Translational modeling of non-traumatic raised intracranial pressure in rodents; the impact of diet-induced obesity on cephalic hypersensitivity and retinal degeneration

Elevated intracranial pressure (ICP) is a feature of critical cerebral disorders. Obesity has been linked to raised ICP, and especially to disorders such as idiopathic intracranial pressure (IIH). We aimed to explore the impact of diet-induced obesity (DIO) on ICP, cephalic sensitivity and structural retinal changes with the dual goal of developing a disease model for non-traumatic raised ICP and IIH. Rats were fed high-fat diet or matched control diet. To assess pain sensitivity, Von Frey and light/dark box testing were performed. Dual energy x-ray absorptiometry scanning was used to measure body composition. Optic nerve head and retinal structures were evaluated using optical coherence tomography. Intraocular pressure was assessed. Rats were then implanted with telemetric device for continuous ICP recording. At the end, eye histology and molecular analysis on choroid plexus (CP) and trigeminal ganglion (TG) were performed. The DIO rats had double the abdominal fat mass. ICP was 55% higher in obese rats (p=0.003). Altered pain thresholds were found in DIO rats as denoted by a lower periorbital threshold (p=0.0002). Expression of Calca and Trpv1 was elevated in TG. Furthermore, a peripapillary retinal nerve fiber layer swelling (p=0.0026) with subsequent neuroretinal degeneration p=0.02) was detected in DIO rats. There was a trend to increased expression of AQP1 and NKCC1 at CP. This study demonstrates for the first time that DIO leads to raised ICP, with clinically relevant sequalae. Our novel model for non-traumatic raised ICP could expand the knowledge regarding disorders with elevated ICP and IIH.


Introduction
Elevated intracranial pressure (ICP) is observed in a range of cerebral pathologies, such as traumatic brain injury (TBI), ischemic stroke, hydrocephalus and idiopathic intracranial hypertension (IIH).
These disorders are associated with significant morbidity and are amongst the most debilitating, costly brain disorders [1]. Currently, limited pharmacotherapies exist to treat patients with elevated ICP, as the mechanisms governing cerebrospinal fluid (CSF) production and ICP regulation and their dysregulation in pathology remain unresolved. The molecular pathway by which water crosses the choroid plexus (CP), the primary site of CSF secretion, is not fully characterized. However the characterization of aquaporin water channels (AQPs) and water co-transporting proteins such as the Na + -K + -2Cl --1 cotransporter (NKCC1) at CP have improved our understanding of this process [2].
Additionally, the effects of different physiological states on ICP and CSF secretion are poorly understood.
In recent years, studies have suggested that obesity is associated with ICP dynamics indicated by association between elevated BMI and ICP [3][4][5]. In support, a direct correlation between BMI, body fat percentage and CSF pressure has been demonstrated [5]. Furthermore, obesity is the major risk factor for development of IIH, which is a condition of increased ICP without identifiable cause. The condition affects mainly women and 90% of the patients are obese [6]. IIH has significant morbidity and the most common symptoms are chronic headache, cognitive impairment and impaired vision due to papilledema [7,8]. The severity of papilledema and visual loss in IIH is closely associated with increased BMI [9,10].
kHz gave a final sampling frequency of 1 kHz. After surgery and the two following days the rats received post-surgical treatments and were allowed to recover up to 7 days as described previously prior to co-housing [13]. On the postoperative days some of the rats lost weight (< 15%) and from day 7 weight increased as observed with this kind of procedure [13,14]. At the end of the experiments all animals were euthanized with an overdose of sodium pentobarbital. Few animals had to be sacrificed earlier due to adverse events such as opening of stiches and injection site reactions.

ICP analysis
ICP was measured for up to 30 days after implantation surgery. Presented means are for day of surgery, day 0 (n=9 CD, n=10 HFD), days 3, 7 and whole period (n=8 control, n=10 high fat). More specific spectral analysis of the ICP waveform was carried out on a 5-min stable section of ICP, i.e. without movement artefacts as previously described [13].

Measurement of hind paw and periorbital thresholds after mechanical stimulation
Von Frey testing was performed from diet start until day of surgery (day 0) (Fig.1). The sensitivity of the frontal region of the head (V1 ophthalmic trigeminal dermatome) and hind paw to static mechanical stimulation was measured using an electronic von Frey device fitted with a rigid plastic tip (IITC LifeScience Inc, USA). Testing was carried out as previously described [15]. The investigator was blinded to the experimental group during the experiment.

Light sensitivity
Photophobia was assessed by quantifying time spend in the light during a light/dark box test as previously described [16,17]. The light/dark box was placed in a dedicated room with video recording . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made where the rats were left to acclimatize before the test (HFD n= 24, CD n=24). Light intensity was set to 500 lux for 20 minutes. Time spend in the light (s) was assessed by blinded observers and defined as the eyes of the rats being visible in the light compartment of the box i.e. exposure to light. Photophobia was tested during their active period and for every round the rats were tested at the same day.

