Differential expression of genes involved in the acute innate immune response to intracortical microelectrodes

Abstract

Higher order tasks in development for brain-computer interfacing applications require the invasiveness of intracortical microelectrodes. Unfortunately, the resulting inflammatory response contributes to the decline of detectable neural signal. The major components of the neuroinflammatory response to microelectrodes have been well-documented with histological imaging, leading to the identification of broad pathways of interest for its inhibition such as oxidative stress and innate immunity. To understand how to mitigate the neuroinflammatory response, a more precise understanding is required. Advancements in genotyping have led the development of new tools for developing temporal gene expression profiles. Therefore, we have meticulously characterized the gene expression profiles of the neuroinflammatory response to mice implanted with non-functional intracortical probes. A time course of differential acute expression of genes of the innate immune response were compared to naïve sham mice, identifying significant changes following implantation. Differential gene expression analysis revealed 22 genes that could inform future therapeutic targets. Particular emphasis is placed on the largest changes in gene expression occurring 24 h post-implantation, and in genes that are involved in multiple innate immune sets including Itgam, Cd14, and Irak4.

1. Introduction

Intracortical microelectrodes have been explored as the interface with the brain for brain computer interfacing (BCI) since the 1990’s. Neural signals captured using intracortical microelectrodes can be used to inform systems that can restore movement and function to patients with sensorimotor disabilities and or other neurological disease. To do so, intracortical microelectrodes acquire single unit neuronal signals that are sent to a computer which decodes neural activity into a patient’s intent to move. Algorithms translate desire to move to a signal instructing a cursor, prosthetic, or a patient’s own limb to move accordingly [1–3].

Due to the level of invasiveness of the implanted BCI technology, the benefits need to outweigh the risks. Specifically, intracortical microelectrodes need to be able to sustain reliable and consistent recorded signals throughout the duration of device implantation [4]. Currently, fluctuations and steady declines in the quality of detected neural signals over time limit the utility for long-term human use [5–7]. Multimodal failure mechanisms contribute to the stability [8], one of the most significant challenges being the self-destructive inflammatory response [9–11]. Upon implantation, the immune system quickly responds resulting in an immediate and imminent neuronal death at the probe interface, ultimately yielding poor neural signal acquisition. The immediate portion of the neuroinflammatory response is facilitated by the innate immune response. The innate immune response is part of the body’s first line of defense against foreign invaders such as pathogens or foreign bodies.

The innate immune response senses and processes external and internal danger signals, initiates a response to clear the danger, and prompts other aspects of the immune system to respond further if necessary [12]. The main components of the innate immune response include physical barriers (i.e. epithelial surfaces), the complement system, effector cells, and subsequent cytokine and chemokine response. Broadly, effector cells such as astrocytes, microglia, and blood-derived macrophages have been implicated in neuronal death at the tissue-probe interface resulting in poor electrode performance [13, 14]. Effector cells have pattern recognition receptors (PRRs), which are used to detect damage and initiate intracellular signaling cascades that coordinate the inflammatory response through cytokine and chemokine production.

There is increasing emphasis in further understanding the intricacies of the biological response at the tissue-probe interface to better inform more targeted therapeutic approaches to decrease neuroinflammation [15]. To accomplish this, one can elucidate the expression of genes involved in the inflammatory response. A temporal delineation of a gene expression profile can detail the activity of a large number of relevant genes providing benefit over traditional means of assessing the inflammatory response, such as immunohistochemistry. High information content can advance the identification of key mechanisms to target the reduction of neuroinflammation leading to improvements of intracortical microelectrode performance.

We have established that two specific pathways are critical to the connection between microelectrode-induced neuroinflammation and microelectrode performance: oxidative stress and innate immunity [16–22]. Oxidative stress is a result of the intense or prolonged inflammatory response associated with microelectrode implantation [21, 22]. We, and others, have previously looked at genes involved in oxidative stress [13, 22–25]. Since oxidative stress is downstream from the initiation of the innate immune response, a profile of the specific innate immune system-related genes could provide a more robust approach to attenuation of microelectrode-induced neuroinflammation. Therefore, here we investigated molecular changes of different subsets of the acute innate immune response driven by damage resulting from the intracortical probe implantation. Using gene expression profiling, changes in innate immune gene expression was measured over time and compared to no-surgery shams. Understanding how gene expression profiles change immediately following probe implantation provides valuable insights into inflammatory mechanisms at the tissue-probe interface. By looking at key players within each subset of the innate immune response and determining if there are any commonalities, we can find key targets to decrease the neuroinflammatory response.

2. Material and methods

2.1. Probes

“Dummy”, non-recording, single shank, uncoated, Michigan-style silicon probes (2 mm × 15 μm × 123 μm) were used as intracortical probes as previously described [26]. Probes were washed in 95% ethanol, three times for five minutes each wash, and sterilized via a cold ethylene oxide gas cycle as previously described in Ravikumar et al. (2014) [27].

2.2. Animals

Prior to surgery, mice were housed in groups of 3–5 with food and water provided ad libitium, while maintained on a 12-hour light/dark cycle. At the time of surgery, all mice were 8 to 10 weeks of age (Male C57/BL6, strain #000,664, Jackson Laboratories). After surgery, mice were individually housed to avoid injury to the implant site. All animal handling was performed in a class II sterile hood using microisolator techniques. There were a total of 15 animals utilized in this study, designated as three animals used per time point (6 h, 24 h, 72 h, and 2 weeks) and three animals used as sham controls. Sham animals are defined as naïve non-surgery controls. Sham animals were all male, age matched, and did not have any pre-, post- or surgical procedures. All procedures and animal care practices were performed in accordance with a protocol approved by the Case Western Reserve University Institutional Animal Care and Use Committee.

2.3. Surgical implantation of probes

Animals were prepped for surgery using established methods [17]. Briefly, after mice reached surgical plane using isoflurane (3% in 1.0 L/min O₂₎, mice were mounted onto a stereotaxic frame for the remainder of the surgical preparations and the surgery itself. A single dose of 0.2 ml of 0.25% Marcaine was administered subcutaneously (SQ) around the surgical site as a topical anesthetic, and Meloxicam (2 mg/kg) and Buprenorphine (0.05 mg/kg) were administered subcutaneously as analgesics. The surgical site was shaved and sterilized with alternating wipes of betadine and isopropanol. Following which, a midline incision was used to expose the skull and four craniotomies were made over the motor and sensorimotor cortex in each brain hemisphere (1.5 mm lateral and 1 mm anterior and posterior to bregma respectively) [28] were created using a high-speed dental drill with a 0.45 mm size bit. Adequate breaks in the drilling were taken to prevent thermal damage [29]. In order to ensure repeatability, the same surgeon manually inserted one dummy probe per craniotomy by hand, for a total of four probes implanted per animal. Insertion speed has been shown to be a contributor to electrode recording quality, likely due to surgical trauma during electrode insertion [30]. Therefore, our method of implantation is a potential limitation to the study. However, the same-trained surgeon, limiting variability, performed all implantations. Nonetheless, precautions were made to ensure repeatable implantations with minimal surgical trauma. Probes were held at the wide tab segment using fine tip forceps and slowly implanted into the cortex at a rate of ~2–3 mm/s until the 2 cm long shank was implanted. The wide tab ensured that the same depth of probe shank was implanted per craniotomy/animal, because it is too wide to be implanted and always remained above the cortical tissue. Following which, the probe was tethered to the skull with silicone elastomer and dental acrylic. Lastly, the incision site was sutured closed using 5–0 monofilment polypropylene suture. Post-operative care and monitoring was provided the week following surgery, with administration of Meloxicam (2 mg/kg) analgesia for three days following surgery.

