Reduced midbrain dopamine neuron number in the adult non-human primate brain after fetal radiation exposure

Abstract

Early gestation is a neurodevelopmental period that is especially vulnerable to environmental insult and one in which neurogenesis features prominently. Prenatal perturbation during early gestation has been linked to neuropsychiatric illnesses such as autism and schizophrenia, and severe environmental insult during this period can result in profound mental impairment. Midbrain dopamine neurons are generated during early gestation and play a key role in the motor, cognitive and reward circuitries implicated in neuropsychiatric disease and addiction. This study examined the impact of curtailing neurogenesis in early gestation on neuron number in the midbrain dopamine group, i.e., the substantia nigra and contiguous ventral tegmental area. Rhesus macaque monkeys were exposed in utero on embryonic days 39-41 to x-irradiation (3-4 exposures of 50 cGy over 3-7 days totaling <200cGy) and allowed to mature to full adulthood. Stereologic cell counts of tyrosine hydroxylase-positive neurons in the midbrain dopamine group were performed in adult monkeys, as were measurements of somal size. Mean total neuron number in the irradiated monkeys was significantly reduced on average by 33% compared to that of the control group. Somal size did not differ between the groups, suggesting that the integrity of survivor populations was not impacted. Reduced midbrain dopamine neuron number in fetally irradiated, adult monkeys indicates that radiation exposure during the critical period of neurogenesis results in an enduring reduction of this population and underscores the susceptibility of early neurodevelopmental processes to irreversible damage from environmental exposures.

Introduction

The first trimester of gestation is a neurodevelopmental period that is particularly vulnerable to a wide array of environmental impacts. For example, exposure to ionizing radiation or viral infection in early gestation causes pronounced disruption of normal neurodevelopment and in extreme cases results in severe mental impairment (Otake et al., 1991; Loganovskaja and Loganovsky, 1999; Ticconi et al., 2016). In addition, considerable evidence suggests that more subtle forms of mental compromise, as found in neuropsychiatric diseases such as schizophrenia and autism, are rooted in disturbances in early development (Brown, 2011; 2012). Cell proliferation is the most prominent developmental process occurring in early human gestation, and therefore interference with neurogenesis during this period is likely to be a prominent, underlying mechanism for abnormal development in neuropsychiatric illness (Selemon and Zecevic, 2015). Notably, many subcortical neurons, including monoamine neuron populations, are generated in the first trimester of gestation in the non-human primate (Rakic, 1977; Ogren and Rakic, 1981; Levitt and Rakic, 1982; Kordower and Rakic, 1990; Kordower et al., 1992).

The impact of radiation on the developing brain has been studied extensively in the mouse (Verreet et al., 2015; 2016). It is particularly noteworthy that the timing of radiation exposure is a critical determinant of the resulting behavioral deficits, as for instance radiation exposure at embryonic day 11 (E11) in the mouse results in both spatial and reference memory deficits whereas following E10 exposure spatial memory is spared (Verreet et al., 2016). Previous studies of fetal irradiation in the non-human primate have also established that the pathology associated with x-irradiation is dependent on the timing of exposure (Algan and Rakic, 1997; Schindler et al., 2002; Selemon et al., 2005; 2009; 2013). Indeed, the prolonged gestational period of neurogenesis in the non-human primate brain allows for targeting of subcortical versus cortical populations of neurons. Timing is key to specificity: this is evident from comparison of cortical pathology following early and mid-gestational exposure to radiation. Radiation exposure before the onset of cortical neurogenesis reduces cortical surface area and volume but not cortical thickness, pathology that is consistent with a reduction in progenitor cell populations (Selemon et al., 2013). In contrast, exposure during the period of cortical neurogenesis reduces cortical neuron number resulting in a thinner cortical ribbon (Selemon et al., 2013). These results are consistent with developmental studies indicating that symmetric division of progenitor cell increases cortical surface area while asymmetric division, i.e. neurogenesis, generates the six layers of neurons in the cortical mantle (Rakic, 1995; Selemon and Zecevic, 2015). Early cell culture studies established that x-irradiation is lethal to dividing cells (Sinclair and Morton, 1966). More recently, in vivo studies in mice have found evidence for radio-sensitivity in non-dividing populations with a single dose exposure of 1 Gy (100cGy)(Vermeet et al., 2015; 2016). Importantly, a previous study of fetal irradiation in the non-human primate showed that exposures ≤200 cGy resulted in specific elimination of cell populations undergoing final mitosis, i.e., those in the process of neurogenesis (Algan and Rakic, 1997). Thus, we were mindful of this threshold in choosing the dosage for this study to avoid lethal effects on non-diving populations. The radiation exposure timeframe in this study E30-E41 occurred during neurogenesis of midbrain dopamine neuron populations, an epoch roughly equivalent to E10-11 in rodents (Specht et al., 1981; Foster et al., 1988).

