Neuronal network function, Decreased leads to Impairment, Learning and memory

Long-term potentiation (LTP) is a long-lasting increase in synaptic efficacy and its discovery suggested that changes in synaptic strength could provide the substrate for learning and memory (reviewed in Lynch, 2004). Moreover, LTP is intimately related to the theta rhythm, an oscillation long associated with learning. Learning-induced enhancement in neuronal excitability, a measurement of neural network function, has also been shown in hippocampal neurons following classical conditioning in several experimental approaches (reviewed in Saar and Barkai, 2003).

On the other hand, memory requires the increase in magnitude of excitatory postsynaptic currents (EPSCs) to be developed quickly and to be persistent for few weeks at least without disturbing already potentiated contacts. Once again, a substantial body of evidence has demonstrated that tight connection between LTP and diverse instances of memory exist (reviewed in Lynch, 2004).

A review on Morris water maze (MWM) as a tool to investigate spatial learning and memory in laboratory rats also pointed out that the disconnection between neuronal networks rather than the brain damage of certain regions is responsible for the impairment of MWM performance. Functional integrated neural networks that involve the coordination action of different brain regions are consequently important for spatial learning and MWM performance (D'Hooge and De Deyn, 2001).

Moreover, it is well accepted that alterations in synaptic transmission and plasticity contribute to deficits in cognitive function. There are a number of studies that have linked exposure to TPO inhibitors (e.g., PTU, MMI), as well as iodine deficient diets, to changes in serum TH levels, which result in alterations in both synaptic function and cognitive behaviors (Akaike et al., 1991; Vara et al., 2002; Gilbert and Sui, 2006; Axelstad et al., 2008; Taylor et al., 2008; Gilbert, 2011; Gilbert et al., 2016), described in the indirect KER "Decrease of TH synthesis leads to learning and memory deficits".

Developmental hypothyroidism reduces the functional integrity in brain regions critical for learning and memory. Neurophysiological indices of synaptic transmission of excitatory and inhibitory circuitry are impaired in the hippocampus of hypothyroid animals. Both hippocampal regions (area CA1 and dentate gyrus) exhibit alterations in excitatory and inhibitory synaptic transmission following reductions in serum TH in the pre and early postnatal period (Vara et al., 2002; Sui and Gilbert, 2003; Sui et al., 2005; Gilbert and Sui, 2006; Taylor et al., 2008; Gilbert, 2011; Gilbert et al., 2016). These alterations persist into adulthood despite a recovery to euthyroid conditions in blood. The latter observation indicates that these alterations represent permanent changes in brain function caused by transient hormones insufficiencies induced during critical window of development.  

Because the adult hippocampus is involved in learning and memory, it is a brain region of remarkable plasticity. Use-dependent synaptic plasticity is critical during brain development for synaptogenesis and fine tuning of synaptic connectivity. In the adult brain, similar plasticity mechanisms underlie use-dependency that underlies learning and memory, as exhibited in LTP model of synaptic memory. Hypothyroidism during development reduces the capacity for synaptic plasticity in juvenile and adult offspring (Vara et al., 2002; Sui and Gilbert, 2003; Dong et al., 2005; Sui et al., 2005; Gilbert and Sui, 2006; Taylor et al., 2008; Gilbert, 2011; Gilbert et al., 2016). Decrease of neuronal network function and plasticity are observed coincident with deficits in learning tasks that require the hippocampus.

- Wang et al., 2012: This study showed that maternal subclinical hypothyroidism impairs spatial learning in the offspring, as well as the efficacy and optimal time of T4 treatment in pregnancy. Female adult Wistar rats were randomly divided into six groups: control, hypothyroid (H), subclinical hypothyroid (SCH) and SCH treated with T4, starting from GD10, GD13 and GD17, respectively, to restore normal TH levels. Results indicate that progenies of SCH and H groups demonstrated significantly longer mean latency in the water maze test (on the 2^nd training day, latency was ~83% higher in H group, and ~50% higher in SCH), and a lower amplification percentage of the amplitude (~15% lower in H group, and 12% lower in SCH), and slope of the field excitatory postsynaptic potential (fEPSP) recording (~20% lower in H group, and 17% lower in SCH), compared to control group. T4 treatment at GD10 and GD13 significantly shortened mean latency and increased the amplification percentage of the amplitude and slope of the fEPSPs of the progeny of rats with subclinical hypothyroidism. However, T4 treatment at GD17 showed only minimal effects on spatial learning in the offspring. Altogether these data indicate direct correlation between decrease of neural network function and learning and memory deficits.