2.4. Tissue processing

Sham and implanted animals were euthanized via cardiac perfusions in order to prepare tissue extracts for analysis. Briefly, deeply anesthetized animals, were transcardially perfused with cold 1X phosphate-buffered saline (PBS) followed by 30% (w/v) sucrose. Brains were immediately extracted and probes (if implanted) were explanted. Perfusion and explantation was done quickly to prevent excessive degradation of RNA. Brain tissue was flash frozen in optimal cutting temperature compound (OCT) on dry ice. Using a cryostat, the cortical brain tissue surrounding the neural probes was sectioned at 150 μm thick frozen slices and collected onto glass microscope slides. A biopsy punch (1 mm deep and 500 μm diameter) centered directly over the implant site of the collected tissue on the glass slide, was used to immediately excise the tissue of the frozen tissue slices, resulting in tissue samples each 150 μm thick and with 250 μm radii from the implant site. This step was repeated for a total of 6–7 tissue slices, resulting in a collection of 900 – 1050 μm total thickness of cortical tissue down the electrode shaft. Tissue collection started at ~150 μm depth, continuing down the length of the device, spanning most of the cortex. Tissue was collected from all four implant sites per animal, but only two implant sites per animal were utilized for further downstream RNA isolation and gene expression analysis to account for technical replicates within each biological animal replicate.

2.5. RNA isolation

Excised brain tissue samples were directly placed in microcentrifuge tubes containing Qiazol (RNA extraction lysate). Extracted tissue was homogenized in Qiazol using 1.5 mm zirconium beads using the Bead Bug Homogenizer at high speed. RNA was immediately extracted and purified using the RNeasy Plus Universal Mini Kit in accordance with manufacturer’s protocol. RNA was stored at −80 °C until further processing. RNA purity and concentrations were determined using Qubit Flourometer. Samples were shipped on dry ice to the Nanostring processing facility (Seattle, WA) for further quality control and quantification.

2.6. RNA quantification

For each sample, 50 ng of total RNA was hybridized to 777 capture probes provided by Nanostring at 96 °C overnight with the nCounter® Mouse Neuroinflammation Plus panel. The nCounter® Mouse Neuroinflammation Plus panel (with custom add-ins genes of interest) interrogated 777 neuro-immune-related mouse genes, associated controls, and 13 housekeeping genes. Hybridized samples were digitally analyzed for frequency of each gene using the nCounter® MAX Analyzer.

2.7. Statistical analysis

Raw mRNA abundance frequency data were imported into nSolver® software Version 4.0. Default quality control checks were run to assess technical assay performance. Raw counts were normalized to internal controls and housekeeping genes to remove any technical variance. Internal controls include positive and negative controls to assess success of hybridization. Background noise was filtered out using thresholding which utilized the mean count of negative controls to substitute all raw counts at or below estimated background to the threshold value. The geometric mean of the positive controls was used to compute the positive control normalization factor. After 2 housekeeping genes were removed due to low average counts, the geometric mean of the housekeeping genes was used to calculate the CodeSet Content Normalization. Low count genes were removed by setting the Threshold Count Value to 25. Observation frequency of the Threshold Count Value was also set to 0.9. The Threshold Count Value was chosen to achieve balance between false positives and false negatives in the data.

Normalized data was Log transformed by base 2 before analysis to achieve a normal distribution of the data. From here on in, log2-fold changes will just be referred to as “fold change” for simplicity. Fold-change in gene expression levels was calculated by dividing individual expression values (normalized reads) for each time point by the mean of values from shams. Each gene was tested for differential expression in response to ‘time’ (covariate). For each gene, a negative binomial regression without dispersion was fit using ‘time’ to predict expression.

Functional analyses were performed using Advanced Analysis nSolver® software Version 2.0.115. Gene sets were curated by NanoString and informed by KEGG analysis pathways. Gene Set Analysis Scores were output by the software. Gene Set Analysis Scores summarize differential expression testing at the gene set level. Gene Set Analysis Scores are calculated as the square root of the mean squared t-statistic for the genes in a gene set (Eq. (1)), with t-statistics coming from the linear regression underlying the differential expression analysis. A correction for multiple comparisons due to the multiplicity given by the many genes was included in the nSolver software analysis. Statistical significance was performed in R 3.6 using the MASS package. The average fold change from the technical replicates within each animal was first calculated and then used in a generalized linear model (GLM) with a Poisson distribution (n = 3). For the GLM, each time point was evaluated as a categorical variable. Genes regulated differentially by more than 1 Log 2 fold from sham condition with a p < 0.05 were considered significant.

$$\text{Gene Set Analysis Score = }{\left(\frac{1}{p}\sum _{i=1}^p{t}_1^2\right)}^{1/2}$$

Eq. (1). Calculation of Gene Set Analysis Scores for each gene set. t_i: t-statistic from ith pathway gene

3. Results

The investigation of gene expression profiles from mice implanted with neural probes (n = 3 animals with a technical duplicate per animals) at various acute time points was compared to sham controls (n = 3 animals with a technical duplicate per animal), resulting in 6 samples tested for each time point. One sham sample was removed due to high technical variance, resulting in 5 sham samples from 3 different animals. Gene expression profiles were assessed using a digital molecular multiplex barcoding array to allow for 777 genes to be analyzed from each sample. Out of the 777 genes of the panel, 590 genes met the Threshold Count requirements and were used for analysis. The present study only examined the genes related to gene sets curated by NanoString involving the innate immune system (101 genes) (Table 1). All genes examined exhibited an up-regulation compared to sham at least one time point due to intracortical probe implantation. None of the genes demonstrated significant down-regulation compared to sham (data not shown).

Table 1.

Genes from the innate immune response included in the gene sets used for analysis: complement cascade, cytokine response, chemokine response, toll-like receptor (TLR) cascade, and pattern recognition receptors (PRRs).

