“Neuropeptides required for Drosophila development under nutritional stress are regulated by the ER-Ca2+ sensor STIM”

Significance Elevation of cytosolic Ca2+ is required for the release of neuropeptides from specialised secretory neurons called neuroendocrine cells. Release of Ca2+ stored in the ER, which further triggers Ca2+ entry from the extracellular milieu is called Store-operated Ca2+ entry (SOCE). STIM, an ER protein, is a key regulator of SOCE and its contribution to neuroendocrine cell functioning is not well studied. Using Drosophila larval development under nutrient restriction as a paradigm, we identified two SOCE regulated neuropeptides, Corazonin and short Neuropeptide F. Reducing STIM alters the level of these neuropeptides in the fed as well as nutrient restricted condition. This study suggests that STIM-triggered SOCE may regulate the release of neuropeptides whose activity ultimately regulates adaptation to nutritional stress. Abstract Cytosolic Ca2+ levels are tightly regulated by the sequestration of Ca2+ within the endoplasmic reticulum (ER). Loss of ER Ca2+ is sensed by STromal Interacting Molecule (STIM), whose translocation to the plasma membrane triggers Store Operated Ca2+ Entry (SOCE), and a subsequent rise in cytosolic Ca2+. Relatively little is known about SOCE’s contribution to neuroendocrine cells; a neuronal sub-type that specializes in the secretion of neuropeptides (NPs) which ultimately regulate animal physiology and behavior. To investigate how SOCE regulates NPs, Drosophila development under nutrient restriction (NR) was used as the biological context. Genetic experiments identified the requirement of two SOCE-regulated NPs-corazonin (Crz) and short neuropeptide F (sNPF) - for the development of NR larvae to pupae and finally, adulthood. Overexpression of SOCE regulators was sufficient to rescue the development of NR larvae with reduced sNPF or Crz levels. To facilitate cellular investigations, a restricted set of neurons that produce sNPF and Crz, and are activated by NR, were identified. Immunohistochemistry on these neurons and mass spectrometric measurements of their projections showed that dSTIM regulates Crz and sNPF levels at steady state and in response to NR, likely by modulating their release. Genetically increasing neuronal output, in the background of reduced dSTIM, robustly rescued development of NR larvae. Because NP action fundamentally underpins how animals adapt to stimuli, regulation of NPs by STIM triggered SOCE is likely to be important. In situations where SOCE is altered, such as neurodegenerative diseases, this regulation may contribute to disease manifestation and progression.


Introduction
Ca 2+ is one of the primary signals utilized by cells for potentiating signal transduction pathways. Depending on cell type and context, elevation of cytosolic Ca 2+ can trigger a diverse array of cellular processes. To keep basal cytosolic Ca 2+ levels low, Ca 2+ is sequestered into various organelles, the largest being the ER. ER-store Ca 2+ is released by ligand activated Ca 2+ channels such as the Ryanodine receptor or inositol 1,4,5trisphosphate receptor (IP 3 R). Loss of ER-Ca 2+ is sensed by STromal Interacting Molecule (STIM), an ER-resident transmembrane protein, which undergoes structural rearrangement to bind Orai, the store-operated Ca 2+ channel on the plasma membrane.
Activation of Orai causes an influx of Ca 2+ from the extracellular milieu which may have up to ~1000 fold higher levels of Ca 2+ as compared to the cytosol. This type of capacitative Ca 2+ entry is termed Store-operated Ca 2+ entry (SOCE) (1). The proteins that regulate SOCE -IP 3 R, STIM and Orai -are ubiquitously expressed in the animal kingdom, underscoring the importance of this pathway to cellular functioning.
Elevation of cytosolic Ca 2+ is central to neuronal function. To achieve context dependent outcomes, neurons possess many different strategies to modulate the frequency and amplitude of Ca 2+ signals (2). Of these, the properties and expression of a large repertoire of activity-dependent voltage gated Ca 2+ channels (VGCCs) and receptoractivated Ca 2+ channels, that bring in external Ca 2+ , is perhaps the most well studied. The contributions of internal ER-Ca 2+ stores to neuronal Ca 2+ dynamics are also well recognized. However, the study of how STIM and subsequently SOCE may influence neuronal functioning is as yet a nascent field. In mice, STIM2 mediates SOCE in cortical (3) and hippocampal neurons (4), while STIM1-mediated SOCE has been reported for cerebellar granule neurons (5) and isolated Purkinje neurons (6). Although both STIM1 and STIM2 are expressed in the hypothalamus (Human Protein Atlas), the major neuroendocrine center in vertebrates, a role for STIM or SOCE is poorly investigated in this tissue.
