Overview

After 3 months of lab work, our experiment data has allowed us to reach the following conclusions:

• Cytoplasmic RFP is a successful transfection reporter in iPSC-derived NPC, however it does not allow cell counting. Our nuclei-bound mCherry reporter successfully allows selection of transformed cells.
• tBDNF and mBDNF cells present higher speeds in differentiation rates in comparison to control cells.
• tBDNF cells present higher cell count compared to control cells without generating uncontrolled growth.
• BDNF expressing cells generate more projections than mBDNF and WT cells.
• BDNF expressing cell cultures contain more active neurons than the control.
• BDNF expression does not affect the formation of neuronal networks.
• BDNF expression levels are 7 times the basal expression of actin.
• Promoter ligation was not achieved, however we have identified the origin of the error and will work on it in the following weeks.

Transfection reporter

---

Our lentivirus had a cytoplasmic transfection reporter (RFP). Using staining 1 we amplified RFP’s signal by using an RFP-antibody bound to Cy3. We successfully observed RFP fluorescence in transfected cells and none in WT cells.

As we can observe, the only red found in the WT samples are noise particles, validating that our cells were successfully transfected and that our reporter works. Moreover, there is no cross-talk between the red fluorescent channel and the green fluorescent channel, which makes RFP (mCherry) a viable fluorescent reporter for our GCamP-GFP expressing cell line.

Nuclei counting

When comparing the use of the RFP-cytoplasmic reporter with the nuclei reporter DAPI, we confirmed our assumptions that this transfection reporter would not allow us to accurately count the transfected cells.

The cytoplasmic expression of RFP is distorted across all the cells in the culture and doesn’t have defined edges that allow us to count the cells for proliferation assays or obtain transfection efficiency rates.

Seeing these results, it is clear that we need a different transfection reporter method for our cell line and future experiments.

Transfection and proof

We have transfected HEK293T cells with our H2B-mCherry construct, and have successfully confirmed nucleus-reported fluorescence.

The localised fluorescence in the cell nuclei allows us to easily identify the transfected cells, meaning that out construct is a great tool to select only transformed cells.

By overlaying the red fluorescence layer with the 2D images of the cell culture, we are able to visualise how fluorescence falls within each cell, demonstrating that the reporter is in the nuclei. We can also see a difference between untransfected cells and transfected cells, as the fluorescence isn’t found in all cells. This was to be expected, as lipofectamine transfection doesn’t have a very high efficiency. This makes our reporter a great choice to select transformed cells from WT.

Effect of BDNF on NPC

---

Assessing the effect of BDNF on our NPC cultures is essential to Reneurish.

We have performed a variety of experiments to test the effect of BDNF-expression on different neural characteristics:

• Differentiation: Immunohistochemical assays and culture observation.
• Projection formation: Chip-cultures.
• Synaptic activity: GECI calcium recordings.

Differentiation

---

BDNF has been shown to promote differentiation and maturation of neural progenitor cells [1]. This is why we wanted to see how it affected our NPC cultures.

The experiment

To test and determine the differentiation and maturation levels in our cell cultures, we took progress pictures via simple microscopy, as well as performed immunohistochemical assays.

These assays would use antibodies to mark different proteins and transporters characteristics of neurons and astrocytes, that would also help determine the cells’ maturity.

In essence, we have 4 different types of markers according to what information they provide of the cultures:

• Transfection reporter: RFP
• Synaptic markers: PSD95, SynI
• Neuronal markers: MAP2, NeuN, TujI
• Astrocyte marker: GFAP

For more information on the staining composition:

Staining	Primary	Secondary	Color
Staining 1	anti-RFP (Mouse)	Anti-Mouse	CY3 (red)
	anti-PSD95 (Rabbit)	Anti-Rabbit	CY5 (magenta)
	anti-MAP2 (Chicken)	Anti-Chicken	488 (green)
	DNA	DAPI (blue)
Staining 3	anti-SynI (Mouse)	Anti-Mouse	CY3 (red)
	Anti-NeuN (Rabbit)	Anti-Rabbit	CY5 (magenta)
	Anti-GFP (Goat)	Anti-Goat	488 (green)
	DNA	DAPI (blue)
Staining 4	Anti-PSD95 (Mouse)	Anti-Mouse	CY5 (magenta)
	Anti-MAP2 (Chicken)	Anti-Chicken	488 (green)
	Anti-GFAP		CY3 (red)
	DNA	DAPI (blue)

For further information regarding the functions of the proteins, consult this section.