Gene symbol	Gene name	Role in the Immune Response
AIM2	Interferon Inducible protein	Pattern recognition receptors
Atm	ataxia telangiectasia mutated	NFκB pathway
Bcl10	B cell leukemia/lymphoma 10	NFκB pathway
Bcl2	B cell leukemia/lymphoma 2	NFκB pathway
Bcl2a1a	B cell leukemia/lymphoma 2 related protein A1a	NFκB pathway
Bcl2l1	BCL2-like 1	NFκB pathway
Birc2	baculoviral IAP repeat-containing 2	Cytokine response, NFκB pathway, TLR pathway
Birc3	baculoviral IAP repeat-containing 3	Cytokine response
Blnk	B cell linker	NFκB pathway
Btk	Bruton agammaglobulinemia tyrosine kinase	NFκB pathway
C1qa	complement component 1, q subcomponent, α polypeptide	Complement and coagulation
C1qb	complement component 1, q subcomponent, β polypeptide	Complement and coagulation
C1qc	complement component 1, q subcomponent, C chain	Complement and coagulation
C3	complement component 3	Complement and coagulation
C3ar1	complement component 3a receptor 1	Complement and coagulation
C5ar1	complement component 5a receptor 1	Complement and coagulation
Casp1	caspase 1	Cytokine response
Casp3	caspase 3	Cytokine response
Casp8	caspase 8	TLR pathway
Cd14	CD14 antigen	Pattern recognition receptors, TLR pathway, NFκB pathway
Cd36	CD36 molecule	TLR pathway
Cd45	CD45 antigen	Cellular markers
Cd68	CD68 antigen	Cellular markers
Cflar	CASP8 and FADD-like apoptosis regulator	NFκB pathway
Chuk	conserved helix-loop-helix ubiquitous kinase	Cytokine response, NFκB pathway, TLR pathway
Csf1	colony stimulating factor 1 (macrophage)	Cytokine response
Csf1r	colony stimulating factor 1 receptor	Cytokine response
Csf2rb	colony stimulating factor 2 receptor, β, low-affinity (granulocyte-macrophage)	Cytokine response
Ctss	cathepsin S	TLR pathway
Ddx58	DEAD (Asp-Glu-Ala-Asp) box polypeptide 58	Cytokine response, NFκB pathway
Dusp7	dual specificity phosphatase 7	Cytokine response, TLR pathway
F3	coagulation factor III	Complement and coagulation
Fos	FBJ osteosarcoma oncogene	Cytokine response, TLR pathway
GFAP	glial fibrillary acidic protein	Cellular markers
Fyn	Fyn proto-oncogene	Cytokine response
Hmgb1	high mobility group box 1	Cytokine response, TLR pathway
Ifnar1	interferon (α and β) receptor 1	Cytokine response
Ifnar2	interferon (α and β) receptor 2	Cytokine response
Ikbkb	inhibitor of kappaB kinase β	Cytokine response, NFκB pathway, TLR pathway
Ikbkg	inhibitor of kappaB kinase gamma	NFκB pathway
Il10rb	interleukin 10 receptor, β	Cytokine response
Il1a	interleukin 1 α	Cytokine response
Il1b	interleukin 1 β	NFκB pathway
Il1r1	interleukin 1 receptor, type I	NFκB pathway
Il1rap	interleukin 1 receptor accessory protein	Cytokine response
Il2rg	interleukin 2 receptor, gamma chain	Cytokine response
Il6ra	interleukin 6 receptor, a	Cytokine response
Inpp5d	inositol polyphosphate-5-phosphatase D	Cytokine response
Irak1	interleukin-1 receptor-associated kinase 1	Cytokine response, NFκB pathway, TLR pathway
Irak2	interleukin-1 receptor-associated kinase 2	Cytokine response, TLR Pathway
Irak4	interleukin-1 receptor-associated kinase 4	Cytokine response, NFκB pathway, TLR pathway
Irf3	interferon regulatory factor 3	Cytokine response, TLR pathway
Irf7	interferon regulatory factor 7	TLR pathway
Itgam	integrin α M	Pattern recognition receptors, Complement and coagulation, TLR pathway
Jun	jun proto-oncogene	Cytokine response, TLR pathway
Lgmn	legumain	TLR pathway
Ltbr	lymphotoxin B receptor	Cytokine response, NFκB pathway
Lyn	LYN proto-oncogene, Src family tyrosine kinase	NFκB pathway
Map2k4	mitogen-activated protein kinase 4	Cytokine response, TLR pathway
Map3k1	mitogen-activated protein kinase 1	TLR pathway
Map3k14	mitogen-activated protein kinase 14	Cytokine response
Mapk10	mitogen-activated protein kinase 10	Cytokine response, TLR pathway
Mapk14	mitogen-activated protein kinase 14	Cytokine response, TLR pathway
Myd88	myeloid differentiation primary response gene 88	Cytokine response, NFκB pathway, TLR pathway
NFκB1	nuclear factor of kappa light polypeptide gene enhancer in B cells 1, p105	Cytokine response, NFκB pathway, TLR pathway
NFκB2	nuclear factor of kappa light polypeptide gene enhancer in B cells 2, p49/p100	Cytokine response
NFκBia	nuclear factor of kappa light polypeptide gene enhancer in B cells inhibitor, α	Cytokine response, NFκB pathway, TLR pathway
Nod1	nucleotide-binding oligomerization domain containing 1	Cytokine response, Pattern recognition receptors, TLR pathway
Osmr	oncostatin M receptor	Cytokine response
Parp1	poly (ADP-ribose) polymerase family, member 1	NFκB pathway
Pik3ca	phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit α	Cytokine response
Pik3cb	phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit β	Cytokine response
Pik3r2	phosphoinositide-3-kinase regulatory subunit 2	Cytokine response
Plcg2	phospholipase C, gamma 2	NFκB pathway, TLR pathway
Prkaca	protein kinase, cAMP dependent, catalytic, α	Cytokine response
Prkcq	protein kinase C, theta	NFκB pathway
Pros1	protein S (α)	Complement and coagulation
Psmb8	proteasome (prosome, macropain) subunit, β type 8 (large multifunctional peptidase 7)	Cytokine response
Ptgs2	prostaglandin-endoperoxide synthase 2	NFκB pathway
Ptpn6	protein tyrosine phosphatase, non-receptor type 6	Cytokine response
RAGE	receptor for advanced glycation endproducts	Pattern recognition receptors
Rela	v-rel reticuloendotheliosis viral oncogene homolog A (avian)	Cytokine response, NFκB pathway, TLR pathway
Relb	avian reticuloendotheliosis viral (v-rel) oncogene related B	Cytokine response, NFκB pathway
Rig-1	retinoic acid-inducible gene I	Pattern recognition receptors
Ripk1	receptor (TNFRSF)-interacting serine-threonine kinase 1	NFκB pathway, TLR pathway
Ripk2	receptor (TNFRSF)-interacting serine-threonine kinase 2	Cytokine response, TLR pathway
S100b	S100 protein, β polypeptide, neural	Cytokine response, TLR pathway
Serpine1	serine (or cysteine) peptidase inhibitor, clade E, member 1	Complement and coagulation
Smarca4	SWI/SNF related, matrix associated, actin dependent regulator of chromatin, subfamily a, member 4	Cytokine response
Sqstm1	sequestosome 1	Cytokine response
Stat1	signal transducer and activator of transcription 1	Cytokine response
Sumo1	small ubiquitin-like modifier 1	Cytokine response
Syk	spleen tyrosine kinase	Cytokine response, NFκB pathway
Tlr2	toll-like receptor 2	Pattern recognition receptors, TLR pathway
Tnfrsf12a	tumor necrosis factor receptor superfamily, member 12a	Cytokine response
Tnfrsf1a	tumor necrosis factor receptor superfamily, member 1a	Cytokine response, NFκB pathway
Tnfrsf1b	tumor necrosis factor receptor superfamily, member 1b	Cytokine response
Tnfrsf25	tumor necrosis factor receptor superfamily, member 25	Cytokine response
Tnfsf12	tumor necrosis factor (ligand) superfamily, member 12	Cytokine response
Traf2	TNF receptor-associated factor 2	Cytokine response, NFκB pathway
Traf3	TNF receptor-associated factor 3	Cytokine response, NFκB pathway, TLR pathway
Traf6	TNF receptor-associated factor 6	Cytokine response, NFκB pathway, TLR pathway
Vamp7	vesicle-associated membrane protein 7	Cytokine response
Vav1	vav 1 oncogene	Cytokine response
Xiap	X-linked inhibitor of apoptosis	NFκB pathway

