measured using ImageJ. For receptors images were taken at x64 (with oil) on a fluorescence microscope (Leica SP5II) after excitation at 488 nm. Using ImageJ, images were converted to RGB files then measurements were taken of the mean grey value of each image providing an average of the relative intensity of the staining. Measurements were verified by cell counts based on DAPI staining. Background levels of fluorescence were ascertained in the absence of cells, primary antibody and/or secondary antibody.

Analysis of conditioned media and plasma

Bone morphogenic proteins (BMPs) 4, 6 and 9 were measured in conditioned media using a cell based assay (Herrera B. and Inman G.J. BMC Cell Biol. 2009 10, 20). BMP2 was estimated with Elisa (Quantikine R&D). Amino acids were measured as previously described (see Curtis D.J. et al., Exp. Neurol. 2014, 261, 386-95).

Small RNAs and MicroRNAs were extracted from conditioned media or foetal plasma (total RNA was extracted from 200 μΙ_ conditioned media using the miRNeasy Mini Kit (Qiagen, Germany) or from 100 μΙ_ foetal blood using the miRNeasy Serum/Plasma Kit (Qiagen) and their levels were measured using the Small RNA Kit on the 2100 Bioanalyzer (Agilent Technologies) at the University of Bristol Genomics Facility. Levels of individual miRNAs were analysed using the nCounter Rat v1 miRNA Expression Assay or the nCounter Human v2 miRNA Expression Assay (NanoString Technologies, USA), which detects 423 or 800 different species-specific miRNA, respectively. Briefly, 3 of each undiluted sample were hybridised with barcoded probes and immobilised on an nCounter Cartridge. Barcode signals were counted using the nCounter Digital Analyzer..

Proteomic analysis was performed by the University of Bristol Proteomics Facility. Briefly, samples were depleted of rat albumin, digested with trypsin and labelled with Tandem Mass Tag (TMT) 10Plex reagents (Thermo Fisher Scientific). The labelled samples were fractionated by high pH reversed-phase chromatography followed by nano LC-MS/MS. The raw data files were processed and quantified using Proteome Discoverer software v1.4 (Thermo Fisher Scientific) and searched against the UniProt Rat database using the SEQUEST algorithm. All peptide data was filtered to satisfy a false discovery rate of 5%. RNA sequencing

RNA was extracted from 50 mg of foetal frontal cortex tissue at GD20 using the RNeasy Mini kit (Qiagen). RNA quality and integrity was measured on the 2100 Bioanalyzer. mRNA sequencing was performed by Edinburgh Genomics. Libraries were prepared from total RNA samples using the lllumina TruSeq stranded mRNA Sample Preparation Kit. Briefly the polyA RNA from 1 ug of total RNA was captured onto Oligo d(T) beads, before fragmentation and elution of the polyA RNA. The RNA was reverse transcribed using random primers and the resulting cDNA was double stranded. The cDNAs were ligated with adapters containing unique barcodes for each sample. Libraries were then assessed for size by electrophoresis and quantified by qPCR. The libraries were sequenced by75 bases paired-end sequencing across two lanes of an lllumina HiSeq 4000. This level of sequencing produced greater than 34 million paired reads per sample. The FASTQ files were generated using the standard lllumina pipeline for bcl2fastq. Bioinformatic analyses

Differential expression analysis of NanoString miRNA data is detailed below. TargetScanHuman v7.052 was used to create a list of potential (human-equivalent) targets of the significant miRNAs (Total Context Score < -0.2). Predicted target genes were analysed for pathway enrichment of CNVs associated with schizophrenia, as described by Pocklington et al. (Neuron 2015, 86, 1203-1214). The analysis was applied to the combined ISC+MGS+CLOZUK dataset, comprising a total of 1 1 ,355 cases and 16,416 controls. The analyses are based on large, rare CNVs (>100kb, frequency < 1 %), as these are both the most robustly called and most enriched in people with schizophrenia. The primary analysis was performed on all CNVs, with secondary analyses performed for deletions and duplications separately. A further analysis was performed on the 'minimal' gene sets previously shown to capture CNS-related gene set enrichment in this CNV dataset (see Pocklington et al., supra).

Peptide spectra from the proteomics data set were analysed using moderated t-statistics from the empirical Bayes method to calculate differential protein abundance. Modified scripts, published by Kammers et al. (EuPA Open Proteom. 2015, 7, 1 1-19) and based on the limma package (Nucleic Acids Res. 2015, 43, e47), were used.

Tophat (Bioinformatics 2009, 25, 1105-1 11 1) was used to align RNA sequencing reads to the rat reference genome Rnor_6.0 (GenBank Assembly ID GCA_000001895.4), and HTSeq (Bioinformatics 2015, 31, 166-169) to generate read counts from the resulting BAM files. Out of 672 million paired-end reads, nearly 557 million (82.8%) concordant pairs mapped to unique locations in the Rn6 version of the genome. Genes were analysed for differential expression and differentially expressed genes were further analysed for interactions with the miRNA data sets (see below).