Midbrain dopamine neurons perform a multitude of functions, including modulation of motor planning, reward processing, and cognition, through projections to the striatum, nucleus accumbens, and neocortex (Haber and Knutson, 2010; Goldman-Rakic, 1999). Dysregulation of the dopamine system has been implicated in neuropsychiatric diseases such as Parkinson’s disease, schizophrenia, and attention deficit hyperactivity disorder and is a central mediator of addictive behavior and substance abuse (Hirsch, 1994; Goldman-Rakic and Selemon, 1997; Tanaka, 2006; Krause, 2008; Selemon, 2014). Moreover, monoaminergic cell populations project transiently into the cortical subcortical plate during development and thereby may play a role in establishing cortical connectivity (Bystron et al., 2005; Zecevic and Verney, 1995). Although Parkinsonism is not generally thought to be a developmental disorder, a study in rats has demonstrated that fetal exposure to a bacteriotoxin reduces dopamine neuron number in adult rats with continued loss of dopamine neurons throughout the lifetime of the animal, suggesting that fetal perturbation can lead to a Parkinson-like pathology (Carvey et al., 2003).

This study examines whether midbrain dopamine neurons of the substantia nigra (SN) and contiguous ventral tegmental area (VTA) are reduced in number and/or size following early gestational exposure to x-irradiation. While rare, radiation therapy is sometimes necessary during pregnancy (Mayr et al., 1998; Fenig et al., 2001; Kal and Struikmans, 2005), yet the long-term effects of this treatment on the brain are not known. Our findings indicate that prenatal exposure to radiation leads to an enduring, if not permanent, deficit in midbrain dopamine populations. Notably, other prenatal exposures, for example elevated cytokines following maternal infection or increased glucocorticoid concentrations in response to maternal stress, also impact cell proliferation (Deverman and Patterson, 2009; Anacker et al., 2013). Furthermore, gene expression of disrupted in schizophrenia 1 is associated with progenitor cell proliferation (Mao et al., 2009). Therefore, the present findings may provide insight into the adult pathology consequent to a variety of genetic and environmental factors as they converge on the process of neurogenesis in the first trimester of gestation.

Experimental Procedures

Subjects

Ten rhesus macaque monkeys were subjects in this study (Table 1). Five were exposed in utero to x-irradiation as described below (FIMs), and five were either sham irradiated (N=4) or non-irradiated (N=1) controls (CONs). The FIM group was comprised of 3 males and 2 females; the CON group was comprised of 2 males and 3 females. Animals are designated in Table 1 by group (CON or FIM) followed by first initial of the animal’s name in capitals and then sex in small letters, as for example CON-Mm. All animals in this study were housed, fed, and experimentally treated in accordance with protocols approved by Yale’s Institutional Animal Care and Use Committee and the National Institute of Health Guide for the Care and Use of Laboratory Animals. The number of animal subjects was kept to the minimum necessary for study; these animals were not subject to any procedures that involved pain or suffering.

Table 1:

Dose and Exposure to Ionizing Radiation

Animal	Sex	Age at Sac (yrs)	Time in Embronic Day (E) and Irradiation Dose (in cGy)	Total Dose Symbol	Graph
CON-Mm	M	5.83	4 doses ketamine - E29,E32,E34,E36	0	Δ
CON-Om	M	6.04	2 doses ketamine - E34,E36	0	▷
CON-Vf	F	5.58	3 doses ketamine - E39,E41, E43	0	◁
CON-Jf	F	5.67		0	▿
CON-Lf	F	5.25	4 doses ketamine - E32,E34,E37,E39	0	◯
FIM-Tm	M	5.96	E34(50), E36(50), E38(50), E41(50)	200	▴
FIM-Sf	F	5.96	E32(50), E34(50), E36(50)	150	▶
FIM-Bm	M	5.67	E35(50), E37(50), E38(50)	150	◀
FIM-Lf	F	6.08	E30(50), E33(50), E35(50), E37(50)	200	▼
FIM-Am	M	5.17	E33(50), E35(50), E37(50)	150	●

