Differences in Neuronal Numbers, Morphology and Developmental Apoptosis in Mice Nigra Provide Evidence of Ontogenic Origin of Vulnerability to Parkinson’s Disease


 Parkinson’s disease (PD) prevalence varies by ethnicity. In an earlier study we replicated the reduced vulnerability to PD in an admixed population, using 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-susceptible C57BL/6J, MPTP-resistant CD-1 and their F1 crossbreds. In the present study we investigated if the differences have a developmental origin. Substantia nigra was evaluated atpostnatal days 2 (P2), P6, P10, P14, P18, and P22. C57BL/6J mice had smaller nigra and fewer dopaminergic neurons than the CD-1 and crossbreds at P2, which persisted through development. A significant increase in numbers and nigral volume was observed across strains till P14. A drastic decline thereafter was specific to C57BL/6J. CD-1 and crossbreds retained their numbers from P14 to stabilize with supernumerary neurons at adulthood. The neuronal size increased gradually to attain adult morphology at P10 in the resistant strains, vis-à-vis at P22 in C57BL/6J. Accordingly, in comparison to C57BL/6J, the nigra of CD-1 and reciprocal crossbreds possessed cyto-morphological features of resilience, since birth. The considerably lesser dopaminergic neuronal loss in the CD-1 and crossbreds seen at P2, P14 and thereafter was complemented by attenuated developmental cell death. The differences in programmed cell death were confirmed by reduced TUNEL labelling, AIF and caspase-3 expression. GDNF expression aligned with the cell death pattern at P2 and P14 in both nigra and striatum. Earlier maturity of nigra and its neurons appear to be better features that reflect as MPTP-resistance at adulthood. Thus variable MPTP-vulnerability in mice and also differential susceptibility to PD in humans may arise early during nigral development.

expression was evaluated at P2 and P14 in the nigro-striatum. Our study provides experimental evidence that developmental apoptosis sculpts the nigra in terms of neuronal numbers and facilitates the identi cation of strain-speci c critical window of maximum susceptibility of DA neurons.

Mice strains:
The pregnant mice were housed in polypropylene cages under standard laboratory conditions with ad-libitum access to food and water. Age and gender matched neonates (n = 6/group/experiment) born following mating of C57BL/6J female and CD-1 males were named F1X1, while the reciprocal ones were named F1X2 [16]. The experiments were performed in the light period (08:00-18:00h) in accordance with the guidelines of the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA), New Delhi, India, that are based on NIH, USA guidelines.

Tissue processing for immunohistochemistry and TUNEL:
We anaesthetized neonates aged postnatal day 2 (P2), P6, P10, P14, P18 and P22 of each strain (n = 6/strain/time-point) by halothane inhalation and perfused intracardially with saline followed by 4% buffered paraformaldehyde (0.1M phosphate buffer; pH 7.4). The brains were quickly harvested and postxed for 24-48h at 4°C. Cryoprotected midbrains (10%, 20%, and 30% sucrose in 0.1M PB) were cryosectioned (Leica Microsystems, Germany) at 40µm. Serial sections were collected on gelatinized slides. Every third midbrain section from P2, P6, and P10 specimens and every sixth midbrain section from P14, P18, and P22 were immunostained, a slight modi cation of earlier reports [ was chosen. We carefully excluded the VTA during quanti cation as it was located more dorso-medial to the substantia nigra [32]. The parabrachial pigmented nucleus segregated the SN and VTA. The red nucleus is furthermoredorso-medial to VTA and has a characteristic circular appearance. Its parvocellular and magnocellular neurons rarely or almost never stained for TH, hence were easily excluded from quanti cation. Further, the retrorubral dopaminergic neurons that are present much caudally [33] were also not included for stereological evaluation. We counted the TH-ir cells using high power objective (100X), with a regular grid interval of 22500µm² (x = 150µm, y = 150µm) and counting frame of size 3600µm² (x = 60µm, y = 60µm). A guard zone of 4µm was implied on both sides resulting in an optical dissector of 17µm. Quanti cation was performed in both hemispheres and pooled to derive total numbers.