Differentially abundant proteins, differentially expressed genes and predicted targets of differentially secreted miRNAs were analysed for enrichment of biological processes, cellular compartments and tissues using the GO-slim feature in PANTHER 11.0 (see Nat. Protoc. 2013, 8, 1551-1566 and Nucleic Acids Res. 2016, 44, D336-342) and the GO Direct and UP_TISSUE features in DAVID 6.8 (see Nat. Protoc. 2009, 4, 44-57 and Nucleic Acids Res. 2009, 37, 1-13). Analysis of NanoString data

NanoString nCounter data consist of discrete sequence counts as a measure of miRNA expression within each sample. These counts are similar to counts from other high-throughput methods such as RNA sequencing, where discrete statistical models such as the Poisson or negative binomial distributions may be used to estimate differences between samples (see Genome Biology 2010, 11, R106 and Bioinformatics 2010, 26, 139-140). Consequently, differential expression prediction methods originally designed for RNA sequencing data may also be used with NanoString data (see Brumbaugh, C. D. et al., BMC Bioinformatics 2010, 12, 479). Following the procedures outlined in Brumbaugh et al., the following pipeline was developed to assess DE-miRNA between samples: First, each raw NanoString output file was converted into a list of counts. Next, the counts for the replicates in each sample were merged into a table with one column per replicate. To compare two samples, the corresponding tables were merged into a single table that was then passed into a differential expression analysis pipeline (below).

Differential expression analysis of NanoString and RNA sequencing data

RUVSeq68 was used to remove unwanted variation and then edgeR66 to predict differentially- expressed RNAs (miRNAs in the case of NanoString data, mRNAs in the case of RNA sequencing data). EdgeR was selected for the pipeline, as it may be more sensitive than DESeq and it facilitates the use of RUVSeq (which was used RUVSeq to eliminate, as far as possible, variation from sources unrelated to the treatment groups, such as differences in blood plasma collection, centrifugation, enrichment, and RNA purification). As no differential expression was expected between replicates within a treatment group and few differentially- expressed RNAs between treatment groups, the relative log-expression should be consistent across all samples. Further, the largest component of variation in the data should reflect RNAs that are differentially expressed between treatment groups. For each comparison, RUVSeq was first used to adjust the counts to account for unwanted variation. The adjusted counts were then used with EdgeR's generalized linear model to yield the final predictions. To mitigate possible false-positives, miRNAs were classed as significant differentially secreted miRNAs if p < 0.05, if count≥ 10 for at least one of the compared conditions (except for foetal plasma miRNAs, where counts were generally very low) and if there was an up or down regulation of at least 25%. Analysis of miRNA-mRNA correlation

Enrichment analysis of the RNA sequencing data for predicted targets (derived from TargetScanHuman; see Free Radic. Biol. Med. 2015, 88, 212-20) of significant miRNAs was performed in R/Bioconductor using the Fisher's exact test and investigating enrichment only. Correlation of abundance changes of significant miRNAs with abundance changes of significant mRNAs was analysed with the miRComb (PloS One 2016, 11, e0151 127 package for R/Bioconductor, using Spearman correlation and Benjamini-Hochberg adjustment for multiple comparisons. In vivo experiments

Three-month-old female Sprague-Dawley rats (Charles River, Wilmington, MA) were maintained on ad libitum standard rat chow and tap water in a 12: 12-h light-dark cycle and acclimatized before breeding. Day 0 of pregnancy was determined by sperm in a vaginal smear.

Rats at day 15 of pregnancy had an average weight of 300 g, with an estimated blood volume of 25 mL (50 mL/kg). At gestational day (GD) 15 of pregnancy the rats were injected intravenously via the tail vein with saline (vehicle control) or with 100 μΙ_ of 125 μΜ MitoQ nanoparticles prepared as described above (to create a blood concentration of nanoparticles of 0.5 μΜ) and exposed for the next 6 days to 21 % or 12% oxygen in an A-Chamber, BioSpherix (see Morton J.S. et al., J. Applied Physiol. 2011 , 110, 1073-1082).

Some rats were sacrificed at GD 20. EDTA plasma collected from foetuses by decapitation was pooled per litter and flash frozen; placenta, maternal and foetal tissues and maternal blood (serum) were collected fresh, flash frozen or fixed in formal saline. Additional placenta were incubated fresh ex vivo in tissue culture medium for 24 hours at 21 % or 2% oxygen to examine their secretions. Other rats were allowed to give birth in normal oxygen conditions. Their birth weights were measured and the brains of the offspring were examined at postnatal day (P) 30.

Statistical tests were not used to predetermine sample size. Experimental groups were allocated according to a preassigned schedule, depending on the order in which successful pregnancy was established. From each litter two males and two females were randomly selected for examination at P30. For subsequent immunohistochemistry analysis of rat tissue, the experimenter was blinded to group allocation.