In utero exposure to x-irradiation

The irradiation procedures used were described previously in detail (Algan and Rakic, 1997) and are only briefly described here. Pregnant monkeys were sedated with ketamine (5-10 mg/kg), administered in conjunction with atropine (0.02 mg/kg). Ultrasound was used to guide a beam of radiation delivered by a 250 kV, 15 mA Stabilipan x-ray tube to the head of the fetus. Informed by the Algan and Rakic (1997) study, which showed that exposures of ≥ 200 cGy resulted in specific loss of populations undergoing neurogenesis, we opted to administer multiple (3-4) low doses (50 cGy) of x-irradiation, usually every other day, for a total dose exposure of 150-200 cGy (Table 1). Fetal irradiation occurred at gestational days E30 - E41 (Table 1). Note that gestation in rhesus monkeys lasts 165 days. Four of the five CONs were subjected to sham irradiation, i.e., the pregnant mothers were sedated with ketamine, and ultrasound was used to localize the fetus but radiation was not administered (Table 1). Following irradiation or sham irradiation, the status of pregnant monkeys was carefully monitored and viability of the fetus was confirmed by ultrasound throughout the remainder of the gestational period.

Primate birth and housing

Eight of the ten monkeys in this study were born naturally at term and allowed to stay with their mothers until 6 months of age. At this time, six of the eight were pair-housed and two were housed alone (CON-Vf and FIM-Sf) in the adult non-human primate colony at Yale University School of Medicine. Because gender and age had to be considered, FIM and CON monkeys were not always housed with each other; some were paired with other animals in the colony. Two of the ten monkeys were raised in the primate nursery from birth. Natural birth of CON-Lf did not occur within a safe interval of the expected delivery date, thereby necessitating a Caesarian section and placement of the newborn in the primate nursery. FIM-Am was born naturally but rejected by his mother, again requiring housing in the nursery. All ten monkeys were healthy throughout life, and body weights did not differ between the groups (Aldridge et al., 2012). These ten monkeys had undergone behavioral and cognitive testing as juveniles and adults (Friedman and Selemon, 2010; Selemon and Friedman, 2013).

Sacrifice, brain perfusion and processing

Monkeys were sedated with ketamine (5-10 mg/kg IM) and deeply anesthetized with sodium pentobarbital (100 mg/kg IV), then perfused intracardially with 0.1 M phosphate buffered saline followed by a mixed aldehyde fixative (4 % paraformaldehyde + 0.1 % glutaraldehyde + 15% saturated picric acid) in 0.1 M phosphate buffer (pH 7.4, 4° C). A block containing the entire left midbrain was embedded in celloidin. The block was serially sectioned coronally at 40 μm, and every tenth section was immunocytochemically stained for tyrosine hydroxylase immunoreactivity.

Tyrosine hydroxylase (TH) Immunocytochemistry

Pairs of brains, one irradiated and one control, were reacted in parallel for immunocytochemical localization of TH. Prior to immunocytochemical reaction, the celloidin embedding medium was removed from the sections by placing the sections in a NaOH solution (50 g NaOH in 170 ml methanol) for 30 minutes until the celloidin was no longer visible and then performing two 100% methanol washes, followed by a 10-minute soak in 10% H₂O₂ (diluted in methanol) and a 15-minute immersion in a 70% methanol solution (diluted in H₂O). Sections were rinsed three times in 0.1M phosphate buffer (PB) for 10 minutes each. A mouse monoclonal antibody (Chemicon International) was used to react the sections. The control sections were reacted without the primary antibody. Sections were first treated in avidin in blocking serum for one hour at 4°C, rinsed in 0.1M PB, and then transferred to biotin in blocking serum for one hour. Sections were reacted with ant-TH mouse immunoglobulin (IgG) at 1:1000 for 48 hours at 4°C. Sections were rinsed three times in 0.1M PB at room temperature, treated with anti-mouse biotinylated IgG (1:200) for two hours at room temperature, again rinsed three times in 0.1M PB, and incubated in ABC solution for 1 hour at room temperature. All steps were performed with agitation. Sections were then rinsed in 0.1M PB and reacted with 3,3’-diaminobenzidine (DAB) for approximately 3-5 minutes at room temperature with agitation until the TH staining appeared as a dark rust color stain against a neutral background. The sections were mounted on slides, and the slides were coded so that analysis could be performed blind to experimental status.