NE cells are secretory neurons that produce NPs, the most diverse group of neuronal signaling agents. Inside the brain, NPs typically modulate neuronal activity and wiring; when released systemically, they act as hormones. Dmel is typical in having a vast repertoire of NPs that together play a role in almost every aspect of its behavior and physiology (17,18). Consequently, NP synthesis and release are highly regulated processes. As elevation in cytosolic Ca 2+ is required for NP release, a contribution of STIM-mediated SOCE to NE function was hypothesized.
Dmel development under nutritional stress was chosen as the biological context to identify SOCE-regulated NPs, for four reasons. First, we previously reported that reducing IP 3 R or dSTIM or expressing a dominant-negative Orai (Orai E180A ), in a large number of Dmel NE cells, reduced larval development under NR conditions (11). Because these manipulations are also known to reduce SOCE in Dmel neurons (7,8), it was likely that SOCE regulates NPs secreted by NE cells that are involved in the adaptation to NR conditions. Second, adaptation to poor nutritional conditions requires a number of coordinated behavioral and physiological responses, each likely needing a distinct NP or a set of NPs, thus using this assay improved the likelihood of uncovering SOCE-regulated NPs. Third, most NPs in the nutritional and feeding context are studied in adult flies, where regulatory mechanisms involved in adaptation to dietary imbalances and starvation, may be different from the larval stage. Adults require nutrition for reproduction and survival, whereas larvae require them predominantly for growth and development. Thus far, a role only for three of eight Dmel insulin-like peptides (dILPs) which are neuropeptides, has been well-established in relation to nutrition and development (19,20). Several NPs directly or indirectly regulate the dILPs, with consequences for maintaining metabolic homeostasis and regulating feeding behaviors, but of note, a majority of these studies are in adults (21). Therefore, the assay used in this study was expected to yield insights on NPs required during development, an uncommon area of study. Finally, a developmental readout allowed for the exploitation of the Dmel genetic tool kit to scale from cellular perturbations to systemic behavior.

SOCE is required in Crz and sNPF producing neurons for development under nutritional stress
The ability to pupariate on nutrient-restricted (NR) media (100mM sucrose) as opposed to a complex mixture of yeast, sugar and corn flour ("normal" food), when transferred in the last stage of larval development (~88 hours post egg laying), was used as a measure of development. Collectively, more than 20 different NPs are made by the NE cells in which reducing SOCE components resulted in poor pupariation upon NR (11). From these NPs, those with established roles in feeding and metabolism were selected and, a curated GAL4-UAS genetic screen undertaken. Neuropeptide GAL4s were used to drive the knockdown of IP 3 R (IP 3 R IR ), and pupariation of the resulting larvae was scored in normal vs NR food conditions (Fig. S1A). On normal food, a significant reduction of pupariation was seen only with sNPF-GAL4. Upon NR, the largest effect was seen with sNPF-GAL4, followed by small but significant pupariation defect with AstA-GAL4, Crz-GAL4 and DSK-GAL4. Neurons that secrete NPs may also secrete neurotransmitters, therefore, a role specifically for sNPF was tested next. Directly reducing the level of sNPF (sNPF IR ) or reducing an enzyme required for neuropeptide processing (amontillado; amon IR ) in sNPFproducing cells, as well as a hypomorphic sNPF mutation (sNPF 00448 ) resulted in impairment of larval development upon NR (Fig. S1B). These indicate that sNPF is required for pupariating under NR conditions. sNPF-GAL4 has a diverse expression pattern in the larval CNS (22) and also expresses in the larval midgut and epidermis. To identify a restricted set of neurons that contribute to the pupariation phenotype observed on NR media, we focused on a Crz-GAL4 strain, as sNPF and Crz are co-expressed in a set of three bilateral neurons in the larval brain lobes ((22); Fig. S1C). Reducing either sNPF or Crz (Fig. 1A) using the Crz-GAL4 driver resulted in larvae with a pupariation defect on NR food. Furthermore, reduction of Crz using sNPF-GAL4 (Fig. S1D) presented a defect upon NR to the same extent as that was observed with Crz-GAL4 (Fig. 1A). Together, these suggest a role for neurons that express both sNPF and Crz in development under nutritional stress.