- Liu et al., 2010 This study assessed the effects of hypothyroidism in 60 female rats who were divided into three groups: (i) maternal subclinical hypothyroidism (total thyroidectomy with T4 infusion), (ii) maternal hypothyroidism (total thyroidectomy without T4 infusion), and (iii) control (sham operated). The Morris water maze tests revealed that pups from the subclinical hypothyroidism group showed long-term memory deficits, and a trend toward short-term memory deficits.

- Gilbert and Sui, 2006 Administration of 3 or 10 ppm PTU to pregnant and lactating dams via the drinking water from GD6 until PND30 caused a 47% and 65% reduction in serum T4, in the dams of the low and high-dose groups, respectively. Baseline synaptic transmission was impaired in PTU-exposed animals: mean EPSP slope (by ~60% with 10 ppm PTU) and population spike amplitudes (by ~70% with 10 ppm PTU) in the dentate gyrus were reduced in a dose-dependent manner in adult offspring of PTU-treated dams. High-dose animals (10 ppm) demonstrated very little evidence of learning despite 16 consecutive days of training (~5-fold higher mean latency to find the hidden platform, used as an index of learning).

- Gilbert et al., 2016 Exposure to PTU during development produced dose-dependent reductions in mRNA expression of nerve growth factor (Ngf) in whole hippocampus of neonates. These changes in basal expression persisted to adulthood despite the return to euthyroid conditions in blood. Developmental PTU treatment dramatically reduced the activity-dependent expression of neurotrophins and related genes in neonate hippocampus and was accompanied by deficits in hippocampal-based learning (e.g., mean latency to find a hidden platform, at 2^nd trial resulted ~60% higher in rats treated with 10 ppm PTU).

- Gilbert, 2011 Trace fear conditioning deficits to context and to cue reported in animals treated with PTU and who also displayed synaptic transmission and LTP deficits in hippocampus. Baseline synaptic transmission was impaired in PTU-exposed animals (by ~50% in animal treated with 3 ppm PTU). EPSP slope amplitudes in the dentate gyrus were reduced in a dose-dependent manner in adult offspring of PTU-treated dams.

BPA, an environmental toxicant known to inhibit NIS-mediated iodide uptake (Wu Y et al., 2016) has been found to cause learning and memory deficits in rodents as described below:

- Jang et al., 2012 In this study, pregnant female C57BL/6 mice (F0) were exposed to BPA (0.1-10 mg/kg) from gestation day 6 to 17, and female offspring (F2) from F1 generation mice were analysed. Exposure of F0 mice to BPA (10 mg/kg) decreased hippocampal neurogenesis (~ 30% decrease of hippocampal BrdU⁺ cells vs control) in F2 female mice. High-dose BPA (10 mg/kg) caused neurocognitive deficit (i.e., reduced memory retention) as shown by passive avoidance testing (~ 33% decrease vs control) in F2 mice. Furthermore, 10 mg/kg BPA decreased the hippocampal levels of BDNF (~ 35% lower vs control) in F2 mice. These results suggest that BPA exposure (NIS inhibitor) in pregnant mothers could decrease hippocampal neurogenesis (decreased number of neurons) and cognitive function in future generations.

In humans, the data linking these two specific KE are much more limited, but certainly clear reductions in IQ, with specific impairments in hippocampus-mediated functions have been observed.

- Wheeler et al., 2015 This study assessed hippocampal functioning in adolescents with congenital hypothyroidism (CH), using functional magnetic resonance imaging (fMRI). 14 adolescents with CH and 14 typically developing controls (TDC) were studied. Hippocampal activation was greater for pairs than items in both groups, but this difference was only significant in TDC. When the groups were directly compared, the right anterior hippocampus was the primary region in which the TDC and CH groups differed for this pair memory effect. Results signify that adolescents with CH show abnormal hippocampal functioning during verbal memory processing, in order to compensate for the effects induced by TH deficit in the brain.