Analysis of total number of TH-positive neurons

Total number of TH-positive neurons in the midbrain dopamine group was estimated using the optional disector method in conjunction with the fractionator method of random sampling (Gundersen et al., 1988). These methods were applied with the aid of a MicroBrightfield Neuroluciva system (Williston, VT) and Stereoinvestigator v10.0 software. In every tenth section through the midbrain dopamine group, contours were drawn at low magnification (2.5x objective) in such a manner as to include all TH-labeled cells and fibers. These contours included the substantia nigra (SN) and contiguous ventral tegmental area (VTA). Sections, initially cut at 40 μm thickness, measured on average 15 μm in thickness following immunocytochemical processing. TH-positive cells were then counted at high magnification (40x objective) with disectors measuring 200μm (x) x 125 μm (y) x 10 μm (z) and a sampling grid of 500 μm x 500 μm. Exclusion planes on the lower, front, and left margins of the disectors were applied, as well as a 2 μm guard zone above and below (z axis) the box within the height of the section. The stereologic analysis generated an estimate of volume of the midbrain DA group. Stereologic recounts were performed to assess reliability of the data.

Analysis of TH-positive cell size

TH neuronal size was examined using the nucleator probe in conjunction with fractionator sampling (Gundersen et al., 1988). The counting frame measured 200 μm width (x), 125 μm height (y) x 10 μm depth (z), and the sampling grid was 500 μm x 500 μm. Two rays were projected onto each cell for measurement.

Statistics

Data are expressed as mean ± standard deviation of the mean. Statistical analyses were performed with the aid of the SPSS Statistics software (ver, 24). Paired sample, 2-tailed t-tests were used to test for significant differences in TH-positive neuron number, TH-positive somal size, and volume of the TH-positive contour between irradiated and control brains. A 2-tailed Pearson correlation was used to assess data validity in repeat analyses of TH-positive neuron number. All tests were considered significant at the p < 0.05 level.

Results

Description of TH staining

The intensity and patterning of TH labeling in the midbrain dopamine group did not differ between control and irradiated groups (Fig. 1, A,B). At the dorsolateral margin of the midbrain group, a small number of large neurons were present in a mesh of sparsely packed fibers (Fig. 1C,G). Ventrolaterally, slightly smaller and less darkly stained TH-positive neurons were densely packed and embedded in a nexus of crossing fibers. In addition, large, darkly stained neurons extended into the ventrally projecting fiber bundles (Fig. 1D,H). Throughout all rostral-caudal levels of the midbrain dopamine group, a compact stratum of TH-labeled soma and fibers was present dorsally (Fig. 1 E,I). Most of the labeled neurons in this layer were large and darkly stained. At the medial margin of this dorsal layer, lighter labeled neurons were densely packed and surrounded by fibers. A sparse distribution of large, darkly stained neurons was present in these fiber bundles as well. The ventromedial most area of the dopamine midbrain group housed a cluster of large, darkly stained cell bodies (Fig. 1F, J). Small and lightly stained neurons, some with elongated dendrites, were scattered throughout this area of labeled fibers. The dorsal stratum of TH-positive neurons described here corresponds to the dorsal tier of dopamine neurons that includes the SN compacta and the VTA whereas the subjacent area of smaller TH-positive neurons and those populating the ventrally projecting fibers correspond to the SN reticulata (Haber, 1995).

Figure 1.

(A) Low power view of midbrain tyrosine-hydroxylase (TH) staining in a sham irradiated monkey (CON-Om), and (B) corresponding staining in a fetally irradiated monkey (FIM-Sf). (C,D,E, F) High power views of TH-positive neurons in CON-Om, and (G,H,I,J) corresponding high power views in FIM-Sf.