Statistical Analysis:
The data was analysed using two-way ANOVA followed by Tukey's post hoc test. The values were stated as mean ± SD and a p-value lower than 0.05 was considered signi cant.\ 3. Results:

Maturation of nigral architecture:
The SNpc and pars reticulata (SNpr) were indistinguishable across the strains at P2 (Fig. 1a1-d1) as described earlier [36]. The DA neurons were randomly distributed throughout the ventral midbrain and occasional clusters were noted along the midline between P2 and P6. Thus till P6 we did not parcellate them as SNpc or SNpr, but considered them as a single entity, i.e. substantia nigra (SN). A gradual distinction between the two tiers was achieved after P6 ( Fig. 1, a2, -d2). By P10 the SNpc DA neurons were identi able ( Fig. 1a3-d3). A clear adult architecture i.e. a compact band of DA neurons rich SNpc was observed at P14 in CD-1 (Fig. 1b4) and the crossbreds ( Fig. c4-d4), whereas this discrimination was relatively delayed in C57BL/6J and appeared at P18 (Fig. 1a4). Qualitative observations of different nigral subsections showed that, the DA neurons in the lateral nigra appeared at P14 in C57BL/6J compared to appearance at P10 in the other strains ( Fig. 1b3-a4, arrows).

Differences in numbers of DA neurons during postnatal development:
The C57BL/6J mice had signi cantly fewer TH -IR nigral neurons when compared to CD-1 and the crossbreds at P2 (Fig. 2a, Table 1, P2, . The numbers increased signi cantly thereafter in all the strains and the maximal increase was noted from P6 to P14 (Fig. 2a). During all the stages studies, the differences between C57BL/6J and rest of the strains were notable.  Thereafter a drastic fall in numbers was noted till P18, selectively in C57BL/6J (Fig. 2a, C57BL/6J, ***p < 0.001, P14 v/s P18). The reduction stabilized at P22 (Fig. 2a, C57BL/6J), and the numbers were comparable with the adult mice described earlier [16]. Despite the reduction between P14-P18, the numbers at P22 were signi cantly higher than those at P2. Interestingly, the CD-1 and crossbreds retained their numbers from P14 to stabilize with relatively more DA neurons at P18 and P22 (Fig. 2a). At P22, the CD-1 and crossbreds had approximately 40% more DA neurons than C57BL/6J (Table 1; Fig. 2a, . Effectively the quantitative differences between the strains persisted throughout the development.

Alterations in nigral volume:
Evaluation of nigral volume revealed two distinct observations. First that, a notable increase occurs in the nigral volume during development i.e. from P2 to P22 across all the mice strains (Fig. 2b). Secondly, at P2 and throughout development, the SN of C57BL/6J was much smaller than CD-1 and the crossbreds v/s P18). Thus the nigral volume was signi cantly higher in CD-1 and crossbreds and stabilized at P14, whereas in C57BL/6J it reduced till P18 and stabilized thereafter (Fig. 2b). Essentially the CD-1 and crossbreds possessed signi cantly voluminous SN at P22 than the C57BL/6J (Fig. 2b, $$ p < 0.01, v/s CD-1, FIX1 and F1X2).

TH expression during development and cellular morphology:
Across all strains, the DA neurons increased in size with developmental age (Fig. 3a1-d5; e). A closer comparison revealed that, the P2 neurons of both the parent strains, i.e., C57BL/6J and CD-1 were much larger than the crossbreds (Fig. 3,  The CD-1 and F1X2 showed relatively higher TH expression at P2 (Fig. 3f). Interestingly, unlike other observations, it was highest at P2 compared to the other stages (Fig. 3f). The expression stabilized to its adult levels at P14 in CD-1 and crossbreds (Fig. 3f), while at P22 in C57BL/6J (Fig. 3f, C57BL/6J, ***p < 0.001, P18 v/s P22).
In view of the appreciable differences in the soma size and the TH expression amongst strains, we quanti ed TH expression per unit soma area. TheF1X2 mice nigra showed higher TH expression/unit soma area at P2 compared to C57BL/6J (Fig. 3g, P2, p < 0.0001, C57BL/6J v/s F1X2). A signi cant decline in the ratio was noted until P10 in all the strains. It stabilized at P14 in CD-1, F1X1 and F1X2, whereas at P22 in C57BL/6J with a further decrement (Fig. 3g, C57BL/6J, **p < 0.01, P18 v/s P22).