- Wheeler et al., 2012 In this study hippocampal neuronal network function was measured based on synaptic performance using fMRI and was altered while subjects engaged in a memory task. Data showed paired word recognition deficits in adolescents with congenital hypothyroidism (N = 14; age range, 11.5-14.7 years) compared with controls (N = 15; age range, 11.2-15.5 years), with no impairment on simple word lists. Analysis of functional magnetic resonance imaging showed that adolescents with congenital hypothyroidism had both increased magnitude of hippocampal activation relative to controls and bilateral hippocampal activation when only the left was observed in controls. Furthermore, the increased activation in the congenital hypothyroidism group was correlated with the severity of the hypothyroidism experienced early in life.

- Willoughby et al., 2013 Analogously, in this study, fMRI revealed increased hippocampus activation with word pair recognition task in CH and children born to women with hypothyroxinemia during midgestation. These differences in functional activation were not seen with single word recognition, but were revealed when retention of word pair associations was probed. The latter is a task requiring engagement of the hippocampus.

A series of important findings suggest that the biochemical changes that happen after induction of LTP also occur during memory acquisition, showing temporality between the two KEs (reviewed in Lynch, 2004).

- Morris et al., 1986 This study found that blocking the NMDA receptor of the neuronal network with AP5 inhibits spatial learning in rats. Most importantly, in the same study they measured brain electrical activity and recorded that this agent also inhibits LTP, however, they have not proven that spatial learning and LTP inhibition are causally related.

Since then a number of NMDA receptor antagonists have been studied towards their ability to induce impairment of learning and memory. It is worth mentioning that similar findings have been found in human subjects:

- Grunwald et al., 1999 By combining behavioural and electrophysiological data from patients with temporal lobe epilepsy exposed to ketamine, involvement of NMDA receptors in human memory processes was demonstrated.

The last KE preceding the AO (learning and memory deficits), i.e. "Decreased Neural Network Function", is also common to the AOP 13, entitled "Chronic binding of antagonist to N-methyl-D-aspartate receptors (NMDARs) during brain development induces impairment of learning and memory abilities" (https://aopwiki.org/aops/13). In this AOP 13, data on lead (Pb) exposure as reference chemical are reported. While these studies do not refer to TH disruption, they provide empirical support for the same KER described in the present AOP.

Pb2+: Exposure to low levels of Pb2+, during early development, has been implicated in long-lasting behavioural abnormalities and cognitive deficits in children (Needleman et al., 1975; Needleman and Gatsonis, 1990; Bellinger et al., 1991; 1992; Baghurst et al., 1992; Leviton et al., 1993; Needleman et al., 1996; Finkelstein et al., 1998; Lanphear et al., 2000; 2005; Canfield et al., 2003; Bellinger 2004; Lanphear et al., 2005; Surkan et al., 2007; Jusko et al., 2008; Neal and Guilarte, 2010) and experimental animals (Brockel and Cory-Slechta, 1998; Murphy and Regan, 1999; Moreira et al., 2001). Multiple lines of evidence suggest that Pb2+ can impair hippocampus-mediated learning in animal models (reviewed in Toscano and Guilarte, 2005).

- Jett et al., 1997 Female rats exposed to Pb²⁺ through gestation and lactation have shown more severe impairment of memory than male rats with similar Pb²⁺ exposures.

- De Souza Lisboa et al., 2005 This study reported that exposure to Pb²⁺ during both pregnancy and lactation caused depressive-like behaviour (detected in the forced swimming test) in female but not male rats.

- Anderson et al., 2012 This study investigated the neurobehavioral outcomes in Pb²⁺-exposed rats (250, 750 and 1500 ppm Pb²⁺ acetate in food) during gestation and through weaning and demonstrated that these outcomes are very much influenced by sex and rearing environment. In females, Pb²⁺ exposure lessened some of the benefits of enriched environment on learning, whereas, in males, enrichment does help to overcome detrimental effects of Pb²⁺ on learning. Regarding reference memory, environmental enrichment has not been beneficial in females when exposure to Pb²⁺ occurs, in contrast to males.

- Jaako-Movits et al., 2005 Wistar rat pups were exposed to 0.2% Pb²⁺ via their dams' drinking water from PND 1 to PND 21 and directly via drinking water from weaning until PND 30. At PND 60 and 80, the neurobehavioural assessment has revealed that developmental Pb²⁺ exposure induces persistent increase in the level of anxiety and inhibition of contextual fear conditioning. The same behavioural syndrome in rats has been described in Salinas and Huff, 2002.

- Finkelstein et al., 1998 These observations are in agreement with observations on humans, as children exposed to low levels of Pb²⁺ displayed attention deficit, increased emotional reactivity and impaired memory and learning.