Total TH-positive neuron number

Total number of TH- positive neurons in the irradiated group (97,800 ± 45,205) was significantly reduced compared to that of the control group (145,680 ± 38,309; t = 6.987; p = .002). Correlation between total neuron number estimated in the initial stereologic analysis and that in the recount was significant (R²= 0.786; p = 0.007). Notably in every pair of irradiated and control brains that were immunocytochemically reacted in parallel, total number was smaller in the irradiated brain (Fig. 2).

Figure 2.

Graphs showing (left) total number of neurons, (middle) volume of the midbrain TH-positive contour, and (right) TH-positive somal size in control monkeys (open symbols) and fetally irradiated monkeys (solid symbols). The symbols identify individual monkeys as shown in Table 1, and the same symbols (open and solid) represent monkey pairs in which tissue was immunoreacted in parallel. Note that for each monkey pair, the total number of TH-positive neurons is smaller in the fetally irradiated monkey compared to the paired control monkey.

Volume of the midbrain TH-positive contour

Volume of the TH-positive contour based on stereologic analysis was not significantly different in the control and irradiated groups (t = 1.35; p = 0.248) (Fig. 2).

TH-positive cell size

Somal size was measured to provide an index of neuronal integrity and to ensure that changes in neuronal counts were not skewed by changes in cell size. Mean areal somal size did not differ between the control (331.2 + 35.0 μm²) and irradiated groups (309.6 + 34.2 μm²; t = 1.717; p = 0.161) (Fig. 2).

Discussion

This study demonstrates that midbrain dopamine neurons are vulnerable to environmental insult in early gestation when this neuronal population is proliferating. Importantly, our findings indicate that when the number of midbrain dopamine neurons is reduced in early gestation, in this instance via exposure to ionizing radiation, the deficit persists into adulthood. These observations are consistent with long-standing developmental studies establishing that each brain region has a limited developmental period in which neurons are spawned from the progenitor population, and that once neurogenesis is complete, no new neurons are generated (Rakic, 2002). Fortunately, prenatal radiation exposure is a rare occurrence in human populations. However, there are circumstances in which pregnant women must undergo radiation treatment for cancer (Mayr et al., 1998; Fenig et al., 2001; Kal and Struikmans, 2005), and therefore, it is important to be able to weigh the risks to the unborn child against the medical needs of the mother. Our findings suggest that radiation exposure prenatally can result in seemingly permanent neuronal deficits.

Two tragic historic events, the atomic bombing of World War II and nuclear reactor accident in Chernobyl, have shown that there are deleterious effects of high-dose radiation on the developing human brain. Individuals exposed in utero at 8-25 weeks gestation to radiation from the atomic had decreased head size, severe mental retardation and seizure disorders (Otake et al., 1991). Those exposed in this same early prenatal period to radiation from the Chernobyl disaster exhibited disturbances in electroencephalographic patterning, decrease intelligence quota, language disorders, as well as behavioral and emotional abnormalities (Loganovskaja and Loganovsky, 1999). While a definitive link between prenatal radiation exposure and neuropsychiatric illness has not been established, it has been proposed that radiation exposure generally may be a risk factor for schizophrenia spectrum disorders (Loganovsky et al., 2005).

The link between radiation exposure and dopamine neuron deficits is not well established; however, there is evidence that dopamine populations are vulnerable to prenatal insult. For example, a study in rats showed that prenatal exposure to a bacteriotoxin caused a permanent loss of dopamine neurons, with neuronal loss increasing with age as it does in Parkinson’s disease (Carvey et al., 2003; Emborg et al., 1998). Moreover, the bacteriotoxin was most toxic when administered at E10, the time when dopamine neurons are generated (Specht et al., 1981). Thus, one possibility is that the toxin induced an immune-activated elevation of cytokines that in turn impacted the cell proliferation of dopamine neurons. It is noteworthy in this regard that immune associated responses in the fetus can have different outcomes. For example, a rodent model of fetal alcohol syndrome has shown that prenatal alcohol exposure triggers an inflammatory response in the rat VTA which in turn results in smaller somal size of VTA neurons, but not a reduction in neuronal number (Aghaie et al., 2020). Interestingly, exposure to bisphenol A in the last two months of gestation in the non-human primate decreases the number of midbrain TH-positive neurons, although likely via a very different mechanism from prenatal irradiation (Elsworth et al., 2013). The authors postulate that antiestrogen effects, perhaps in conjunction with anti-thyroid actions of the compound, may reduce the number of TH-expressing neurons (Elsworth et al., 2013).