As reduction of IP 3 R resulted in a weak phenotype (Fig. S1A), the focus was shifted to dSTIM, as both proteins are known to positively regulate SOCE in Drosophila neurons (7,12). Knockdown of dSTIM (dSTIM IR ) with either sNPF-GAL4 or Crz-GAL4, resulted in reduced pupariation under NR to the same extent, with no effect on development on normal food (Fig. 1B). Over-expression of a dominant-negative version of Orai (Orai E180A ), known to reduce SOCE in neurons (8), also reduced purariation rates in NR larvae when expressed in the Crz + neurons ( Fig S1E). Thus, SOCE is required in Crz + neurons for development under nutritional stress.
Loss of either IP 3 R or sNPF has previously been shown to affect larval feeding (23,24). Hence, to eliminate the possibility that developmental defects on sucrose arise from compromised feeding, the intake of dye-colored food was measured. Age-synchronized larvae with knockdown of either dSTIM, IP 3 R, Crz or sNPF in Crz-GAL4 expressing neurons exhibited no difference in feeding over a 2-hour period (Fig. S1G), suggesting that developmental defects in the NR assay are not arising from a fundamental feeding problem.
Reduction in either NPs -sNPF or Crz -or SOCE regulators -IP 3 R and dSTIM -in Crz + neurons (Fig. 1A,B, S1A) resulted in reduced pupariation upon NR. To understand if these NPs and SOCE regulators interact genetically, compensation experiments were undertaken. In the background of reduced NPs, SOCE-regulators were over-expressed in Crz + neurons. NR larvae with this genetic make-up not only showed significant improvement in pupariation (Fig. 1C), but also in their ability to eclose as adults (Fig. 1D).

Enhancing the expression of SOCE regulators leads to increased SOCE in Drosophila
neurons (12), suggesting that reduced neuronal output of NPs can be overcome by enhancing intracellular Ca 2+ signaling.

Activation of Crz + and sNPF + DLP neurons is required to pupariate on NR diet
In the larval CNS, Crz is expressed in 3 pairs of DLPs (Dorso Lateral Peptidergic neurons) in the pars lateralis region of the brain lobes, 1 pair of neurons in dorso medial region and 8 pairs of interneurons in the VG (ventral ganglion) (25). Except the medial neurons, the Crz-GAL4 used in this study recapitulates the known expression pattern of Crz + neurons ( Fig. S2A and C; Cartoon: Fig. 2A). Low ectopic expression of Crz-GAL4 was observed in 3-4 neurons adjacent to DLPs (Fig. S2B). Crz + DLPs also express sNPF (22). The Crz + and sNPF + DLPs have two major branches: the anterior branch culminates in a dense nest of projections at the ring gland (RG) which is considered the neurohaemal site for systemic release, while the posterior branch terminates in the gnathal ganglion (GNG). The VG neurons form a network amongst themselves with two parallel branches emerging that terminate in the brain lobes, close to the anterior branch of the DLPs, suggesting possible neuromodulation by the VG neurons. Atrial neuropeptide (ANF), a rat neuropeptide coupled to GFP (26) was used as a proxy to visualize the distribution of neuropeptides in the Crz + neurons ( Fig S2D). Firstly, ANF intensity was higher in the cell bodies of the DLPs than VG neurons. Secondly, within the DLPs, ANF::GPF was either in the cell bodies or RG projections, but not in the projections terminating at the GNG, suggesting that the RG is the major site of neuropeptide release for the DLPs (Fig. S2D).
Functional requirement of Crz + neurons to survive NR was tested genetically in different ways: 1) by ablation of the neurons (using UAS-hid::UAS-rpr); 2) by reducing vesicle recycling using temperature-sensitive dynamin mutant (shibre ts ); 3) by impairing vesicle release using tetanus toxin (TeTxLc). All three perturbations resulted in reduced pupariation on NR media (Fig. 2C).