Alterations in developmental apoptosis and cell death markers:
Developmental cell death was correspondingly evaluated, by TUNEL, at P2, P6, P10, P14, P18 and P22. TH immunoreactive (brown) cells labelled for TUNEL reaction (black) in the nucleus were identi ed as apoptotic DA neurons (Fig. 4. Ab). TUNEL-ir cells that were TH negative, but present within the nigral region were also considered ( Fig. 4. Aa). A positive control (Fig. 4. Ac) showed TUNEL reaction in almost all cells in the eld. Peak apoptosis was noted at P2 and P14 ( Fig. 4B & 4C, P14 *p < 0.05 C57BL/6J vs. F1X2) across the strains. The C57BL/6J had more TUNEL-TH co-labelled neurons, and persisted till P22; while in the CD-1 and the crossbreds, they subsided by P14.
The caspase-3 and AIF expression was evaluated exclusively in TH co-labelled neurons so as to assess the alterations in the surviving dopaminergic neurons.

GDNF expression in striatum and nigra:
A signi cant upregulation in GDNF expression was noted in C57BL/6J between postnatal days 2 and 14, in the striatum (

Discussion:
A few neuropsychiatric diseases like autism and attention de cit hyperactive disorder etc. are thought to be neurodevelopmental disorders while others like Huntington's, Alzheimer's, Parkinson's disease are designated as adult or elderly onset disorders [37][38][39][40][41]. It is being increasingly recognized that deregulation of developmental neuronal apoptosis or glial death may result in functional de cits and associated diseases in the later years of life [42]. In the absence of human tissues for analysis, animal models are valuable resources for mechanistic explanations of a disease. Ours is the rst study using unbiased stereology on developing nigra that compares the developmental trajectories of nigral DA neurons in two distinct mice strains and the F1 progeny of their crossbreds. The studies were conducted in MPTP-sensitive C57BL/6J, MPTP resistant CD-1 mice and the F1 generation of their reciprocal crossbreds. Our model offers the advantage of "no use of neurotoxins", since most other studies on PCD relied on chemical lesions to investigate target dependence in development.
The rst two postnatal weeks are crucial for the rodent midbrain DA system, for being associated with culmination of neuronal migration, maturation of perikarya, extension of axons and their terminal differentiation, which occurs alongside the apoptotic death of neurons that fail to mature [43,25,44,26].
Stressors during this period prime the DA system to subsequent insults like prenatal stress, maternal separation, ischemia, hypoxia and excitotoxicity etc.
resulting in the delayed attainment of DA phenotype [45], reduction in the number of DA neurons, exaggerated response to neurotoxins [46,47] and affect the DA neurotransmission at adulthood [48]. In fetal non-human primates [24] peak PCD of DA neurons occurs at E80 and approximately at mid-gestation in humans, when approximately 50% neurons are pruned. The time period P2 in rats equates with E79/80 in monkeys [24], thus P2 in mice is a critical window of vulnerability and protective modalities could be developed if the relevant molecular underpinnings are understood.
The clusters of TH-ir neurons found along the midline during early postnatal development i.e. P2 and P6 probably represent the migrating neurons or other dopaminergic clusters. The gradual increase in numbers till P14 point at the culmination of neuronal migration [49][50][51], or addition of TH-ir neurons by postnatal neurogenesis [52][53][54]49]. The third likelihood is that progressively more neurons show detectable levels of TH expression i.e. they attain DA phenotype [55,56,45]. These factors reduce the likelihood of extensive cell loss during this phase, thus maintaining the steady increase in numbers. In an earlier study using unbiased stereology, Jackson-Lewis et al. (2000) also reported a gradual increase in nigral dopaminergic neurons till P14. However, the small decline in the numbers after P14 observed in our study was contrasted by absence of cell loss in their study. Our observation is corroborated by the ndings of Oo and Burke [25], who showed a smaller second peak of dopaminergic neuronal loss around P14. The data of Jackson-Lewis et al. [23] represent cumulative results of two different mice strains i.e. C57BL/6J and CD-1, which may underlie the discrepancy. The delay in attainment of adult nigral architecture in C57BL/6J i.e. at P18, may be attributed to the ongoing cell loss vis-à-vis the early maturation at P14 in CD-1 and the crossbreds. Our observations on C57BL/6J, match those of Oo and Burke [25], who showed two peaks of loss i.e. at P2 and P14 in rats, proposing a biphasic nature of programmed cell death. We propose an additional minor peak at P22 in C57BL/6J, complemented by TUNEL reaction, caspase-3 expression and reduction in TH-ir dopaminergic neurons. Both caspase-mediated and caspase-independent cell death pathways, associated with mitochondria and endoplasmic reticulum respectively, work in unison to cause neurodegeneration [57]. The enhancement of AIF corroborates with the role of caspase independent cell death in developing human midbrain [58]. Elevated AIF expression at P14 in addition to high caspase-3 expression at P22, corroborated by TUNEL-expression, further illustrate the alliance between the two cell death pathways in the instance of DA neurons.