3.1. Cellular response to implanted microelectrode

We have previously shown that the inhibition of specific innate immunity pathways reduces the glial scar and neuronal death and the probe/tissue interface, and improves recording quality [16–18]. To connect our previous studies that utilized traditional immunohistochemistry to the gene expression profiles measured here, we first looked at the genes commonly used to code for proteins with immunohistochemistry: astrocytes (Gfap), infiltrating leukocytes (Cd45), and activated microglia/macrophages (Cd68) [13, 31, 32]. It was found that all three genes were significantly upregulated at all time points examined (6 h, 24 h, 72 h, and 2 w post-implantation) compared to sham controls (Fig. 1, Table 2). Furthermore, all three cellular genes were also significantly upregulated at 24 h, 72 h, 2 w compared to 6 h post-implantation, suggesting that transcriptional changes due to the implantation of the probe are not maximized until at least after 6 h post-implantation (Fig. 1, Table 2).

Fig. 1.

Differential expression of genes for commonly used markers of astrocyte and microglial/macrophage activity compared to sham. A. Gfap, commonly used as anastroglial marker B. Cd45, an infiltrating leukocyte marker C. Cd68, commonly used as a marker for microglia/macrophage activation; Fold change is a log2-fold changecompared to sham for each time point post-implantation of dummy probe. * indicates (log2-fold change > 1and p < 0.05).

Table 2. Statistical comparisons between time points of genes commonly used as markers of astrocyte and microglial/macrophage activity.

Bold values represent conditions in which p value is <0.05, indicating statistical significance. n = 3.

	Sham v 6 h	Sham v 24 h	Sham v 72 h	Sham v 2 w	6 h v 24 h	6 h v 72 h	6 h v 2 w	24 h v 72 h	24 h v 2w	72 h v 2 w
Gfap	2.33E-04	8.50E-27	4.07E-41	3.68E-24	9.74E-13	4.46E-23	5.85E-11	5.49E-03	5.55E-01	7.62E-04
Cd45	6.05E-03	8.37E-08	4.79E-13	2.46E-08	6.30E-05	2.00E-13	1.21E-05	8.05E-06	6.49E-01	5.01E-05
Cd68	1.14E-02	1.28E-12	9.99E-31	2.28E-16	1.24E-06	2.58E-22	1.37E-09	1.85E-07	2.03E-01	7.40E-05

3.2. The complement cascade

The complement system is a component of the innate immune system comprised of plasma proteins which can opsonize foreign substances for clearance and destruction by phagocytes, such as microglia and macrophages, and induce inflammation. As with any injury, the complement system is activated when an intracortical probe is implanted. There has been minimal research investigating the role of the complement system in the response to implanted probes, thus leaving a gap of knowledge that is necessary to be addressed.

A heatmap depicting the genes expressed in the complement cascade is shown in Fig. 2A. The Complement Cascade Gene Set Differential Expression Score (Eq. (1)), a summary score that describes the collective expression of complement associated genes compared to sham, fluctuates throughout the time course of the study (Fig. 2A, B). At 6 h, there is a collective upregulation in complement factors, followed by a further increase at 24 h post-implantation. By 72 h post-implantation, the extent of upregulation slightly subsides compared to 24 h post-implantation, remaining fairly consistent through the 2 w time point (Fig. 2A, B).

Fig. 2.

Differential expression of gene set involved in complement cascade compared to sham. A. Heat map of differential expression of each gene of this set significantly upregulated at least one time point (log2-fold change > 1and p < 0.05) compared to sham. B. Gene Set Differential Expression Score, a composite score for gene set. C.Top differentially expressed genes from this gene set. Fold change is a log2-fold change and compared to sham for each time point post-implantation of dummy probe. * indicates (log2-fold change > 1 and p < 0.05).

It was found that the genes that had the highest differential expression within the complement cascade included: C3ar1, C3, C5ar1, and Itgam (Fig. 2A, C). There was a significant fold increase for C3ar1 at all time points assessed compared to sham (Fig. 2C, Table 3). Additionally, there was a significant increase of Itgam upregulation from 6 h to 24 h post-implantation which decreased by 2 weeks (Fig. 2C, Table 3). C3 gene expression was significantly upregulated at 2 weeks compared to sham 6, 24 and 72 hrs. (Fig. 2C, Table 3). Of note, by 2 weeks post-implantation, there was a significantly less up-regulation of C3ar1, C5ar1, and Itgam compared to 24 h post-implantation detected (Fig. 2C, Table 3). Thus, the complement cascade is a subset of the innate immune response that warrants further investigation at early time points.

Table 3. Statistical comparisons between time points for genes for complement cascade.

Bold values represent conditions in which p value is <0.05, indicating statistical significance. n = 3.

	Sham v 6 h	Sham v 24 h	Sham v 72 h	Sham v 2 w	6 h v 24 h	6 h v 72 h	6 h v 2 w	24 h v 72 h	24 h v 2w	72 h v 2 w
C3ar1	4.65E-09	4.51E-15	1.29E-16	7.15E-06	1.84E-02	3.91E-03	1.19E-01	5.94E-01	1.01E-04	1.04E-05
C3	2.98E-01	1.53E-01	6.15E-02	2.45E-02	9.24E-02	1.16E-04	1.08E-08	6.32E-03	1.91E-07	6.06E-03
C5ar1	1.32E-07	1.09E-12	1.18E-06	8.50E-02	1.75E-03	5.10E-01	7.35E-06	1.67E-04	2.90E-12	8.64E-05
Itgam	1.97E-01	1.56E-08	9.46E-19	1.11E-04	2.42E-06	7.39E-17	6.29E-03	7.55E-06	2.94E-02	1.30E-10

3.3. Cytokine response

Cells involved in the innate immune response, post intracortical probe implantation, release proinflammatory mediators such as cytokines and chemokines. Cytokines are pleiotropic polypeptides (8–26 kDa) that are chemical messengers of the innate immune response. Cytokines orchestrate cellular events and are the key players in the dynamic cross-talk among all the cell types involved in the innate immune response to intracortical probe implantation [33]. Chemokines are discussed in the following section. The Gene Set Analysis scores summarize differential expression testing compared to sham at the gene set level. At acute time points, the genes associated with cytokine expression were at their highest upregulated levels at 24 h post-implantation (Fig. 3A, B). Interestingly, there was a decline in the collective cytokine response gene expression from 24 h to 72 hr, which was followed by an increase at 2 w post-implantation (Fig. 3A, B).

Fig. 3.