To probe the activation state of Crz + neurons, the UV light-activated genetically encoded calcium sensor, CaMPARI (27), was utilised. The sensor fluoresces in the GFP range (F green ) and is converted irreversibly to fluoresce in the RFP range (F red ), when exposed to UV light and in the presence of Ca 2+ . The level of conversion positively titrates with Ca 2+ concentrations. Larva expressing CaMPaRI in Crz + neurons were placed in either normal or NR food for 24 hours (NR). Whole larvae were then exposed to UV light for 2mins. Control larva (Fig. 2E,F; no UV) were subject to the same treatment but without being exposed to UV light. Increased F red /F green ratios were observed in the DLPs after 24 hours on NR media, suggesting increased cytosolic Ca 2+ levels and therefore, enhanced neuronal activity (Fig. 2D,E). F red /F green ratios did not appear to change in the VG neurons suggesting either that either they do not participate in nutritional stress response (Fig. 2F,   S2E) or there is insufficient penetration of UV light into the VG. Therefore, expression of dSTIM was restricted to the DLPs by using tshGAL80 (28) that prevented expression of Crz-GAL4 in the VG (Fig. 2G). The level of pupariation under NR conditions observed with restricted expression was similar to that seen with full expression (Fig. 1B). Similar levels of reduced pupariation of NR larvae is also seen upon DLP restricted knockdown of Crz by sNPF-GAL4 (Fig. S1D). Together, these results suggest that a major role in the nutrient stress response is played by the Crz + and sNPF + DLP neurons.

dSTIM affects Crz and sNPF peptide levels
The identification of a restricted set of neurons, along with specific neuropeptides, provided a system to interrogate how dSTIM might regulate NPs at the cellular level. As both neuronal activation and dSTIM-regulated SOCE can regulate transcription (30,31), relative levels of Crz mRNA in larval brains were measured by qRT-PCR. Larvae (~90hAEL) were fed either normal or NR food for 24hrs followed by measurement.
Interestingly, elevated levels of Crz mRNA were found in the dSTIM knockdown condition as compared to controls in both normal and NR conditions (Fig. 3A). Notably, NR did not alter transcript levels of Crz in control brains (Fig. 3A). Systemic feedback from reduced Crz signaling has been reported to increase Crz transcript levels (32). To understand if the increased Crz mRNA levels in fed state of dSTIM knockdown might be explained by such feedback, we measured levels of Corazonin receptor (CrzR), which in larvae are predominantly expressed in the CNS (33). Reducing Crz results in increased levels of CrzR in larval brains on normal food, and a similar increase is seen when dSTIM is reduced in Crz + neurons (Fig. S3A). This indicates that loss of dSTIM in Crz + neurons results in reduced systemic Crz signaling; with likely feedback resulting in increased Crz mRNA levels.
Under conditions similar to that used for measuring transcript changes, we next Thus, increased activation of these neurons by NR (Fig. 2E), appears to result in Crz peptide mobilization. In comparison, Crz + DLPs with reduced dSTIM displayed increased Crz peptide levels on normal food itself (Fig. 3B,C, S3C,D). Also, unlike in the control, NR did not appear to significantly alter peptide levels in Crz + DLPs with reduced dSTIM (Fig.   3B,C, S3 C,D; See dSTIM KD Normal vs NR).
Because sNPF is expressed broadly in several neurons of the larval CNS (Fig.   S1C), and the phenotype under observation derives from a small set of neurons, the strategy used to assess Crz levels could not be used for sNPF. Instead, semi-quantitative direct mass spectrometric profiling of dissected RGs was employed. This technique can measure peptide levels relative to stable isotopic standards at single neurohemal release sites (34). As Crz peptide levels measured by IHC were consistent between the cell bodies and projections, and Crz + DLPs are major contributors to sNPF on the RG, this technique allowed us to infer sNPF peptide levels from a restricted set of neurons. RGs were collected from larvae with the same background and conditions as used for RT-PCR and IHC experiments. In the control, although starvation increased the average level of sNPF peptide in the RGs, the increase was not statistically significant; perhaps a resolution limit of the technique (Fig. 3D). However, when dSTIM was reduced in the Crz + DLPs, RG preparations showed a sharp increase in sNPF levels on normal food and remained unaltered with NR (Fig. 3D). While Crz peptide was detectable in the RG preparations, it was of much lower intensity; thus, despite increases in average levels of Crz peptide as seen for sNPF, statistically higher levels of Crz were observed only under NR, dSTIM KD condition (Fig. S3E).