Localisation of NPs

Pregnant rats were injected with 125 μΜ NPs conjugated with Alexa Fluor 488 and culled at GD16 to collect flash-frozen foetal, maternal and placental tissue. Sagittal sections of placental labyrinth and junctional zone, sections of foetal and maternal cortex and liver were produced using a cryostat and viewed with confocal microscopy for the localisation of fluorescent NPs.

Analysis of placental, foetal and maternal tissues

Placenta were fixed in 10% formal saline and processed for paraffin wax histology. Sections were cut at 3 μηι and stained with Haematoloxylin and Eosin and at 2 μηι for avidin-biotin immunocytochemistry for CD 34 (ab81289; Abeam, UK) using a full automated Bond3 immunostaining machine with bond polymer refined detection (Leica) to ensure the same immunostaining method was applied to all sections. . Images were captured on a fluorescent microscope (Leica SP5II) and analysed with ImageJ software. For measurements pertaining to placenta vasculature, values for each biological replicate are averages of technical replicates taken from multiple fields of view. Levels of reactive oxygen species (ROS) were measured in foetal, maternal and placental tissues using the 2',7'-dichlorofluorescein diacetate (DCFDA) assay. Sagittal sections (10 μηι) were cut on the cryostat and exposed to 20 μΜ DCFDA solution in HBSS at 37°C in a humidifying chamber. After counter-staining with DAPI, slides were immediately imaged using a confocal microscope (excitation and emission wavelengths of 495 nm and 529 nm, respectively). Using Image-Pro Premier 9.2 (Media Cybernetics, USA), fluorescence levels of DCF, the product of DCFDA deacetylation by cellular esterases and oxidation by ROS, were quantified in labyrinth and junctional zone of the placenta, maternal and foetal liver, cortex and cerebellum.

Fresh placenta were homogenised (manually in 1.5 ml_ RIPA buffer) and protein carbonyls were measured with OxyBlot assay protein oxidation detection kit (Merck Millipore, Germany) according to manufacturer's protocol. Briefly, 20 μg tissue lysate in 50 mM dithiothreitol was subjected to a derivatisation reaction with 2,4-dinitrophenylhydrazine. The resulting 2,4- dinitrophenyl moieties were detected using Western blotting. Six bands were selected for densitometry analysis using ImageJ software. Density was normalised to total protein loading as measured by Ponceau S staining.

Offspring Brains

At postnatal day 30, rats were anaesthetised with 4% isoflurane and perfused with 4% paraformaldehyde (PFA). The brains were post fixed in 4% PFA, placed in 30% sucrose until sunk. Three brains per condition were randomly selected, one from each litter. The brains were then placed in an aluminium foil cup and O.C.T. solution (Tissue-Tek®) was added. The aluminium foil cup was then frozen using dry ice and placed in a -80 degrees centigrade freezer.

Immunohistochemistry of sections of brain

12 μηι cryostat sections, mounted as contiguous triplicates, stained with antibodies for neuronal and astrocyte markers (as described below). Images were captured using confocal or conventional fluorescent microscopy and analysed with image analysis.

12 μηι cryostat sections of offspring brain prepared as described above were exposed to super cold methanol (2% PFA for MAP2 staining) at -20°C for 10 min for fixation. For GluN1 staining, sections were fixed in 2% PFA, followed by permeabilisation in 0.3% Triton X-100 in PBS for 15 min. All slides were washed in PBS after every incubation step (once with PBS and once with PBS+tween-20 for 5 minutes (each wash) at 4°C in a humidifying chamber). Each slide was then exposed inside the pap-pen lines to 0.5 ml_ of Quench solution (26.74 mg Ammonium chloride in 10 mL PBS) for 5 minutes. A blocking solution was then prepared as 10% BSA in PBS with 5 μΙ_ triton and 25 μΙ_ tween- 20. 1 mL of blocking solution was added and slides were left for 4 hours at 4°C. The slides were then washed twice in PBS and once with PBS+tween-20 for 5 minutes (each wash) at 4°C. Slides were blocked with 5% goat serum (Sigma-Aldrich), 0.3% Triton X-100 in PBS for 2 h at 4°C, followed by an overnight incubation (or 48 h incubation for GluN1 staining) at 4°C in primary antibody in PBS with 1 % BSA, 0.3% Triton X-100. A solution of primary antibody was then made in blocking solution. The antibodies used were raised against MAP2 (1 :500, Abeam ab32454), NeuN (1 :1000, Abeam ab177487), neurofilament (1 :500, ab24575; Abeam), GFAP (Cell Signalling Technology, Ab #3670 1 :500), Parvalbumin (Abeam ab1 1427 1 :500), Tyrosine hydroxylase (Abeam ab112 1 :500) and GluN1 (1 :200, MAB1570, Merck Millipore). The sections were exposed overnight at 4°C. The next day the slides were washed twice in PBS and once with PBS+tween-20 for 5 minutes (each wash) on a moving shelf. A solution of secondary antibody (Alex Fluor 555) was then made in the dark in blocking solution. Slides were incubated with secondary antibody Alexa Fluor 555 anti-rabbit IgG, Alexa Fluor 488 anti-mouse IgG or Alexa Fluor 568 anti-mouse IgG (Thermo Fisher Scientific) at 1 :500 for 2 h at 4°C. In the dark slides were then washed twice in PBS and once with PBS+tween-20 for 5 minutes (each wash) at 4°C. Slides were then mounted (as contiguous triplicates) using Vectashield Mounting Medium with DAPI (Vector Laboratories, USA).