Culture Evolution

---

Routine microscopy observations of our cultures allowed us to observe physiological changes of our neurons and oscillations in cell confluency.

Pre-transfection

The first images were taken on May 29th, right before lentiviral transfection and 13 days after thawing the It-NES cells.

At this stage, cells were still in a very early stage of development but already started presenting neural-like characteristics, such as the association of cell bodies in rosettes. We can also observe the cells starting to generate projections, but it is still too early to tell whether they will become axons or dendrites.

Post-transfection

The day after transfection (May 30th), no significant changes in cell confluence or survival were observed.

This means that the cell transfection has not affected cell viability.

Day 16

On day 16 post-differentiation, we observed the progress of our cultures. The confluence of the cells is much greater, and the clusters formed by the neural bodies have become much larger.

We can also observe how the size of the neural projections, such as the axons, have extended significantly, some even seeming to connect different neural rosettes.

Immunohistochemical Assays

---

Cell Quantification

Cell quantification allows us to see whether BDNF promotes neural survival or affects cell count. This also helps us observe whether constitutive expression of BDNF generates uncontrolled growth and presents risk of teratomas.

To quantify the number of cells per image, we isolated the DAPI (blue) layer marking the cell nuclei, and used imageJ for automatic cell counting.

This process required all images to be inverted and converted to 8-bit. We then increased contrast by adjusting the threshold to reduce system error. We automatised the process by using the Image-based Tool for Counting Nuclei (ICTN) plugin. We used the same settings for all counting, using 6 px as the set diameter and a minimum distance of 8 px.

Day 16: Control and tBDNF

By day 16, the amount of cells is greater in tBDNF cultures than control cultures, and the size and depth of the neural aggregates is also much greater in tBDNF cultures. This is a positive sign, as bigger neural aggregates are a sign of higher differentiation.

On day 16 post-differentiation we fixed a control ibidi plate and a tBDNF ibidi plate.

The biggest noticeable difference between these two cultures is the size of the neuronal aggregates and the width of the different cultures. The confocal microscope we used to take the images analyses the cultures and takes pictures at different Z levels (different depth planes), and whilst the control culture needed 3 planes, the tBDNF culture needed 7.

We estimate that the cell count of these images might present a significant error due to the images containing large neuronal clusters. This most likely affects the tBDNF culture most, as the depth of the culture is much greater.

Nonetheless, the program estimated that the control plate contained 6398 cells, whilst the tBDNF plate contains a total of 9236 cells . In this case, the tBDNF culture has 30% more cells than its WT counterpart. However, due to the size of the neural aggregates, the estimation is not entirely reliable.

For further count imaging, we will try to select areas of the plates with the least possible amount of neural aggregates to obtain more reliable data.

Day 52: Control, mBDNF, and tBDNF

Day 52 plates contain much more mature cells and stable cultures. In this case, cells from both stain 3 and stain 4 ibidi cells were counted using the automated method mentioned previously. We observed how, on average, tBDNF cultures contain 80% more cells than their control counterparts.

Control cell count is within the 1010-960 interval, averaging 989 cell nuclei per ibidi plate. In comparison, tBDNF count belongs in the 1640-1860 interval, and averages 1781 . This means that tBDNF cells have around 800 more cells per plate than the control cells.

However, the difference in cell count is not great enough to suggest highly proliferative cells nor uncontrolled growth, meaning that despite an increase in cell number, the risk of teratoma formation is very low.