The non-human primates examined in the present study exhibited adult-onset cognitive impairment and behavioral disturbances (Friedman and Selemon; Selemon and Friedman, 2013). Our previous analyses have also shown that there is widespread, though subtle, pathology in the cortex and thalamus following early gestational radiation exposure (Schindler et al., 2002; Selemon et al., 2005; Selemon et al., 2009; Aldridge et al., 2012; Selemon et al., 2013). While not dismissing the contribution of cortical and thalamic pathology to these cognitive and behavioral disturbances, the present findings of diminished midbrain dopamine neuron number in these animals suggest that circuitry involving midbrain dopaminergic neurons, e.g., mesocortical and mesostriatal projections may in part account for the cognitive and behavioral abnormalities. Although dopaminergic neurons are generated in early gestation, dopaminergic projections to limbic, striatal, and cortical brain regions remain immature until late adolescence (Rosenberg and Lewis, 1995; Brenhouse et al., 2008). Thus, prenatal environmental factors that impact the neurogenesis of dopamine neurons in early gestation might go undetected until adolescence or adulthood when these dopaminergic systems reach maturity. As such, these findings underscore the potential for prenatal environmental exposure to have an insidious impact on mental capacity and brain function.

Relevance to schizophrenia

In a non-human primate model of schizophrenia, we have explored the anatomic and behavioral consequences of curtailing neurogenesis in early gestation via exposure to x-irradiation. Our previous findings have documented widespread volumetric deficits in nonhuman primates that had been irradiated in early gestation (Schindler et al., 2002; Selemon et al., 2005; Aldridge et al., 2012). Neuron number is reduced in the thalamus, particularly in the mediodorsal nucleus (Selemon et al., 2009), and the reduction in volume of the prefrontal cortex appears to be due to a decrease in cortical progenitor populations (Selemon et al., 2013). As previously mentioned, adult non-human primates exposed to x-irradiation in early gestation exhibit deficits on a spatial delayed working memory task, motor stereotypies and cognitive perseveration, despite that fact that their performance and behavior had been indistinguishable from controls as juveniles (Friedman and Selemon, 2010; Selemon and Friedman, 2013). These findings recapitulate key anatomic and behavioral aspects of schizophrenia, as for example cortical volume deficit, reduced number of neurons in the mediodorsal nucleus of the thalamus, and late adolescent onset of symptomatology in schizophrenia (Pakkenberg, 1990; Zipursky et al., 1992; Popken et al., 2000; Young et al., 2000)

While many of the features of schizophrenia are reproduced by early gestational radiation exposure, reduction in the population of midbrain dopamine neurons has not been found in postmortem studies of subjects with schizophrenia. When the SN/VTA in subjects with schizophrenia was compared to that of controls, decreased TH labeling of neuronal processes, reduced TH messenger ribonucleic acid, and decreased TH protein expression were observed in rostral regions of the midbrain dopamine group, but notably no deficit in the number of TH-positive neurons was found (Perez-Costas et al., 2012; Rice et al., 2016). These recent cell counts are consistent with an early study of melatonin-containing neurons in the SN/VTA that also found equivalent numbers of neurons in schizophrenia subjects and controls (Bogerts, 1983). Thus, fetal irradiation produces dopamine dysregulation that may mirror some aspects of the dysfunction present in schizophrenia, however, it appears that the mechanism of dysfunction in fetally irradiated monkeys differs from that present in subjects with schizophrenia.

Technical considerations

Prenatal exposure to x-irradiation is a useful tool to examine the effects of curtailing neurogenesis. The advantage of using radiation, rather than other epidemiologically identified exposures, is that we know how radiation impacts neurodevelopment. Radiation is lethal to dividing cells (Sinclair and Morton, 1966), and at relatively low doses, radiation specifically impacts only dividing cells (Algan and Rakic, 1997). The SN is generated in the period E36-E43; the VTA is generated E38-E43 (Levitt and Rakic, 1982). Notably, exposure to x-irradiation in this study would have impacted only the early stages of neurogenesis in these areas (Table 1). An even greater reduction of midbrain dopamine neuronal number might have been observed if the radiation exposure had spanned the entire period of neurogenesis.