- Kumar and Desiraju, 1992 In Wistar rats fed with lead acetate (400 µg/g body weight/day) from PND 2 until PND 60, EEG findings showed statistically significant reduction in the delta, theta, alpha and beta band EEG spectral power in motor cortex and hippocampus, but not in delta and beta bands power of motor cortex in wakeful state. After 40 days of recovery, animals were assessed for their neurobehaviour, and revealed that Pb²⁺ treated animals showed more time and sessions in attaining criterion of learning than controls.

Further data obtained using animal behavioral techniques demonstrate that NMDA mediated synaptic transmission is decreased by Pb²⁺ exposure (Cory-Slechta, 1995; Cohn and Cory-Slechta, 1993 and 1994).

- Xiao et al., 2014 Rat pups from parents exposed to 2 mM PbCl₂ three weeks before mating until their weaning (pre-weaning Pb²⁺) and weaned pups exposed to 2 mM PbCl₂ for nine weeks (post-weaning Pb²⁺) were assessed for their spatial learning and memory by MWM on PND 85-90. The study revealed that both rat pups in pre-weaning Pb²⁺ and post-weaning Pb²⁺ groups performed significantly worse than those in the control group. The number of synapses in pre-weaning Pb²⁺ group increased significantly, but it was still less than that of control group. The number of synapses in post-weaning Pb²⁺ group was also less than that of control group, although the number of synapses had no differences between post-weaning Pb²⁺ and control groups before MWM. In both pre-weaning Pb²⁺ and post-weaning Pb²⁺ groups, synaptic structural parameters such as thickness of postsynaptic density (PSD), length of synaptic active zone and synaptic curvature increased, whereas width of synaptic cleft decreased compared to controls.

The last KE preceding the AO (learning and memory deficits), i.e. "Decreased Neural Network Function", is also common to the AOP 17, entitled " Binding of electrophilic chemicals to SH(thiol)-group of proteins and /or to seleno-proteins during brain development leads to impairment of learning and memory" (https://aopwiki.org/aops/17). In this AOP 17, data on mercury exposure as reference chemical are reported. While these studies do not refer to TH disruption, they provide empirical support for the same KER described in the present AOP.

Sokolowski et al. 2013. Rats at postnatal day 7 received a single injection of methylmercury (0.6 microgr/g, that caused caspase activation in the hilus of granule cell layer in hippocampus. At PD 21, a decrease in cell number or 22% in hilus and of 27% in granule cell layer, as well as a decreased proliferation of neural precursor cells of 25% were observed. This was associated with a decrease of spatial memory as assessed by Morris water maze.

Eddins et al., 2008. Mice exposed during postnatal week 1-3 to 2-5 mg/kg mercury chloride in 0.01 ml/g of NaCl injectd s.c. The behavioral tests at 3 months of age revealed learning deficits (radial maze), which was associated with increased levels of monoamines in frontal cortex.

Zanoli et al., 1994. Single injection of methylmercury (8 mg/kg by gavage) at gestational day 15. Offsprings analyzed at 14, 21, and 60 days of age exhibited a decrease in the number of muscarinic receptors at 14 and 21 days and a decrease in avoidance latency at 60 days, indicating learning and memory deficits.

Zanoli et al., 2001. Single injection of methylmercury (8 mg/kg) at gestational day 8. Brain was removed at PD 21 and 60. An  increase in tryptophan level in hippocampus was detected at both days. At PD 21, a decrease in anthranilic acid and an increase in quinolinic acid was found. No change in glutamic acid nor in aspartic acid were detected.

Montgomery et al., 2008. C57/B6 mice exposed during pregnancy (GD 8-18) with food containing methylmercury (0.01 mg/kg body wheight). Tested when adult, they showed deficits in motor function, coordination, overall activity and impairment in reference memory.

Glover et al., 2009. Balb mice exposed to methylmercury in diet (low dose: 1.5 mg/kg; high dose: 4.5 mg/kg) during 11 weeks (6 weeks prior mating, 3 weeks during gestation and 2 weeks post-partum). Offsprings tested at PD 15 showed an accumulation of Hg in brain (0.08 mg/kg for low dose and 0.25 mg/kg for the high dose). At hte cellular level, there was alterations in gene expression for cytoskeleton, cell processes, cell adhesion, cell differentiation, development), which could be all involved in cellular network formation. This was associated with behavioral impairment, i.e. a decrease in exploratory activity measured in open field.