Differential expression of gene set involved in cytokine response compared to sham. A. Heat map of differential expression of each gene of this set significantly upregulated at least one time (log2-fold change > 1and p < 0.05) compared to sham. B. Gene Set Differential Expression Score, a composite score for gene set. C. Top differentially expressed genes from this gene set. Fold change is a log2-fold change and compared to sham for each time point post-implantation of dummy probe.*indicates (log2-fold change > 1and p < 0.05).

Notably, the genes that had the highest differential expression within the cytokine response pathway, compared to sham, included: Il1β, Il2rγ, Irak4, Osmr, Psmb8, Ptpn6, and Tnfrsf1a (Fig. 3A, C). There was a significant –fold increase of Irak4, Osmr, and Tnfrsf1a at all time points assessed, compared to sham (Fig. 3C, Table 4). Moreover, there was a significant upregulation of Il2rγ, Irak4, and Psmb8 at 24 h, 72 h, and 2 w post-implantation compared to 6 h post-implantation (Fig. 3C, Table 4). Inversely, it was observed that Il1β was at the highest upregulation at 6 h post-implantation (Fig. 3C, Table 4). Furthermore, Ptpn6 was significantly upregulated at 24 h, 72 h, 2 w post-implantation compared to sham (Fig. 3C, Table 4). By 2 weeks post-implantation, there was significantly less up-regulation of Osmr, Ptpn6, and Tnfrsf1a compared to 24 h post-implantation (Fig. 3C, Table 4). Many signaling molecules involved in the cytokine response are strongly upregulated early in the neuroinflammatory response to probes, suggesting that targeting strategies need to be present near or before the time of implantation.

Table 4. Statistical comparisons between time points for genes in cytokine response.

Bold values represent conditions in which p value is <0.05, indicating statistical significance. n = 3.

	Sham v 6 h	Sham v 24 h	Sham v 72 h	Sham v 2 w	6 h v 24 h	6 h v 72 h	6 h v 2 w	24 h v 72 h	24 h v 2w	72 h v 2 w
Il1β	1.96E-05	4.03E-02	1.03E-02	7.64E-02	1.71E-09	2.19E-08	1.44E-09	2.40E-01	5.92E-01	9.58E-02
Il2rγ	5.79E-02	9.75E-06	6.50E-10	1.63E-07	2.10E-03	5.28E-08	3.22E-05	4.39E-03	2.08E-01	1.04E-01
Irak4	2.08E-02	3.13E-06	1.02E-08	3.59E-05	5.83E-03	4.18E-05	3.50E-02	1.44E-01	4.95E-01	3.31E-02
Osmr	2.08E-07	1.41E-08	1.61E-05	9.08E-04	5.80E-01	3.22E-01	3.82E-02	1.23E-01	8.83E-03	2.76E-01
Psmb8	4.25E-01	8.40E-05	1.88E-11	1.03E-08	1.08E-03	4.16E-10	2.17E-07	8.16E-04	3.31E-02	2.16E-01
Ptpn6	2.82E-01	4.58E-07	4.55E-09	5.82E-05	1.85E-07	5.22E-09	9.73E-06	1.99E-01	1.45E-01	6.68E-03
Tnfrsf1a	5.07E-06	3.42E-09	2.29E-09	1.18E-03	1.02E-01	8.63E-02	1.35E-01	9.34E-01	1.96E-03	1.49E-03

3.4. Chemokine response

Chemokines are a class of cytokines which specifically mediate immune cell activation and migration to and from injured tissue. There are three main groups of chemokines based on the location of cysteine (C) residues: the CXC, CC, and CX3C [34]. The Gene Set Analysis scores summarize the expression compared to sham at the gene set level. As a set, the genes involved in the chemokine signaling are at their highest upregulation at 24 h post-implantation (Fig. 4A, B). Similar to the cytokine response gene expression trends, we observed a continual decline in the extent of up-regulation of the collective chemokine response genes from 24 h to 72 h and from 72 h to 2 w post-implantation (Fig. 4A, B).

Fig. 4.

Differential expression of gene set involved in chemokine response compared to sham. A. Heat map of differential expression of each gene of this set significantly upregulated at least one time (log2-fold change > 1and p < 0.05) compared to sham. B. Gene Set Differential Expression Score, a composite score for gene set. C. Top differentially expressed genes from this gene set. Fold change is a log2-fold change and compared to sham for each time point post-implantation of dummy probe.* indicates (log2-fold change > 1and p < 0.05).

The top differentially expressed chemokine response genes compared to sham were: Ccl2 and Dock2. There was a significant fold increase of Ccl2 at all time points assessed compared to sham (Fig. 4C, Table 5). Dock2 was significantly upregulated at 24 h, 72 h, and 2 weeks compared to sham (Fig. 4C, Table 5). Compared to the 6 h time point, Ccl2 was significantly less up-regulated at 24 h, 72 h, and 2 weeks (Fig. 4C, Table 5). Conversely, Dock2 had a significant fold increase at 24 h, 72 h, and 2 weeks compared to 6 h (Fig. 4C, Table 5). Similar to the cytokine response, chemokines were strongly upregulated within a day post implantation indicating chemokine targeting strategies need to be present at the site of implantation immediately following implantation.

Table 5. Statistical comparisons between time points for genes in chemokine response.

Bold values represent conditions in which p value is <0.05, indicating statistical significance. n = 3.

	Sham v 6 h	Sham v 24 h	Sham v 72 h	Sham v 2 w	6 h v 24 h	6 h v 72 h	6 h v 2 w	24 h v 72 h	24 h v 2w	72 h v 2 w
Ccl2	4.94E-09	1.13E-06	6.91E-07	8.95E-06	2.85E-02	4.78E-02	1.88E-03	8.32E-01	3.53E-01	2.54E-01
Dock2	6.27E-01	2.16E-08	1.03E-11	7.27E-06	1.59E-07	8.42E-11	4.34E-05	1.40E-01	1.92E-01	5.68E-03

3.5. Toll-like receptor (TLR) signaling cascade

TLRs are some of the most well characterized among PRRs that detect damage due to injury, such as that occurring during intracortical probe implantation. This class of receptors recognize molecular patterns on exogenous or endogenous damaged moieties. Ligand binding to TLRs result in necrosis factor κB (NF-κB) activation leading to inflammatory mechanisms [35]. When activated, the TLR signaling cascade leads to the secretion of proinflammatory cytokines and chemokines such as TNF, IL1β, IL1R, IL6, IL8, IL10, IL23, MIP-1 α/1β, and IFNβ [36, 37]. After probe implantation, some of these cytokines and chemokines have been implicated in the facilitation of the entry of blood-derived inflammatory cells to the brain, increases of blood brain barrier (BBB) permeability, and activation of glial cells [33]. Gene Set Analysis scores in Fig. 5A are a summary score that describes the collective expression of TLR-associated genes compared to sham. The TLR-associated gene set was elevated at all time point examined, and collectively at their highest expression at 24 h post-implantation (Fig. 5A, B). The collective expression of genes involved in the TLR signaling cascade was less up-regulated at 72 h than at 24 h, followed by a slight increase in up-regulation at 2 w, from the 72 h time point (Fig. 5A, B).

Fig. 5.