Counts of TH-labeled neurons ranged from <50,000 in an irradiated monkey to >190,000 neurons in a control. However, these numbers were much more closely aligned for pairs of cases immune-processed in parallel, suggesting that variation in immunoreactivity between different processing runs is substantial and that our approach, e.g., to use a paired t-test as a statistical tool, is a valid strategy for dealing with this variability. Indeed, for every pair analyzed, TH-positive counts were greater in the control member of the pair. In contrast, volume of the midbrain dopamine group was not different in the irradiated and control groups. This negative finding implies that neuronal density is also reduced, though we did not specifically measure cell density and the reduction in density was not visibly evident in the stained sections.

The neuron counts in this study comport with previous counts of TH-positive neurons in normal, young adult non-human primate brain: 79,000 to 111,000 neurons (Emborg et al., 1998; Blesa et al., 2012). Although the mean number for control brains in this study is higher (~146,000), the additional counts in part may reflect the counting of contiguous ventral tegmental neurons (A10) in addition to the TH-positive neurons in the substantia nigra (A9), as the counts cited above were limited to A9 only. In the Blesa et al. (2012) study, when counts of A9 (79,000) and A10 (28,000) were combined, the total summed to 107,000 neurons. Emborg et al. (1998) did not include A10 neurons in their counts. In comparison, a quantitative analysis of pigmented neurons in the human substantia nigra reported a mean of 550,000 neurons (Pakkenberg, 1991).

Because the neurons counted were immunocytochemically labeled, one must consider whether the reduction in number observed in the irradiated monkeys represents an actual decrease in number of neurons or rather a reduction in TH staining. That is, if the TH protein content of normally TH-positive neurons is reduced to such an extent that these neurons no longer appear positive for TH, the reduction of number could reflect this protein decrease rather than a decrease in neuron number. However, the fact that neither the intensity nor the patterning of staining was different in the two groups (Fig. 1) supports the premise that the actual number of TH-neurons was smaller in the irradiated group. The fact that TH-positive somal size was not smaller in the irradiated group is further evidence indicating that TH staining properties did not account for the reduction in neuronal number.

Somal size was not different in the irradiated and control monkeys. This finding may indicate that the survivor population of dopamine neurons differentiated and grew normal dendritic trees, as the expanse of the dendritic arbor is generally thought to correlate with somal size (Coleman and Friedlander, 2002; Liu et al., 2017). Animal studies have shown that prenatal hypoxia, prenatal malnutrition, and fetal alcohol exposure can reduce neuronal somal size (Diaz-Cintra et al., 1994; Briscoe et al., 2006; Aghaie et al., 2020); these insults may also impact dendritic growth. The fact that somal size was not altered after fetal irradiation in this study is further evidence for the specificity of precisely timed radiation to impact neurogenesis of neuronal populations while not affecting later maturational processes.

Ionizing radiation exposure in early gestation causes a long-lasting, if not permanent, reduction of midbrain dopamine neuron number in the non-human primate brain. Furthermore, early gestational reduction of midbrain dopamine neurons may be associated with adult-onset working memory deficits, cognitive perseveration and motor stereotypies (Friedman and Selemon, 2010; Selemon and Friedman, 2013), suggesting that prenatal environmental exposures can have an insidious effect on mental functioning. These findings emphasize the delicate nature of early neurodevelopment and the incontrovertible need to protect the fetus from adverse external influences.

Cover Figure.

Tyrosine hydroxylase positive cell bodies and processes are shown in the substantia nigra of an adult macaque monkey.

• Early gestational exposure to x-irradiation decreased midbrain dopamine neuron number in the adult non-human primate

• Stereologic counts of tyrosine -positive neurons revealed on average a 33% reduction in midbrain dopamine neuron number

• Somal size of dopamine neurons did not differ between irradiated and control groups

• Environmental insult to the developing fetal brain can result in enduring, if not permanent, brain pathology

Abbreviations

DAB

3,3’-diaminobenzidine

CON

control monkey

FIM

fetally irradiated monkey

IgG

immunoglobulin

PB

phosphate buffer

SN

substantia nigra

TH

tyrosine hydroxylase

VTA

ventral tegmental area