Differentiation

Stain 1 was used to visualise maturity during the early stages of differentiation (day 16), whilst stains 3 and 4 allowed us to tell maturity at later stages (day 52)

Early stages:stain 1

There is no expression of PSD95 (magenta) in neither control nor tBDNF cultures, meaning the cells are still in an early stage of development. No significant differences are observed between control and tBDNF cells.

Late stages: stain 3

We calculated the proportion of NeuN expressed in the cells by normalising the cell count by the number of cell nuclei detected in the cell, which we’d obtained previously from our cell quantification.

The results show that the tBDNF culture doubles the control culture in proportion of NeuN expressing cells. Whilst 3,86% of control cells express NeuN, 7,86% of tBDNF cells are NeuN expressing. This means that NeuN is found two-fold in tBDNF cultures in comparison to the control.

NeuN, being often used as a late differentiation marker, confirms that BDNF-expressing cells are proportionally more mature in comparison to the negative control cells.

Late stages: stain 4

Differences in PSD95 (Pink) allow us to confirm that BDNF favours cell differentiation. As we can observe, PSD95 staining in BDNF cells (mBDNF, tBDNF) highlights the axons, demonstrating presence of the post-synaptic channel. In contrast, control cells’ axons are not stained with Cy5, and we mostly observe the stain in the cell body.

The increase of PSD95 in cells under exposure to BDNF helps us cement that it speeds up neural differentiation and maturation.

This is the only stain we have marked astrocytes with GFAP. This was interesting as our NPC are supposed to differentiate into all neural cell types, not only neurons. In these images, and clearly standing out in red, we were able to verify that our iPSC-derived NPC are able to produce more than one neural cell type.

Projection Formation

---

The generation of projections by the different cell types is analysed via the chip culture plates assay. By quantifying cells on each side of the channels and counting the amount of axons entering and exiting them, we determined that BDNF exposure helps improve the formation and extension of projections, in comparison to WT cells.
Stain

These chips were all stained with the staining 2 solution. We decided to implement the staining two solution to better visualise the neurons, as it contains markers for both mature (MAP2) and immature (Tuj1) neuron markers stained in green (488). This gives a more universal view of the neurons present in the culture, and ensures visualisation of all neurons.

Staining 2
anti-NeuN	Rabbit	Anti-Rabbit	Cy5 (magenta)
	anti-MAP2	Mouse	Anti-Mouse	488 (green)
	anti-tubulin	Mouse
	DNA			DAPI (blue)

Axon count

Axon counting was performed on images with the entire plane of the chip plates, with a total of 120 channels per plate.

Whilst tBDNF and mBDNF cells emit at least 1 axon per channel, control cells only emit 2 axons per every 5 channels . This confirms that BDNF exposure to cells, whether in media or by cell expression, increases the amount of axons generated by neurons.

Due to the presence of plastic from the plate separating the channels, counting of the axons that cross in the middle of the channels is not recommended. Primarily as interference from the fluorescence and light dispersion is not picked up by the microscope, but also because due to the small size of the channels the antibody solution struggles to flow through and stain the axons. This is why we counted the axons that were entering the channels from either sides of the cell compartments instead of the axons crossing. This is also a more reliable method, as in the channels axons get compacted and can’t be counted adequately.

It is important to note, that despite the control chip having nuclei close to the axon channels, this has not translated to the presence of MAP2 and Tuj (green stain). As observed in the image, very few green projections are found close to the channel entrances, which may have affected the low results of the control chip.

In contrast, mBDNF had a very high density of axons close to the channels, generating the opposite problem. The high density of axons and their association might have resulted in overestimating the amount of axons that enter the channels. This specially affected the entering axons on the left side of the chip.

	Left	Right	Total
WT	28	48	76
tBDNF	55	140	195
mBDNF	148	96	244

The “right” value for each respective chip plate corresponds to WT cells for the control chip, tBDNF cells for the tBDNF chip, and mBDNF cells for the mBDNF chip. In contrast, the “left” value always corresponds to WT cells.