Cumulatively, these results show that reducing dSTIM affects both sNPF and Crz peptide levels similarly, suggesting that dSTIM alters function of specific NE cells, at a fundamental level. NR activates Crz + and sNPF + DLPs (Fig. 2E), causing a mobilization of Crz and sNPF peptide (Fig. 3C,D, S3C,D). This mobilization is likely required for the sustained release of Crz and sNPF from the neurohemal endings of DLPs on the RG, and promotes development in NR larvae. Loss of dSTIM in Crz + DLPs increases peptide levels and this increase remains unchanged even under starvation. One interpretation of greater retention is impaired release, which is consistent with the reduced systemic Crz signaling seen upon loss of dSTIM (Fig. S3A). Relative total sNPF peptide levels measured on dissected ring glands (N atop bars) and quantified using MALDI-MS. Externally added heavy standard (Hug-PK*) was used to normalise peptide levels between samples. Data represents mean ± SEM. Two-way ANOVA with Sidak multi comparison test p<0.05 for (A). Kruskal-Wallis Test with Dunn's multicomparison correction p<0.05 for (C) and (D). Bars with the same alphabet represent statistically indistinguishable groups.

Development of NR larvae with reduced dSTIM in Crz + neurons can be rescued by increasing neuronal output
Cellular studies suggested that reducing dSTIM decreases the release of Crz and sNPF in normal and NR food conditions. Enhancing neuronal activation leads to increased vesicle exocytosis and therefore serves as a means to increase NP release. This formed the basis of employing genetically encoded, externally triggered channels to enhance neuronal activity of Crz + neurons. Two types of channels were over-expressed in Crz + neurons: a) TrpA1, a temperature sensitive Ca 2+ channel and b) Channelrhodopsin (ChR2-XXL), a light-sensitive cation channel. In the dSTIM IR background, over-expression of dTrpA1 and its activation by raising the temperature to 30 o C acutely for 24 hours at the time point of transfer to NR media, resulted in a rescue of the pupariation as well adult eclosion defect of NR larvae (Fig. 4A). Over-expression of ChR2-XXL and exposure to white light post-transfer to NR media, resulted in modest rescue of pupariation of dSTIM IR NR larvae (Fig. S4A). Poorer rescue with ChR-XXL could be because sustained activation of this channel depress synaptic transmission (35). Similar genetic manipulation but in a hypomorphic IP 3 R mutant (itpr ku ) resulted in a small but significant rescue of pupariation on NR media (Fig. S4B,C). Together, these results strongly suggested that defects in intracellular Ca 2+ signaling molecules may be overcome by increasing neuronal activity, which likely increases peptide release. Another strategy to increase NP release relied on the over-expression of the small GTPase Ral (crz-GAL4>Ral WT ) which functions as part of the exocyst complex and positively regulates NP release from dense core vesicles (15). NR Larvae with reduced dSTIM but over-expressing Ral displayed higher pupariation than controls (Fig. S4D).
Besides neuronal activation, neuronal output of NPs may be enhanced by increasing NP production. The simplest way to accomplish this is by NP overexpression via the GAL4-UAS system. However, in the peculiar case of NPs, for enhanced biosynthesis of active peptide to translate to enhanced release, proteins involved in NP processing as well as regulated secretory system would need to be up-regulated.
Furthermore, regulatory feedback from peptides to their transcription may complicate overexpression. To get around these issues, an indirect strategy to enhance overall protein synthesis was adopted. Over-expression of Insulin receptor (InR) increases the cell size and peptidergic capacity of DIMM + NE cells with functional consequences (36). As the Crz + DLPs are DIMM + (37), we reasoned that Crz + neurons may be sensitive to InR levels.
Supporting this was the observation that reducing InR (InR IR ) significantly decreased pupariation of NR larvae (Fig. S4E). Also, over-expressing InR not only increased cell size (Fig. S4F), but also increased Crz peptide levels ( Fig. 4B,C). Finally, in the dSTIM knockdown background, over-expression of InR resulted in increased pupariation as well as adult eclosion of NR larvae (Fig. 4D).
Thus, increasing either neuronal activity or protein synthesis, could override the loss of dSTIM in Crz + neurons and result in a functional rescue of development of NR larvae.