Coronal sections were chosen to show thalamic reticular nucleus, primary somatosensory, primary auditory and retrosplenial granular cortex (see Paxinos G., Watson C. (2007) The Rat Brain in Stereotaxic Coordinates, 6th Ed San Diego, Elsevier Academic Press). Damaged sections, due to problems with freezing or storage of the samples, were excluded from analysis after visual confirmation. Analysis of brain sections was done with the experimenter blind to the in vivo exposure. For each site, in each brain, 5 fields of view were examined for each of 3 sections in both hemispheres using a LASX (Leica) widefield microscope or a SP5II (Leica) confocal microscope at x40 magnification using oil. This resulted in a minimum of 30 fields of view being analysed for each site. Images were captured using Zeiss AXIO Imager.AI with Q-capture software and analysed using Image-Pro Premier and Image-J software. The analysis of the intensity of the IHC staining of GluN1 was determined using a similar approach to that used previously (see Hughes, E. G. et al., J. Neurosci. 2010, 30, 5866-5875). Briefly, nine sections per condition were imaged and 10-12 images were collected for every brain region per case and using the same microscopy settings of intensity and magnification to allow for a meaningful comparison. Photos were sorted and converted to greyscale then the total pixel number was determined using a corrected macro for ImageJ. The total pixel number was subsequently subtracted from the background.

Immunocytochemistry of cortical cultures

Cortical cultures that had been exposed to conditioned media were fixed in supercold methanol (-20°C). They were washed with PBS twice and blocked with 5% BSA, 5% NGS in PBS for 30 min. They were incubated with primary antibodies against MAP2 (1 :2000, #188 004; Synaptic Systems, Germany), GFAP (1 :1000, #3670; Cell Signalling Technology), tyrosine hydroxylase (1 :500, ab1 12; Abeam) and GluN1 (1 :500, ab9864; Merck Millipore), overnight at 4°C. The sections were stained with secondary antibodies Alexa Fluor 488 anti-rabbit IgG, Alexa Fluor 488 anti-mouse IgG or Alexa Fluor 488 anti-guinea pig IgG (all Thermo Fisher Scientific, diluted 1 :500) for 2 h at room temperature under minimal light conditions, washed with PBS and mounted in DAPI mounting media. Five images per cover slip were taken on a confocal microscope (SP2-AOBS, Leica) at x40 magnification. Dendrite lengths were measured using ImageJ. For receptors images were taken at x64 (with oil) on a fluorescence microscope (Leica SP5II) after excitation at 488 nm. Using ImageJ, images were converted to RGB files then measurements were taken of the mean grey value of each image providing an average of the relative intensity of the staining. Measurements were verified by cell counts based on DAP1 staining. Background levels of fluorescence were ascertained in the absence of cells, primary antibody and/or secondary antibody.

Immunohistochemistry of P30 brain sections

Frozen P30 brain sections were exposed to cold methanol at -20°C for 10 min for fixation. For NR1 staining sections were fixed in 2% PFA, followed by permeabilisation in 0.3% Triton X- 100 in PBS for 15 min. All slides were then washed in PBS and after every incubation step. The slides were blocked with 5% goat serum (Sigma-Aldrich), 0.3% Triton X-100 in PBS for 2 h at 4°C, followed by an overnight incubation (or 48 h incubation for NR1 staining) at 4°C in primary antibody in PBS with 1 % BSA, 0.3% Triton X-100. Antibodies used were raised against MAP2 (1 :500, ab32454; abeam), NeuN (1 : 1000, ab177487; abeam), neurofilament (1 :500, ab24575; abeam), GFAP (1 :500, #3670; Cell Signaling Technology), parvalbumin (1 :500, ab1 1427; abeam), tyrosine hydroxylase (1 :500, ab1 12; abeam) and GluN1 (1 :200, MAB1570, Merck Millipore). Slides were incubated with secondary antibody Alexa Fluor 555 anti-rabbit IgG, Alexa Fluor 488 anti-mouse IgG or Alexa Fluor 568 anti-mouse IgG (Thermo Fisher Scientific) at 1 :500 for 2 h at 4°C. Vectashield Mounting Medium with DAPI (Vector Laboratories, USA) was used to mount coverslips. Coronal sections were chosen to show thalamic reticular nucleus, primary somatosensory, primary auditory and retrosplenial granular cortex. For each site, in each brain, 5 fields of view were examined for each of 3 sections in both hemispheres using a LASX (Leica) widefield microscope or a SP5II (Leica) confocal microscope at x40 magnification using oil. This resulted in a total of 30 fields of view being analysed for each site.