We were pleasantly surprised by the increase in emitted projections by the BDNF-expressing neurons in comparison to the positive control (mBDNF). Not only that, but the tBDNF cells have generated three times the projections than their WT counterpart.

Taking into account how the mBDNF cells have also emitted more projections than the WT cells, we can confirm that BDNF exposure positively affects the emission of projections by neurons, and that constitutive expression by said neurons is preferred over media exposure to BDNF.

This means that our transfected neurons are better suited for integration into brain tissue than neurons treated with BDNF.

Synaptic Activity

---

To study our cells’ synaptic activity, we performed multiple recordings of GFP expression bound to GCamP, which allowed us to visualise changes in action potential, and therefore when and where synapses were taking place.

We used the program HoKaWo from Hamamatusu to perform the recordings, and analysed the data with Netcal and custom MatLab scripts.

ROI Selection and individual Calcium spikes

In order to collect and interpret the data from the culture recordings, we used Netcal program to identify active cells (ROI), by detecting changes in brightness intensity.

This program generates traces of individual fluorescent intensity of each ROI, and visualises them into a spike-like map. Each spike represents an action potential of the cell, detected by an increase and decrease in fluorescence.

The spike profiles of the different neuronal cultures present normal patterns, meaning that no frequency aberrations are observed in presence of BDNF.

Active cells

As shown in the diagram, the number of active cells in tBDNF (purple) cultures doubles the amount of active control (orange) neurons. This pattern is not very significant at early stages of differentiation, but increases with neuron maturation. .

Community synapses

The analysis of community synapses takes into account the global interaction of neurons with each other in the establishment of networks. We have observed how BDNF exposure does not negatively affect the formation of the neuronal network nor does it generate aberrations, meaning that the functionality of the networks stays the same as the control.
Due to size differences between tBDNF and control communities, the following data only takes into account synaptic activity, but not cell amount. For these analysis, we had to downsize the tBDNF sample so that the number of active tBDNF cells matched the number of active control cells in order to avoid introducing any bias to the data. In our particular case, we had to downsize the sample size to 75 cells for both control and tBDNF cultures.

Analysis - Network connectivity maps

What we have observed is that the differences in synaptic networks are negligible. tBDNF cells are establishing the same networks as their control counterparts. The size of the communities is variable, but it is not a significant parametre.
The network connectivity maps use the information derived from the adjacency connectivity matrices by connecting all the cells that have established a connection in a 3D space. This allows us to “see” the shape of the synaptic network to assess whether there are any significant changes in their composition. The different colours of the graph represent the neurons belonging to different communities.

These results allow us to confirm how expression of BDNF in neurons does not change the composition of synaptic networks, meaning that BDNF does not have a negative effect in their constitution.

Raster Plot

From the spike maps obtained from each individual ROI, the program extracts a raster plot, which represents every single active cell within the recording, were the blue dots represent when they experience an active potential, namely, when they synapse.

Each horizontal line represents a cell, and the blue dots that make it up represent the timepoint during the recording at which they synapse. Blue dots belonging to the same vertical line are synapses that occur at the same time, and are a sign of cell communication, specially if the behaviour is presented repeatedly.

Adjacency connectivity matrices


Adjacency connectivity matrices are derived from the raster plots. The raster plot matrix vertical axis, which is used to identify individual neurons, is positioned in both the vertical and horizontal axis. In this case, the black dots in the matrix represent an effective connection established between two different neurons (or, in other words, a simultaneous synapse). This analysis allows us to define neuron communities, which are shaded in different colours. This has allowed us to observe that whilst control cultures have larger and fewer cultures, tBDNF cultures have generated more neuron communities but of smaller size.

Other network parameters


These parameters were calculated using MATLAB’s Brain Connectivity Toolbox [2]

GEFF_TH Global efficiency accounts for the capacity of neurons to exchange information across the entire network.

When Gₑ ≃ 0, the ROI poorly communicate with each other, whilst Gₑ ≃ 1 indicates higher capacity of information exchange throughout the global network.