Differential expression of gene set involved in TLR cascade compared to sham. A. Heat map of differential expression of each gene of this set significantly upregulated at least one time (log2-fold change > 1and p < 0.05) compared to sham. B. Gene Set Differential Expression Score, a composite score for gene set. C. Top differentially expressed genes from this gene set. Fold change is a log2-fold change and compared to sham for each time point post-implantation of dummy probe. * indicates (log2-fold change > 1and p < 0.05).

The top differentially expressed genes from the TLR-associated gene set were: Casp8, Cd14, Irak4, Irf7, and Itgam. Itgam and Irak4 were previously discussed in Sections 3.2 and 3.3, respectively. There was a significant fold increase of Cd14 at all time points assessed compared to sham (Fig. 5C, Table 6). Whereas Irf7 was significantly up-regulated at 72 h and 2 weeks compared to sham (Fig. 5C, Table 6). Upregulation of Cd14 was significantly higher at 6 h and 24 h compared to 72 h and 2 w post-implantation (Fig. 5C, Table 6). There was a significant fold increase of Casp8 at 24 h, 72 h, and 2 weeks compared to 6 h (Fig. 5C, Table 6). By 2 w post-implantation, Cd14 was significantly less up-regulated, and Irf7 was significantly up-regulated, compared to 24 h post-implantation (Fig. 5C, Table 6). Overall, the TLR pathway is upregulated immediately following probe implantation suggesting that TLR associated genes need to be further explored for potential therapeutic targeting immediately after implantation.

Table 6. Statistical comparisons between time points for genes in TLR cascade.

Bold values represent conditions in which p value is <0.05, indicating statistical significance. n = 3.

	Sham v 6 h	Sham v 24 h	Sham v 72 h	Sham v 2 w	6 h v 24 h	6 h v 72 h	6 h v 2 w	24 h v 72 h	24 h v 2w	72 h v 2 w
Casp8	1.79E-01	1.37E-04	1.86E-08	1.95E-03	4.27E-03	2.61E-07	4.66E-02	4.22E-03	3.36E-01	2.04E-04
Cd14	4.04E-09	1.49E-07	1.04E-03	2.07E-02	2.73E-01	5.68E-05	4.64E-07	2.70E-03	4.21E-05	2.09E-01
Irak4	2.08E-02	3.13E-06	1.02E-08	3.59E-05	5.83E-03	4.18E-05	3.50E-02	1.44E-01	4.95E-01	3.31E-02
Irf7	7.17E-01	3.98E-01	9.41E-05	7.34E-04	6.25E-01	2.51E-04	1.87E-03	1.07E-03	7.00E-03	5.40E-01
Itgam	1.97E-01	1.56E-08	9.46E-19	1.11E-04	2.42E-06	7.39E-17	6.29E-03	7.55E-06	2.94E-02	1.30E-10

3.6. Pattern recognition receptors

Pattern recognition receptors (PRRs) are the molecular links between tissue damage and innate immunity. When certain molecular patterns or structures, including both PAMPs and DAMPs from serum proteins or damaged cells resulting from implantation of the probe, bind to PRRs, signaling of inflammatory processes ensues. Thus, the gene expression of PRRs was of particular interest to investigate in this study. Of specific interest were the PRRs residing on the cell surface, as they are released after probe implantation. Because of the role of receptor binding in cell phenotype, it is hypothesized that PRRs will be feasible future therapeutic target which can be used to block downstream cellular processes.

Fig. 6 A provides Gene Set Analysis scores, a score summarizing the collective expression of PRRs compared to sham. Genes associated with PRRs are upregulated at all time points included in this study. Collectively, the PRRs gene expression was at the highest upregulation at 24 h post-implantation, and return to approximately the same level of up-regulation at both 72 h and 2 w, as to that detected at 6 h post-implantation (Fig. 6 A, B).

Fig. 6.

Differential expression of gene set of pattern recognition receptors compared to sham. A. Heat map of differential expression of each gene of this set significantly upregulated at least one time point (log2-fold change > 1and p < 0.05) compared to sham. B. Gene Set Differential Expression Score, a composite score for gene set. C. Top differentially expressed genes from this gene set. Fold change is a log2-fold change and compared to sham for each time point post-implantation of dummy probe. * indicates (log2-fold change > 1and p < 0.05).

Aim2, Cd14, and Itgam were the top differentially expressed genes compared to sham. Because of the biological redundancies in pathways, Itgam and Cd14 were previously discussed in Sections 3.2 and 3.5, respectively. There was a significant fold increase of Aim2 at 24 h, 72 h, and 2 w post-implantation compared to sham (Fig. 6 C, Table 7). Moreover, Aim2 expression levels were significantly higher at 72 h and 2 weeks compared to 6 h (Fig. 6 C, Table 7). Since PRRs were upregulated post intracortical probe implantation and remain to be upregulated up for 2 weeks, methods to reduce PRR signaling should continue to be explored to reduce inflammation.

Table 7. Statistical comparisons between time points for genes of PRR.

Bold values represent conditions in which p value is <0.05, indicating statistical significance. n = 3.

	Sham v 6 h	Sham v 24 h	Sham v 72 h	Sham v 2 w	6 h v 24 h	6 h v 72 h	6 h v 2 w	24 h v 72 h	24 h v 2w	72 h v 2 w
Aim2	1.49E-01	1.61E-02	4.71E-08	4.41E-05	3.10E-01	9.60E-06	4.27E-03	3.71E-04	5.76E-02	8.01E-02
Cd14	4.04E-09	1.49E-07	1.04E-03	2.07E-02	2.73E-01	5.68E-05	4.64E-07	2.70E-03	4.21E-05	2.09E-01
Itgam	1.97E-01	1.56E-08	9.46E-19	1.11E-04	2.42E-06	7.39E-17	6.29E-03	7.55E-06	2.94E-02	1.30E-10

4. Discussion

There has been widespread interest investigating the neuroinflammatory response after intracortical probe implantation [10, 31, 38, 39] Recent progress has utilized gene expression analysis to evaluate the molecular pathways following implantation [22, 40, 41]. However, a comprehensive assessment of the molecular sequelae directed at the acute innate immune response post probe implantation, was yet to be performed. Therefore, this study quantitatively analyzed gene expression profiles from implanted mice at four time points during the acute inflammatory response and compared them to a non-surgical sham control. In order to have a comprehensive molecular analysis of the acute innate immune response, genes involved in the inflammatory cellular markers, the complement response, the cytokine response and the chemokine response were investigated. Since our lab has previously found targeting the TLR/CD14 pathway yields promising results in both improving intracortical microelectrode recording and neuronal density [16, 17], it was critical to continue investigation of this pathway. Furthermore, we explored PRRs that could also be potential therapeutic targets, as we have previously shown CD14 to be [18].

There were 101 different genes examined from the innate immune response. As expected, our findings indicate that genes associated with the innate immune response are significantly upregulated in the neuroinflammatory response following probe implantation. This is consistent with previous studies demonstrating the inflammatory response to intracortical probe implantation [31, 42, 43]. Within the subsets of the innate immune response, we identified 22 genes that were significantly upregulated during the first two weeks post intracortical probe implantation. The Gene Set Analysis score for all the gene within those five pathways demonstrated a robust upregulation of genes involved in the innate immune response 24 h post-implantation (Figs. 2B, 3B, 4B, 5B, 6B), suggesting that the innate immune response should be targeted at this time point.