Statistics

Data are presented as means ± s.e.m. For all statistical comparisons, variances were similar in magnitude between the compared groups. One-way or two-way ANOVA were performed in Prism 6.0 (GraphPad, USA) or SPSS 21.0 (IBM Corp., USA) with post-hoc analysis using Bonferroni correction or Tukey's test for multiple comparisons. Two-way ANOVA was used to test for main effects of drug and of oxygen conditions and for interaction effects. For GluN1 staining intensity analysis, normality of data distribution was tested using the D'Agostino- Omnibus normality test in Prism 6.0 (p > 0.05) and parametric testing was applied to test for differences of means between groups using multiple unpaired, two-sided Student's test post- hoc testing adjusted for multiple comparisons using Bonferroni correction.

For receptors: data distribution was assumed to be normal and tested using the D'Agostino- Omnibus normality test on Graphpad prism version 6 (p>0.05) and parametric testing was applied to test for differences of means between groups using multiple unpaired student-t-test post-hoc testing.

Results

Physical properties of drug-loaded nanoparticles

Gap 26, MitoQ and Compound 17 were adsorbed by hydrophobic and electrostatic interaction with nanoparticles which were made of an amphiphilic copolymer of poly(Y-glutamic acid) and L-phenylalanine ethylester (Phe) (γ-PGA-Phe NPs). There was a sustained release of the drug from these nanoparticles into both PBS and tissue culture medium containing serum. The adsorption efficiency of the drug was between 30 to 40% and the amount adsorbed depended on its initial concentration. The sizes of the Gap 26 (50 μg/NP 1 mg), MitoQ (45 μg/NP 1 mg) or Compound 17 (20 μg/NP 1 mg)-adsorbed NPs were 260 nm (PDI = 0.14), 180 nm (PDI = 0.09) and 220 nm (PDI = 0.08), respectively. The drug-adsorbed NPs showed a negative zeta potential in PBS (-20 mV), possibly due to the ionized carboxyl groups of the γ-PGA on the NP surfaces. The NPs were stable under physiological conditions, exhibiting no aggregation, precipitation, or dissociation for at least 6 weeks.

When the NPs were applied to bilayered BeWo barriers or primary human trophoblast cells ('trophoblast barriers') for up to 24 hours, they were, predominantly located in the top layer of the barrier. There was no evidence of a passage of NPs, or release of MitoQ, across the barriers into the media below except at very high NP dose (2 mg/mL, 2580-fold higher than in vivo or in vitro doses), applied for 24 hours. Signalling through barriers in vitro

Bilayered BeWo barriers, bilayered primary human cytotrophoblast barriers and explants of first trimester human placenta were exposed them to altered oxygenation, to imitate a stimulus during pregnancy that might be linked with adverse effects on the foetus and its future development. Conditioned tissue culture media collected from beneath the barriers or around the explants in vitro were investigated to determine whether they had secreted factors that would cause DNA damage in fibroblasts and dendrite shortening in dissociated neurones from rat embryonic cerebral cortex (which forms of damage have relevance to psychological disorders). Whilst the conditioned medium caused DNA damage and dendrite shortening, this was almost completely prevented by applying the nanoparticles bound to MitoQ (MQ-NPs) to the bilayered BeWo or primary human trophoblast barriers or human placenta at the start of the hypoxia reoxygenation. Nanoparticles bound to Gap26 (Gap-NPs) were effective when applied to BeWo barriers but not to explants of human placenta. Dendrite shortening was also prevented by MQ-NPs and Gap-NPs in experiments where BeWo barriers were exposed to chemical toxins (benzoquinone and hydroquinone). Nanoparticles bound to the ATP receptor antagonist PPADS (PPADS-NPs) were less effective. MitoQ or Gap26 bound to nanoparticles were found to be as effective as the soluble drug without nanoparticle attachment. In all cases, unloaded NPs were ineffective. In order to explore the potential safety of such NPs, should they pass through a placental trophoblast barrier in vivo with or without a toxin, fibroblasts and cerebral cortical cells were exposed directly to NPs. Neither MQ-NPs and/or Gap-NPs caused DNA damage or dendrite shortening on their own. MQ-NPs did not alter the amount of DNA damage in fibroblasts which had been caused by direct exposure to benzoquinone and hydroquinone, whereas Gap-NPs slightly reduced that damage.

MQ-NPs, unloaded NPs or MitoQ alone had no direct effect on cell viability of fibroblast or cortical cultures. In particular, total cell number, neuron count and astrocyte count of rat cortical cultures were all unaffected by direct application of unbound MQ-NPs, unloaded NPs or MitoQ alone.

Effects in vivo Drug delivery nanoparticles were investigated for their ability to prevent responses to hypoxia in vivo, using an animal model been developed by Davidge and co-workers (Morton, J.S. et al. J. Appl. Physiol. 2011 , 110, 1073-1082). Here pregnant rats are exposed to 1 1 % oxygen (equivalent to high altitude) for the last 6 days of pregnancy, which reduces foetal birth weight and causes abnormal foetal cardiovascular programming.