Onishchenko et al., 2007. Pregnant mice received 0.5 mg methylmercury/kg/day in drinking water from gestational dy 7 until day 7 after delivery. Offspring behavior was monitored at 5-15 and 26-36 weeks of age. Mercury-induced alterations in reference memory were detected.

Cagiano et al., 1990. Pregnant rat received at GD 15 8mg/kg of methylmercury by gavage. Offsprings were tested at day 16, 21 and 60. A reduced functional activity of glutamatergic system associated with disturbances in learning and memory were observed.

Rice, 1992. Female monkeys exposed to 10, 25 and 50 microg/kg/day to methylmercury. Male unexposed. Infants separated from mother at birth and exposed to similar doses did not show gross intellectual impairment, but interferences with temporal discrimination.

Sahin et al., 2016. Exposure of rat pups for 5 weeks or 5 months with mercury chloride (4.6 microg/kg as first injection, followed each day by 0.07 microg/kg/day). Learning and memory impairment measured by passive avoidance and Morris-water-maze was found in 5-weeks group, but not in the 5-month group. This was accompanied by hearing loss.

In humans:

Orenstein et al., 2014. Maternal peripartum hair mercury level was measured to assess prenatal mercury exposure. The concentrations of mercury was found in the range of 0.3-5.1 microg/g, similar to fish eating population in US. However, statistical analyses revealed that each microg/g increase in hair Hg was associated with a decrement in visula memory, learning and verbal memory.

Yorifuji et al., 2011. A survey of the Minamata exposed population made in 1971 to assess pre- and post-natal exposure revealed a methylmercury-induced impairment of intelligence as well as behavioral dysfunction.

One of the most difficult issues for neuroscientists is to link neuronal network function to cognition, including learning and memory. It is still unclear what modifications of neuronal circuits need to happen in order to alter motor behaviour as it is recorded in a learning and memory test (Mayford et al., 2012), meaning that there is no clear understanding about how these two KEs are connected.

The direct relationship of alterations in neural network function and specific cognitive deficits is difficult to ascertain given the many forms that learning and memory can take and the complexity of synaptic interactions in even the simplest brain circuit. Linking of neurophysiological assessments to learning and memory processes have, by necessity, been made across simple monosynaptic connections and largely focused on the hippocampus. Alterations in synaptic function have been found in the absence of behavioral impairments. This may result from measuring only one component in the complex brain circuitry that underlies 'cognition', behavioral tests that are not sufficiently sensitive for the detection of subtle cognitive impairments, and behavioral plasticity whereby tasks are solved by the animal via different strategies developed as a consequence of developmental insult.

Finally, in order to provide empirical support for this KER, data on the effects of lead (Pb) exposure are reported. Several epidemiological studies where Pb2+ exposure levels have been studied in relation to neurobehavioural alterations in children have been reviewed in Koller et al. 2004. This review has concluded that in some occasions there is negative correlation between Pb2+ dose and cognitive deficits of the subjects due to high influence of social and parenting factors in cognitive ability like learning and memory (Koller et al. 2004), meaning that not always Pb2+ exposure is positively associated with learning and memory impairment in children.

Mercury

Olczak et al., 2001. Postnatal exposure of rats to Thimerosal (4 injections with 12, 240, 1440 and 3000 microgHg/kg per injection). Effects were measured in adult, which exhibited alterations in dopaminergic system with decline in the density of striatal D2 receptors, with a higher sensitivity for males. No alterations in spatial learning and memory was observed, but impairments of motor activity, increased anxiety (open fiel measurment), which are other symptoms of autism spectrum disorder.

Franco et al., 2006. Lactational exposure of mice to methylmercury in drinking water (10 mg/L). Analysis at weaning revealed only impairment in motor performances.

Franco et al., 2007. Lactational exposure of mice with mercury chloride (0.5 and 1.5 mg/kg,  i.p. injection once a day).. At weaning , animals exhibited an increased level of mercury in cerebellum associated with motor deficit.

Cardenas et al., 2017 showed that maternal red blood cell mercury of 3.8 ng/g was associated to increased DNA methylation of PON1 in umbilical cord blood only in male and observed deficit in cognitive performances, such as visual motor ability, vocabiary and verbal intellgence.