Leff_TH Local efficiency is a similar measurement to Global efficiency, where the neuronal connection is judged by singular nodes instead of the community.

Q_TH Modularity index account the tendency of neurons to form functional modules.

In essence, it evaluates wether the neurons of a community have higher tendency to generate connections within their own community, or whether they have a greater tendency to reach out to other communities.



High Q values indicate low connectivity between different neuron communities, whilst low Q values indicate high connectivity within the same community.

Comm_Size_5more Community size calculates the relative size of the different neuronal communities.

SmallWorldness_SW Small Worldness is a parameter used to compare internal organisation of the neuronal communities.

Ave_L Ave_L helps quantify the number of connexions established between neurons.

Overall, averages of the aforementioned parameters show no significant differences in magnitude. This means that BDNF-expressing cells in culture do not alter the generation and interactions of the neuronal networks nor the neuronal communities.

Conclusions

---

The different assays we have performed have allowed us to determine that BDNF expression in iPSC-derived NPC has great potential for neural transplants.

Not only do BDNF expressing NPC differentiate at a slightly faster rate than WT NPC, but they also generate cultures with higher neuron counts, without generating excessive growth (teratomas).

We have also confirmed that BDNF expressing cells are able to emit a high number of projections, indicating that these cells can easily integrate into brain tissue, facilitating the incorporation of the transplanted cells into the patients’ brains.

Not only that, but BDNF has a positive effect in neurons by increasing the neuronal activity without affecting the composition of the neuronal network nor generating aberrations.


Constitutive BDNF expression

In order to determine the magnitude of BDNF expression under constitutive promoter EF1a, we extracted cell culture media from control and BDF-expressing cells at multiple points during the experiment to perform a qPCR assay. qPCR data

qPCR analysis Reneurish
Conditions	Average Experimental Ct Value (BDNF)	Average Experimental Ct Value (GAPDH)	Average Experimental Ct Value (Actin)
Conditions	Testing Gene (BDNF)	Housekeeping Gene (GAPDH)	Housekeeping Gene (Actin)
WT	26,03444917	23,90796934	21,52323232
BDNF	30,46665792	31,29373045	31,58479845
deltaCt Value
(BDNF-GAPDH)	(WT-GAPDH)	(BDNF-Actin)	Value (WT-Actin)
-0,827072525	2,126523829	-1,118140528	4,51116995
	ΔΔCt GAPDH	ΔΔCt Actin
	-2,953596354	-5,629310478
	2^(-ΔΔCt)	2^(-ΔΔCt)
	7,746777769	49,49841686

qPCR analysis

Using the actin housekeeping gene as reference, we can determine that BDNF is expressed sevenfold in comparison to the control (WT) cells. This means that the constitutive expression of BDNF under EF1a promoter is 7 times the basal expression of actin, whilst the wild-type presents no BDNF expression.

Plasmid Assembly

---

We want to test BDNF expression under three different promoters. We ordered one construct that contains EF1a constitutive promoter, so our plan was to digest the construct to remove EF1a, and then ligate the different promoters into the backbone.

We are currently working on the ligation step as we have obtained inconclusive results.

Ligation verification
To verify that the ligation had occurred as expected, we used Sanger sequencing to sequence the promoter region of the plasmid, where the different gene fragments (promoters) should have inserted. Unfortunately, the results obtained from the sequencing service were ambiguous and hindered the possibility to draw definitive conclusions.

Instead, to verify ligation we designed a new digestion protocol of the plasmid using new enzymes followed by a gel electrophoresis. In essence, in presence of the desired promoter (successful ligation), we would obtain two bands in the gel. If the desired promoter wasn’t in the plasmid, we would obtain 1 single band of the uncut plasmid.