The investigations with nanoparticles examined whether there were changes in what was released from the placenta ex vivo and whether there were alterations in birth weight and in the brains of the offspring in later life. The NPs were also investigated for their ability to prevent any changes that were seen, if injected intravenously at the start of the hypoxia.

MQ-NPs (a single dose) showed the best response in experiments with model barriers. Effects on rat placenta and foetus

Birth-weights were decreased following maternal hypoxia. MQ-NP injection rescued over 60% of this deficit. Hypoxia and MitoQ-NPs had no effect on body weight at P30, or placenta or brain weight.

The rat placenta differs from the other model systems used, in having three layers. Two layers are syncytiotrophoblast and these form the barrier. In addition, the syncytiotrophoblast layer in the rat placenta faces the foetus, unlike in the human placenta, and it is the cytotrophoblast that faces the maternal circulation.

NPs were detected within the placenta, most prevalently in the labyrinth but also in the junctional zone. The NPs were particularly found in cytotrophoblasts, which face the maternal circulation, and less commonly seen in syncytiotrophoblast cells, which face the foetus. NPs were not found in the foetal brain or in thoracic or abdominal tissues including liver. NPs were detected in the maternal liver, particularly in Kupffer cells and hepatocytes. Areas of NP localisation were detected in the maternal brain, and fluorescence was mainly detected within neuronal processes, but these areas were very sparse throughout the brain. Neither maternal hypoxia nor nanoparticle injection caused a gross change in placental fine structure in sections stained with haematoxylin and eosin or in the width of labyrinth or decidua. Nor was the relative area, number, perimeter or diameter of the foetal capillaries altered in the labyrinth of the placenta as shown by CD 34 immunostaining. Levels of oxidative stress (as determined by measurements of levels of fluorescent dichlorofluorescein) were significantly increased in the labyrinth and junctional zones of the placenta following gestational hypoxia. Treatment with MQ-NPs significantly reduced levels of oxidative stress. Analysis of maternal brain, liver and foetal liver showed elevated levels of oxidative stress. After MitoQ treatment, increased oxidative stress was still observed in foetal liver but not in maternal liver, whilst oxidative stress in maternal brain was rescued following MQ-NP administration. In contrast, no changes in oxidative stress were detected in the foetal brain after maternal hypoxia or NP treatment.

Both 6 days of maternal hypoxia in vivo and 1 day of placental hypoxia ex vivo caused a higher level (albeit not significant) of protein carbonyls, a marker of hypoxia, in the rat placenta. The level with hypoxia was lower after MQ-NP treatment. No change was seen in p53.

These observations suggest that the MQ-NPs injected into the maternal circulation reduce oxidative stress in maternal tissues and placenta but not in foetal tissues following gestational hypoxia. Moreover, they rescue foetal birth weight although they do not reach the foetus.

Effects on BeWo barrier and placental secretions

Levels of molecules were measured in media conditioned by bilayered BeWo barriers or rat placenta.

BeWo barriers had been exposed to varying oxygen conditions with or without application of MQ-NPs. Rat placenta and foetal plasma was collected from Dams breathing 21 % and 12% oxygen after maternal administration of saline or MQ-NPs. The placenta was placed in tissue culture medium to condition it. This was done for 24 hours at 21 % as well as at 2% oxygen, in order to imitate a further stress of extreme (e.g. perinatal) hypoxia.

Three sets of identified molecules in the conditioned medium were measured that might have relevance to foetal development, the aetiology of psychiatric diseases and hypoxia of the placenta respectively, namely miRNAs, amino acids and bone morphogenetic proteins (BMPs). The biological effects of unidentified factors in the conditioned medium upon dissociated cells from embryonic rat cerebral cortex in tissue culture was also investigated.

Changes to oxygen levels (hypoxia or hypoxia-reoxygenation vs. normoxia) altered microRNAs and small RNAs, and potentially small extracellular vesicles (such as exosomes), released from BeWo barriers into the medium. There was no change in total microRNA or small RNA in medium conditioned by rat placenta after gestational hypoxia. However, in media conditioned by barriers or rat placenta and in foetal plasma there was a complex pattern of increased and decreased levels of individual microRNAs. The application of MQ-NPs partially reversed these changes for the majority of differentially secreted microRNAs.