Such different functional outcomes are likely regulated by controlling NP release amount and timing, from fast quantile release to slow secretion (48). Because Ca 2+ is required for NP vesicle exocytosis, various combinations of Ca 2+ influx mechanisms can be used to fine tune the amount and timing of Ca 2+ elevation and therefore, NP release. For example, in Dmel neuromuscular junction, octopamine elicits NP release by a combination of cAMP signaling and ER-store Ca 2+ , and the release is independent of activity-dependent Ca 2+ influx (49). In the mammalian dorsal root ganglion, VGCC activation causes a fast and complete release of NP vesicles, while activation of TRPV1 causes a pulsed and prolonged release (50). In primary neuronal cultures from Dmel, loss of dSTIM not only reduces SOCE (7) but also KCl-stimulated vesicle exocytosis (15). Thus, it is conceivable that elevation of cytosolic Ca 2+ resulting from dSTIM mediated SOCE is a strategy used by NE cells to fine tune peptide release in vivo across phyla.
Because Ca 2+ levels monitored by CaMPARI (Fig. 2 D,E) show that Crz + DLPs have basal activity in the fed state too, together, these results suggest that dSTIM regulates Ca 2+ dynamics and therefore, NE cell activity, under basal as well as stimulated conditions. This is consistent with observations that basal SOCE contributes to spinogenesis, ER-Ca 2+ dynamics as well as transcription (51). However, in our case, this regulation appears to have functional significance only in NR conditions as pupariation of larvae with reduced levels of dSTIM in Crz + neurons is not affected on normal food (Fig 1B).
In the context of nutritional stress, this study provides evidence for a hypothesis first proposed by Jan Veenstra that Crz + neurons play a central role in conveying nutritional status information (52). Crz + DLPs are known to play a role in sugar sensing (53,54) and express the fructose receptor Gr43a (53). Additionally, they express receptors for neuropeptides DH31 (55), DH44 (55) and AstA (52)  In a broader context, STIM is a critical regulator of cellular Ca 2+ homeostasis as well as SOCE, and a role for it in the hypothalamus has been poorly explored. Because STIM is highly conserved across the metazoan phyla, our study predicts a role for STIM and STIM-mediated SOCE in peptidergic neurons of the hypothalamus. Certain NPs also exhibit functional conservation between flies and humans (dILPs/insulin; AKH/ glucagon).
The intriguing similarity between Crz/GnRH (gonadotrophin-releasing hormone) and sNPF/PrRP (Prolactin-releasing peptide), at the structural (56) and developmental (57) level, along with the structural conservation of their receptors, leads to the speculation that GnRH and PrRP might play a role in mammalian development during nutrient restriction.
Of note, like sNPF, PrRP has roles in stress response and appetite regulation (58). Most often, studies on mammalian NPs are performed in the postnatal phase. We hope this study provides an impetus to look at NPs at an early mammalian developmental time point, like the fetus.
Finally, there is growing evidence that SOCE is dysregulated in neurodegenerative diseases (59). In neurons derived from mouse models of familial Alzheimer's' disease (60) and early onset Parkinson's (51), reduced SOCE has been reported. How genetic mutations responsible for these diseases manifest in NE cells is unclear. If they were to also reduce SOCE in peptidergic neurons, it could be speculated that physiological and behavioural symptoms associated with these diseases, may in part stem from compromised SOCE in NE cells.

Fly Husbandry
Flies were grown at 25 o C in 12h:12h L:D cycle. Recipe for normal food (11). For nutritional stress assay, flies were allowed to lay eggs for 6 hours on normal food. After 88hours, larvae were collected and transferred to either normal or NR (100mM Sucrose) food.
Pupae and adults were scored after 10 days of observation. Fly strains used in this study are tabulated in supplementary information.

Measuring neuronal activation using CaMPARI
Early third instar larvae were transferred to either normal or NR food. After 24 hours, larvae were recovered and immobilized on double sided tape. UV light from a Hg-arc lamp was focused on the larvae through a 10X objective on Olympus BX60, for 2 minutes.
Larvae were them immediately dissected in ice-cold PBS, mounted in PBS and imaged using Olympus FV-3000 Confocal microscope using a 40X objective and high-sensitivity detectors. Each experiment always had a no UV control, in which larvae were subject to immobilisation but not UV light. Fluorescence intensity was calculated for each cell body using Image J.