In all cases the first restriction site is found in the plasmid backbone and the second one is found inside the promoters pA, pB and pC but not in EF1a. Thus, if the promoter of interest has been ligated to the vector two bands should appear in the gel. Otherwise, only one band will appear. In the following table we can see the sizes of the expected bands for each situation:

Digestion	Sizes of Expected Fragments (if Successful Ligation)	Sizes of Expected Fragments (if Original Plasmid)
Digestion to check for pA ligation	6150 bp + 850 bp	8000 bp
Digestion to check the pB ligation	700 bp + 6800 bp	1000 bp + 7000 bp
Digestion to check the pC ligation	6250 bp + 950 bp	8000 bp

Digestion results
After the digestion, we ran 1% agarose gels and loaded 2 replicas of each promoter ligation. Plasmids are named pVB432-(+inserted promoter + replica nº + “ ‘ ” if their ligation is still not validated). For each of the plasmid replicas, we loaded a digested sample (“+ enz”) and a not digested replica as control.

For the expected inserted promoters A, B and C we always see the same pattern:

• Digested sample: a single band between 7 kb and 10 kb.
• Control sample: a single band bigger than 10 kb.

This pattern is not the expected 2 bands always smaller than 7 Kb for promoters A, B and C so the expected ligation has not occurred.

Moreover, the digested band size is around 8 Kb, which matches the size of the opened pVB432-EF1a vector. The DNA in the control band appears to be bigger than 8 kb, a plasmid size that is bigger than any of pVB432-EF1a, pVB432-A, pVB432-B or pVB432-C vectors. The explanation for this result is that circular plasmids migrate differently than linearised plasmids and depending on several factors, this migration is slower than the linearised form of the plasmid. [3]

Taking into consideration that we decided not to purify only the DNA fragments of interest, it makes sense that the results indicate that our ligated vectors contain mainly EF1a and not our promoter sequences.

Lastly, before designing a new approach for our plasmid assembly we tried to contemplate other possible limiting factors for the ligation. Initially only a 1% agarose gel had been run to check the ligation (see figure 3). This step was enough to validate the pVB432-EF1a was effectively digested but there was not enough resolution to conclude the same for the promoter's digestion (labelled as A, B and C in the Figures 3 and 4). Therefore, we ran a 4% agarose electrophoresis gel with the digested and control samples of the promoters. There is little resolution for promoter B, as it is the biggest of all (close to 800 bp) and the two digested fragments are about 20bp on each side of the promoter, so a 40bp difference is barely noticeable. Regarding A and B promoters, the digested vs control size difference can be clearly noticed.

We conclude that the limiting factor of our assembly lies in the preparation of the ligation reaction: while the digested DNA fragments of interest are successfully obtained during the double digestion, the ligation reaction contains an excess of unwanted fragments.

Aiming to overcome the latter issue our next steps in the lab are the following:

• Repeat the double digestion step ensuring the conditions to obtain a complete digestion of the DNA.
• After digestion, proceed with a gel electrophoresis DNA recovery of the promoters and vector bands to ensure that we are only adding the desired DNA in the ligation reaction.
• Set several ligation reactions at different insert:vector ratios to find the most optimal.

We expect to obtain successful results in the coming weeks.

References

---

[1] Gibson, E. G., Oviatt, A. A., & Osheroff, N. (2020). Two-Dimensional Gel Electrophoresis to Resolve DNA Topoisomers. Methods In Molecular Biology, 15-24. https://doi.org/10.1007/978-1-0716-0323-9_2
[2] Rubinov, M., & Sporns, O. (2010). Complex network measures of brain connectivity: Uses and interpretations. NeuroImage, 52(3), 1059–1069. https://doi.org/10.1016/j.neuroimage.2009.10.003
[3] Waterhouse, E. G., An, J. J., Orefice, L. L., Baydyuk, M., Liao, G.-Y., Zheng, K., Lu, B., & Xu, B. (2012). BDNF Promotes Differentiation and Maturation of Adult-born Neurons through GABAergic Transmission. Journal of Neuroscience, 32(41), 14318–14330. https://doi.org/10.1523/JNEUROSCI.0709-12.2012