The pattern of secretion of BMPs from BeWo barriers was also altered following exposure to reduced oxygen. This was prevented by the MQ-NPs. Individual protein levels in foetal plasma were changed in response to gestational hypoxia. These were significantly enriched for exosomes and lipoproteins, the latter being of interest in view of the susceptibility of offspring to cardiovascular disease after gestational hypoxia (see Giussani & Davidge, J. Devel. Orig. Health Dis. 2013, 4, 328-337 and Hutter, D. et ai, Int. J. Pediatrics 2010, 401323). No change was seen after NP treatment. In contrast, in media conditioned by placenta, there was little change after gestational hypoxia but significant change after NP treatment. The rat placenta released more microRNAs and small RNAs into the medium when the mother had been hypoxic for 6 days compared to when the mother breathed atmospheric oxygen, especially when the placenta was placed in 21 % oxygen compared to 2% oxygen. If the mother had been injected with MitoQ nanoparticles at the start of the 6 day hypoxia, the levels were similar to control. Quantitative analysis of all individual microRNAs was performed by nanostring analysis. There was a similar pattern in the media conditioned by rat placenta with maternal hypoxia in vivo, with or without additional 2% hypoxia in vitro, as in media conditioned by BeWo barriers with hypoxia in vitro and significantly in samples of the rat foetal blood at GD 20 in which there was maternal hypoxia with or without MitoQ NP injection. In all experiments some microRNAs were increased by the hypoxia whilst others were reduced compared to maternal normoxia. However in all, including in the foetal blood, those microRNAs that were increased were reduced towards or to normal levels by MitoQ NP injection/application whilst those that were reduced were now increased. This pattern did not depend on the method of analysis used. A bioinformatics analysis of the targets of these differentially expressed microRNAs in the conditioned media indicated that they are enriched in brain compared to other organs (p<0.001).

The levels of amino acids in the conditioned media were altered in the opposite direction. Overall levels tended to be lower in samples after maternal hypoxia (with or without additional hypoxia in vitro) but were significantly increased by the in vivo injection of MitoQ nanoparticles (p<0.05), including aliphatic and aromatic amino acids.

Biological effects of conditioned media in vitro and effects on the offspring in vivo

The placenta, as the interface between mother and foetus, is important in foetal programming and in the aetiology of neurodevelopmental disorders. The biological effects of the tissue culture media, as conditioned by the placenta ex vivo, were therefore compared with any effects that might be seen in the foetus or the offspring in later life. microRNAs regulate mRNA abundance (Nat. Rev. Genet. 2011 , 72, 99-110) and have been shown to be involved in neurodevelopment and disease (Experimental Neurology 2015, 268, 46-53). Moreover exosomes, extracellular vesicles known to carry microRNAs, can cross the blood-brain barrier and reach the brain (Nature Biotech. 2011 , 29, 341-345).

Bioinformatic analyses of predicted targets of the differentially expressed microRNAs identified in rat placenta conditioned media and foetal plasma indicated enrichment in the brain compared to other organs (p<0.001) as well as an enrichment in biological processes linked to development. Furthermore, predicted microRNA targets were enriched within copy number variants (CNVs) found in schizophrenia cases compared to CNVs found in matched controls. An additional analysis conditioned on a minimal set of CNS-linked gene-sets previously shown to be enriched for schizophrenia CNVs in the same case-control data (Neuron 2015, 86, 1203- 1214); this indicated that while some of the enrichment seen in the above-mentioned, predicted targets was shared with these known associated gene sets, a significant proportion appeared to be independent.

Expression (mRNA) of 510 genes was significantly altered in the foetal brain after maternal hypoxia. MQ-NPs reduced these effects. Genes down-regulated following hypoxia were enriched for developmental processes. Differentially expressed genes in the foetal brain were significantly enriched for predicted targets of those microRNAs that had been found to be significantly altered in foetal plasma and conditioned medium, suggesting a potential link between secreted microRNAs and cortical transcriptome changes following a hypoxic insult. Negative correlation with significant mRNA changes in the brain was detected for a subset of significant microRNAs, including miR-17-5p, which has been associated with psychiatric disease phenotype (PloS One 2014, 9, e86469).

Levels of amino acids that are regulated by transporters, which are impaired in foetal growth restriction or hypoxia, were less in the media conditioned by the placenta where there was maternal hypoxia. The levels were significantly increased with MitoQ-NP injection. Birth weight was less after maternal hypoxia, but this was rescued by MitoQ-NP injection. The brain weights at birth and the brain at body weights at PD 30 were unchanged.

In very early development, the brain's supply of transmitter serotonin (Nature 2011 , 472, 347- 350), and possibly others (Nature 2011 , 472, 298-299) is derived from the placenta. Obstetric challenge, maternal inflammation, alters this secretion from placenta to brain and disrupts foetal 5HT axonal outgrowth within the forebrain (J. Neurosci. 2016, 36, 6041-6049). The potential link between placental secretions and changes in the offspring brain was therefore explored, by investigating whether foetal plasma and media conditioned by model barriers and rat placenta had similar biological effects on embryonic cortical neurones in culture and whether these effects might also be noted within the offspring brains at a later stage of development. Parameters chosen were those having relevance to psychological disorders (see, for example: Schizophrenia Bull. 2012, 38, 920-926; Arch. Gen. Psychiatry 2012, 69, 776-786; and Am. J. Psychiatry 2003, 160, 13-23). Like the media that was conditioned by human placenta under in vitro hypoxia, the media conditioned by rat placenta after in vivo maternal hypoxia released factors that were investigated for their effects on brain cells (including effects on dendrite length of dissociated cells from embryonic rat cerebral cortex). Exposure to conditioned media and plasma collected following a hypoxic episode caused a shortening of dendrites, a reduction of tyrosine hydroxylase (TH)-positive process lengths, a loss of GluN1 receptor subunit staining and an increase in astrocyte-to-neuron ratio in cortical cultures. In general, such changes were also seen in the brain. The overall length of dendrites measured in each field of view was less in reticular nucleus after maternal hypoxia but not in somatosensory auditory and retrosplenal cortex combined. The decrease in dendrite length was thus restricted to the thalamic reticular nucleus (TRN) although thick, abnormally branched dendrites were noted in the somatosensory cortex (SSC). These changes were not noted after maternal MitoQ-NP injection.