Some highly upregulated genes of this study were involved in multiple subsets of the innate immune response. Itgam, Cd14, and Irak4 genes had the highest upregulation compared to sham, and are involved in multiple innate immunity pathways. Itgam is involved in the complement cascade, TLR signaling, and is a PRR (Figs. 2, 5, 6). Cd14 is involved in TLR signaling and functions as a PRR (Figs. 5 and 6). Irak4 is involved with both cytokine and TLR signaling (Figs. 3 and 5).

ITGAM [44], CD14 [45], and IRAK4 [46] are all involved in NF-κB signaling. ITGAM (CD11b) is a subunit for the Mac-1 (complement receptor 3 (CR3)) integrin. ITGAM is expressed on both blood-derived cells and microglia, playing a role in neutrophil and macrophage adhesion, activation, and fusion [44, 47, 48]. Activated by TLR pathway [49], ITGAM is also involved in the phagocytosis of complement coated particles [50]. The integrin is a fibrinogen [51], fibronectin [52], and collagen [53] receptor which mediates cell adhesion to these proteins, all of which can adsorb onto the probe surface [54]. Notably, Davalos et al. demonstrated that targeting ITGAM can decrease neuroinflammation. They demonstrated that blocking the fibrinogen binding motif recognized by ITGAM decreases microglial response and axonal damage in experimental autoimmune encephalomyelitis (EAE) [55]. Monoclonal antibodies to ITGAM have been developed and future studies could test the ability of this antibody to reduce intracortical probe induced inflammation [56].

Another of the strongly upregulated genes common to multiple gene sets in this study is Cd14. CD14 is a PRR involved in the recognition of TLR2 and TLR4 by its ligands. The CD14/TLR4 complex recognize DAMPs originating from cells or the extracellular matrix [37] and CD14 is mandatory for microglial reactions to DAMPs [57]. Some DAMPs CD14 is a co-receptor for include hsp70, a common DAMP released by necrotic cells, and S100A9, another DAMP released by neutrophils (blood derived cells present in the few days post-implantation) during inflammation [58, 59]. We have previously demonstrated that mice lacking Cd14 had decreases in both neuronal death and astroglial scar around the probe interface [17]. We have also shown that targeting CD14 can improve intracortical microelectrode performance [16, 18]. Since Cd14 is at its highest upregulation at 6 h and 24 h post-implantation, the best therapeutic window to target CD14 is suggested prior to 24 h post-implantation (Figs. 5C, 6C). Future studies should evaluate this recommended dosing regimen to avoid unnecessary treatment.

Irak4 is also highly upregulated at acute time points post intracortical probe implantation (Figs. 3C and 5C). Irak4 is involved in intracellular signaling cascades downstream of ligand binding to TLRs [60]. This molecule transduces pro-inflammatory downstream signals through both physical protein-protein interaction and kinase activity [61]. There are several IRAK4 inhibitors currently being developed through small-molecule screening of the IRAK4 kinase which were reviewed by Wang et al. [62]. Additionally, Nyrada is developing an IRAK4 inhibitor which is able to cross the BBB which may show promise to mitigate intracortical probe induced inflammation [63].

The following sections highlight in detail the specific genes within each innate immunity pathway that were significantly upregulated.

4.1. The complement system

The complement system is integral to the innate immune system and has not been explored with regards to intracortical probe implantation. Complement induces phagosome mediated loss of neurons in the substatnia nigra as well as microglial associated synapse loss, suggesting that the complement cascade could lead to these same results in the cortex post intracortical probe implantation [64]. The complement system can have deleterious effects [65]; however, there are studies which demonstrate that the complement system can be neuroprotective through its clearance of damage [66]. Conversely, even though the complement pathway can promote the inflammatory response, there is support for its beneficial role in promoting the health of neurons

To enhance the understanding of the complement system, genes involved with this system were investigated temporally (Fig. 2). Interestingly, the complement cascade had a slightly different time course than the other four pathways evaluated in this study. Although the genes involved in the complement cascade were similarly at the highest upregulation at 24 h, the complement cascade experienced a much slower decrease in the extent of up-regulation from 24 h to 2 w post-implantation compared to the other four pathways analyzed (Figs. 2B, 3B, 4B, 5B, 6B). Moreover, C3 and C3ar1 had the highest upregulation compared to the other genes involved in the complement response (Fig. 2A, C). C3 is a downstream effector of NF-κB, found primarily in astrocytes [67] and involved in phagocytosis [68], especially of the elimination of unwanted synapses [69]. Upregulation of C3 around the implant site suggests that further experiments are needed to determine whether synapses are being removed around the probe via C3 opsonization. C3 can also mediate macrophage adhesion to different implanted surfaces, suggesting that C3 is a protein which adsorbs onto the implanted probe, inducing a cellular response through detection by its receptor, C3ar [70]. C3 could potentially be involved with the initial stages of the innate immune response around the probe to clean up cellular debris resulting from implantation. The chemotactic receptor, C3ar, has been shown to be upregulated following CNS injury potentially promoting the migration of microglia, macrophages, and astrocytes towards the probe [71, 72].

As can be seen by the roles the upregulated genes from the complement cascade have in neuroinflammation, this pathway is a major factor that influences the outcome to intracortical probe implantation. Unfortunately, due to controversial reports, it is still unclear whether the complement system’s role in neuroinflammation are beneficial or detrimental. Therefore, before approaches can be developed to “target” any of the genes identified in this grouping, a more mechanistic study must be completed to further elucidate the function in response to this “injury” model.

4.2. Cytokine response

There were numerous molecules associated with cytokine signaling significantly upregulated post neural probe implantation compared to sham. Specifically, Il1β, Il2rγ, Irak4, Osmr, Psmb8, Ptpn6, and Tnfrsf1a (Fig. 3) were all expressed at higher levels in implanted animals than in sham. Irak4 was discussed above as it is involved with multiple gene sets examined in this study. IL1β is a hallmark cytokine involved in the acute inflammatory response and produced by both blood-derived cells and microglia. [73]. IL1β is commonly thought of as a first response cytokine [74], correspondingly, it was found that the highest upregulation of IL1β in this study was at the 6 h post-implantation (Fig. 3C). In the brain, IL1β activates endothelial cells of the BBB, enabling the trafficking of blood-derived cells into the parenchyma [11, 75]. We have previously demonstrated that monocyte trafficking into the brain following probe implantation dominates the cell population at the probe tissue interface [13], ultimately perpetuating the inflammatory response. Similarly, Karumbaiah et al. examined the gene response of several cytokines for multiple probe designs 72 h post-implantation. They found that Il1β was significantly upregulated for all probe designs examined, as demonstrated in our study for single-shank 15 μm Michigan style probes (Fig. 3C) [33]. Targeting IL1β immediately after probe implantation might be neuroprotective and lead to more robust recordings as targeting IL1B has been found to reduce neural ischemic damage [76, 77].