Classical studies suggest that GDNF protects nigral neuron from apoptosis [59,60]. Cellular hypertrophy following second peak of apoptosis noted exclusively in C57BL/6J may be a physiological process corresponding to the expansion of nerve terminals, also seen in other amine systems [49,[66][67][68]. Alternatively it may be a predisposing factor; since larger neurons are vulnerable to degeneration [69]. Hypertrophy of DA neurons in aging human nigra was considered as a compensation for sub-threshold neurodegeneration [10] and its absence as a marker for resilience [12]. The gradual decrease in cellular TH over development alongside an increase in soma area, suggests that overall dopamine synthesis remains stable through development. Amongst the strains, F1X2 showed best preservation of TH levels.
A correlation between the nigral volume and the neuronal numbers through development and across strains indicates that volume can be an indirect measure of DA neuronal number. Nigral volumetry by MRI is a diagnostic tool for PD, based on a similar premise. The striatal volume varies in healthy individuals [70,71] and imparts differential susceptibility to psychiatric disorders [72][73][74]. Anthropometric studies suggest that children with smaller head circumference and cerebral volume may develop Huntington's disease [37] re ecting the developmental origin of the disease. Cognitive capabilities were better protected in Alzheimer's disease patients with relatively larger head circumference [75]. Therefore, nigral volumetry may be a valuable tool to identify the predisposed individuals/populations.
A hypothetical division of the postnatal period in three phases reveals that the second phase that de nes the neuronal numbers is more critical. Such differences may be envisioned in the Caucasian, Asian-Indians, and their admixed population imparting them varying degrees of susceptibility to develop PD [9,12,11,6,15,34,35].
Finally, the ndings of fewer neurons throughout development in C57BL/6J nigra; posit the presence of excessive endogenous toxins or pro-apoptotic factors at these stages. This possibility was strengthened by corresponding high levels of caspase-3 and AIF. The higher cellular packing density at P14 in C57BL/6J may represent increased competition for trophic factors. Whereas, minimal loss of DA cells, higher GNDF and controlled expression of pro-apoptotic factors in CD-1 and the crossbreds, hint at better endogenous milieu. Developmental apoptosis sculpts aberrant neural connections [76,77]to regulate the nal neural numbers in adult multicellular organisms [78,79]. Since both the mice strains are non-transgenic and genetically distinct, we hypothesize that genetic constitution may predispose individuals or populations even in absence of major mutations. In late eighties, Fahn [5] hypothesised that the brain wiring during development may in uence the number of neurons at birth and other cellular features which may be unique to the individuals, making them either vulnerable or resilient to PD. He also proposed that the metabolic processes in the brain may govern vulnerability to develop PD [5]. Both these hypotheses hereby stand validated, hinting at the developmental origin of vulnerability to MPTP. It may therefore be possible that susceptibility to PD originates during nigral development in humans.

Declarations:
Declarations Ethics approval and consent to participate: the study was approved by NIMHANS Institutional Animal Ethics committee (AEC/50/312/N.P.). Since this study presents no human data, the "consent to participate" is not applicable.
Consent for publication: All authors have approved the work described in this manuscript and consented for its submission and publication.
Availability of data and materials: The data will be made available upon reasonable request. We purchased commercially available antibodies hence we do not have any biological material to share.
Competing interests: The authors declare that the research was conducted in the absence of any commercial or nancial relationships with any company that could be construed as a potential con ict of interest.  p<0.01, C57BL/6J & CD-1 v/s F1X1 & F1X2). The soma size was comparable across strains at P10, when they attained adult phenotype i.e. clear nuc Cellular TH expression (f): At P2 the TH expression in CD-1 and F1X2 was higher compared to C57BL/6J and F1X1. Stabilization of TH expression to TH expression per unit soma area (g): TheF1X2 mice showed signi cantly more TH expression per unit soma area at P2 than C57BL/6J (P2,