The numerical density of parvalbumin-containing cells was slightly reduced in reticular nucleus and cortex combined after maternal hypoxia and this was less significant or not seen after MitoQ-NP injection. Process lengths (overall length per field of view) of tyrosine hydroxylase positive processes was increased in reticular nucleus and cortex combined after maternal hypoxia and was unchanged or less after nanoparticle exposure. This contrasted with the reduction in process length observed after exposing cortical cultures to foetal plasma or rat placenta conditioned medium. However, the directionality of the effect observed in vivo was replicated when cortical cultures were exposed to BeWo conditioned medium and when neuron-only cultures were exposed to media conditioned by rat placenta. The conditioned media after maternal nanoparticle injection caused an increased length of astrocyte processes in culture. In the brains the injection of MitoQ-NPs increased the overall length per field of view after maternal hypoxia but reduced it after maternal normoxia.

In general, the actions of MitoQ-NP treatment were also similar across the different models and either significantly prevented the effects of hypoxia, or caused hypoxia to no longer have significant effects, on dendrite length, TH ⁺ cell process length, GluN1 receptor subunit staining and astrocyte-to-neuron ratio. Additionally, the numerical density of parvalbumin-containing cells was significantly reduced in TRN, CTX and SSC of offspring following maternal hypoxia and rescued in the TRN by MitoQ-NP injection. Total cell count was not affected by either hypoxia or MitoQ-NP application in vitro or in the TRN, whereas a reduction in cell number was observed in the CTX. The number of TH ⁺ cells were reduced in cortical cultures only after additional severe hypoxia was applied to the placenta ex vivo and no change was observed in vivo.

It has been demonstrated that lipopolysaccharide, which shows minimal transport across the blood-brain barrier, may nonetheless have toxic actions within the brain by acting on astrocytes located at the blood-brain barrier (Hasegawa-lshii, S. et al. Scientific reports 2016, 6, 25457). It has also been noted that glutamate may play a role in mediating the effects of hypoxia- conditioned media on cortical cultures (Curtis, D.J. et al., Experimental Neurology 2014, 261, 386-395). In the light of these observations, the effects of media conditioned by rat placenta after maternal hypoxia on astrocyte signalling to neurons were tested with and without glutamate blockade. Dendrite lengths and GluN1 staining intensity were decreased in mixed astrocyte/neuron cultures, unchanged in neuron-only cultures and decreased if astrocyte cultures were first exposed and then this astrocyte-conditioned medium was applied to neuron- only cultures. Moreover, these effects were blocked by the NMDA receptor antagonist MK- 801 (dizocilpine). This suggests a role for astrocyte-to-neuron signalling via glutamate after exposure to the conditioned media. In contrast, TH ⁺ cell processes, although decreased in mixed cultures and after exposure to the astrocyte conditioned media, were increased in neuron-only cultures, an effect that was not altered by the glutamate antagonist. This points to a more complex regulation.

Conclusion

Oxidative stress and hypoxia during pregnancy causes placental and foetal pathology including low birth weight and disease in later life. The placenta plays a key role in foetal programming.

Gestational hypoxia in a rat model results in a series of adverse outcomes in the offspring, including a reduction in birthweight, structural/morphological changes in neurons, which are relevant to neurodevelopmental disorders, and changes in the brain transcriptome, despite the absence of oxidative stress in the foetal brain. The composition of microRNAs and proteins, of relevance to brain, foetal development and neurodevelopmental disorders, was altered in the foetal plasma and in media conditioned by placenta after maternal hypoxia or by model placental trophoblast barriers with altered oxygen. Exposing embryonic cortical cultures in vitro to foetal plasma or conditioned media after hypoxia resulted in changes that mimicked those observed in the offspring brain in vivo, suggesting a potential mechanistic link between molecules secreted from the placenta and the effects of gestational hypoxia on the offspring brain.

A single intravenous injection of MitoQ nanoparticles at the start of hypoxia has been demonstrated to reach the placenta but not the foetus, to prevent oxidative stress in maternal brain and liver and placenta (without effect in foetal brain and liver) and to prevent or alleviate the observed outcomes of maternal hypoxia (including altered release of molecules from the placenta and changes in the foetus including low birth weight and structural changes in the brain in later life). In a variety of models of the placenta in tissue culture, a common theme of change with hypoxia and with nanoparticle exposure was observed.