Other notable cytokines that were upregulated include, Il2rγ and Tnfrsf1a, which comprise portions of the IL-2 receptor and TNF cell surface receptor, respectively. The IL-2 ligand/receptor complex is involved in the signaling of other cellular players of the immune response [78]. Both the IL-2 and the TNF complex are known to lead to downstream NF-κB activation and are viable targets for as therapeutic methods to reduce intracortical probe induced inflammation [79, 80]. Conversely, Osmr and Ptpn6 negatively regulate the inflammatory response [81]. OSMR signaling inhibited both recruitment of monocytes and NF-κB signaling in a peritoneal model of acute inflammation. Furthermore, the ligand for OSMR, oncostatin M, regulates neuronal function and survival [82, 83]. PTPN6 is expressed both in astrocytes and neurons. PTPN6 negatively regulates astrocyte proliferation post ischemic injury [84] supporting functional recovery after brain injury [85]. Thus, controlled modulation of OSMR and PTPN6 signaling could be of therapeutic benefit after probe implantation.

4.3. Chemokine signaling

Chemokines mainly attract blood-derived cells and microglia to the area of the probe after implantation; however, they orchestrate other cell to cell interactions as well. The top differentially expressed genes compared to sham within the chemokine signaling pathway include: Ccl2 and Dock2. CCR2/CCL2 is a ligand-receptor pair that mediates macrophage infiltration into the brain resulting in neuronal loss [86, 87]. In fact, Hsieh et.al. implicated CCR2 in macrophage induced neuronal loss after traumatic brain injury (TBI), which is in accordance with other models of CNS injury, including experimental autoimmune encephalomyelitis (EAE) and hemorrhagic stroke [88–90]. The main ligand for CCR2 is CCL2 which was highly upregulated as early as 6 h post-implantation (Fig. 4C). Since it has been shown that the amount of blood-derived cells at the tissue-probe interface temporally correlate with neuronal loss [13], targeting the CCL2/CCR2 pathway to mitigate blood-derived cell infiltration might be a potent therapeutic to reduce neuronal loss following probe implantation. Sawyer et al. demonstrated that targeting the CCL2/CCR2 pathway, both via a knock-model and CCR2 antagonist, improves the neuroinflammatory response to intracortical probes. Targeting the CCL2/CCR2 pathway reduced BBB permeability at 2 weeks. Furthermore, targeting CCR2 resulted in higher neuronal density within 100 μm of the probe [91]. Next steps should include the administration of small-molecule drugs available to target the CCR2/CCL2 ligand-receptor pair, to see if effective dosing can be determined to improve microelectrode performance.

4.4. TLR cascade

The TLR family detects endogenous tissue damage and elicits innate immune responses through NF-κB. The highest upregulated genes involved in this cascade are: Casp8, Cd14, Irak4, Irf7, and Itgam. Itgam, CD14, and Irak4 were already discussed above.

Caspase 8 is a protease involved in several cell death mechanisms and in microglial activation [92]. Caspase 8 is expressed in neurons [93], macrophages, and microglia [94] and has been found to directly regulate IL1β in response to TLR3 and TLR4 activation [95]. Caspase 8 can prevent necroptosis [96], regulate cytokine transcription, and is a negative regulator of inflammasomes [97]. Furthermore, neuron specific casp8 deletion in mice resulted in decreased neurodegneration and neuronal cell death in a TBI model [93]. Thus, caspase 8 could be a key mediator of neuron death in intracortical probe induced inflammation. Considering that neuronal death by the probe limits the functionality of the device, future studies should explore caspase 8 and other potential neuronal death mediators in more detail.

4.5. Pattern recognition receptors

PRRs play an essential role in coordinating the innate immune response to clear damage. Given that cell-surface receptors are a popular target for therapeutic agents, we explored the families of PRRs (regardless or not if they are cell-surface receptors). The different families of PRRs converge with similar signaling cascades which induce the activation of pro-inflammatory transcription factors such as NF-κB. As mentioned above, the activation of NF-κB mediates the production of cytokines and chemokines which promote blood-borne cell infiltration and cellular activation.

In addition to Cd14 and Itgam (discussed in detail above), Aim2 was the highest upregulated gene of the PRRs analyzed. AIM2 is part of the AIM2 inflammasome, a cytosolic sensor which detects mislocalized self-DNA, suggesting a loss of cellular integrity [98]. It is unclear what DNA AIM2 is detecting post intracortical probe implantation. However, we speculate that cellular damage which occurs due to the implantation trauma could cause mislocalized DNA. Hyperactivation of AIM2 has been shown to mediate pyroptosis, a mechanism of proinflammatory cell death in neurons [99]. Thus, pyroptosis could be a possible neuronal death mechanism at the neural interface that needs explored.

5. Conclusion

A crucial step to mitigate the inflammatory response is to identify the molecular mediators involved. This study comprises an important stride towards understanding temporal changes to the biology at the tissue-probe interface. Here, the use of gene expression analysis was utilized to better understand the molecular players in the innate immune response to intracortical microelectrodes and discover new targets to mitigate neuroinflammation to intracortical microelectrodes. Since the innate immune system is fast responding, we first focused on acute time points post-implantation. Our findings corroborate studies that demonstrate the critical role the innate immune response plays in the neuroinflammatory response to intracortical probes. An increased temporal understanding of the innate immune response can inform more targeted intervention strategies to mitigate inflammation at the tissue-probe interface. This work identified Cd14, Irak4 and Itgam, among others, to be promising candidates to reduce the intracortical microelectrode mediated inflammatory response, and to potentially improve the function and stability of the device. CD14 has already been demonstrated to be a key player in the inflammatory response to intracortical microelectrodes, however future studies need to be conducted to confirm of biological relevance of IRAK4 and ITGAM.

Upregulated genes could be further explored using knock-out models or RNA interference (RNAi) to further validate ITGAM and IRAK4 as viable, potent targets to reduce neuroinflammation. Future studies also need to explore the effects of knock out models such as Itgam^−/−, and Irak4^−/− on electrophysiology in order to correlate the gene expressions with the decline of detectable neural signals commonly observed in control animals. Further analysis deciphering the contribution of specific cell types, including peripheral immune, on the changes of gene expression and protein levels, is a next step towards uncovering the neuroinflammatory response to implanted microelectrodes. Additional analysis could also warrant the collection of appropriate data for future modeling to fit kinetic or spatial profiles of the inflammatory response over time. The findings of this study can inform local delivery of immunomodulatory agents, such as dissolvable probe coatings, to reduce inflammation at the probe interface. Collectively, this study provides an important resource for future studies exploring the role innate immunity plays in intracortical microelectrode failure.

Statement of Significance.

Current understanding of the cellular response contributing to the failure of intracortical microelectrodes has been limited to the evaluation of cellular presence around the electrode. Minimal research investigating gene expression profiles of these cells has left a knowledge gap identifying their phenotype. This manuscript represents the first robust investigation of the changes in gene expression levels specific to the innate immune response following intracortical microelectrode implantation. To understand the role of the complement system in response to implanted probes, we performed gene expression profiling over acute time points from implanted subjects and compared them to no-surgery controls. This manuscript provides valuable insights into inflammatory mechanisms at the tissue-probe interface, thus having a high impact on those using intracortical microelectrodes to study and treat neurological diseases and injuries.