Optical coherence tomography and optic disk photography, image capture and analysis
Optical coherence tomography (OCT) and optic disk photography (ODP) of both eyes was performed prior to implantation of the ICP probe on day 0 (HFD n= 14, CD n=10, 48 eyes). OCT and ODP images were captured with a combined digital fundus microscope and image-guided spectral domain OCT system (Phoenix MICRON IV In Vivo Imaging Microscope and OCT2 system, Phoenix Technology). The corneal surface of eyes was topically anesthetized with Phenylephrine Hydrochloride 10% and pupils were dilated with Tropicamide 1%. The optic nerve head (ONH) was centered for the best OCT image. The OCT image was aligned horizontally, and optimized for best image quality through adjusting the polarization, contrast, gamma and dispersion. Peripapillary circular B-scans (1024 A-scans with a radius of 515 µm) were performed with 3-10 image averaging to reduce image noise. Finally, an ODP was captured with the optic disc centered and in focus.
From each eye, the best quality OCT images were analyzed using semi-automatic retinal segmentation software (InSight, Phoenix Technology). The inner limiting membrane (ILM), plexiform layer (IPL) and Bruch's membrane (BM) were automatically segmented and manually fitted for best accuracy, and additionally two layers (retinal nerve fiber layer (RNFL) and inner nuclear layer (INL)) were manually segmented. The thickness of the total retina (TR), the RNFL, the ganglion cell complex (GCC, which consists of the RNFL, the ganglion cell layer (GCL) and the IPL). All OCT segmentations and ODPs . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made were used to assess gene expression. Reactions were carried out in 384 well plates, singleplex 10µl . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint quantified using ImageJ (NIH, USA). The relative protein expression was normalized to the loading control (Lamin-B1) and presented as fold change from control.

Blood glucose
Blood glucose was measured on day 30, where the rats had been fasted overnight for at least 14h.
Blood from the lateral tail veins was used to assess blood glucose (mmol/L) using glucometer ACCU-CHEK Aviva (Roche Diagnostics, Denmark).

Statistics
Statistical analysis was performed using Graphpad Prism 9 (Graphpad Software Inc, USA). Data presented as mean ± SEM or mean ± SD as stated. Data normality assessed via Shapiro-Wilk normality test. Where data was normally distributed unpaired two-tailed t-tests (equal variance) or Welch's test (unequal variance) was employed, whereas non-parametric data was assessed via Mann-Whitney U test. Pearson's correlation coefficient (r) was used for assessing correlations. Results were judged significant at P<0.05.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made

DIO increases ICP
ICP telemeters were implanted in DIO and control rats in which ICP was continuously measured for 30 days. ICP was higher in DIO rats over the first 14 days (P=0.003, Fig 2A), where ICP on the day of surgery, day 0, correlated with abdominal fat percentage (r=0.54, P=0.016, Fig 2B). On day 0, DIO rats had raised ICP relative to control rats (2.77±0.57 mmHg vs -0.17±0.73, P=0.0052, Fig2 C, D). This raised ICP was associated with a right shift in ICP pressure frequency ( Fig 2E). On day three, a day without the influence of post-surgical drugs, ICP was more than double in the DIO rats (4.6±0.74 mmHg vs1.46±0.06, P=0.0052, Fig 2F,G) also with a right shift in ICP (Fig 2I,H). Similarly, on day 7 when the rats have fully recovered from surgery [13], DIO rats had a higher ICP (5.17±0.47 vs 3.53±0.55 mmHg, P=0.03, Fig 2I,J) with a right shift in ICP (Fig 2K). From day 15 to 30 of recording there was no difference in ICP (P=0.2, Fig 2L). In the final 7 days recording, DIO rats had a trend to higher ICP than the control rats (5.94±0.55 vs 4.32±0.6 mmHg, P=0.07, Fig 2M). This loss of . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint significance in ICP could be mediated by weight loss. Although both groups lost weight as expected [13,14], DIO rats lost more weight before regaining weight to a lesser amount (-7.5±0.6% vs 0.6±1.7%, P=0.0004, Fig 2N).
To further analyze the ICP phenotype, ICP waveforms were assessed. On the day 0, there was no difference in any ICP frequencies, including those of slow non-respiratory ICP waves

DIO induces headache like behavior
All rats had cephalic and peripheral mechanical thresholds assessed during course of the diets. In the rats that started the diet at 12 weeks, the DIO rats had a lower threshold over the course of the diet (P=0.0038, Fig 4A), demonstrating cyclical periods of sensitivity. Rats that started the diet at age 6 weeks, showed no difference in threshold (P=0.4). This difference may be due to the challenges of getting the younger animals habituated to the test. The hind paw data showed that no difference over the course of the diet (P=0.4, Fig 4B). (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint abdominal fat percentage (r=-0.65, P=0.0005, Fig 4E). ICP was not correlated with cephalic withdrawal thresholds on day 0 (r=-0.236, P=0.4). We assessed light sensitivity as a measure of photophobia. The control and DIO rats spend equal time in the light during the test (22.8 % ± 2.5 vs 24.3 % ± 2.4., P=0.69, Fig 4G).
To further interrogate the headache phenotype, we assessed gene expression in TG on day 30 (Fig 4F).
In the DIO rats, there was a 1.

Retinal swelling and degeneration in DIO
The ODPs in both groups appeared normal without any notable pathology or any signs of acute papilledema (Fig. 5). The OCT scans showed no significant thickness difference between right and left eye in any of the retinal layers (Table 2). No difference in thickness between DIO and control animals was found for TR, GCC, IRL and ORL (Fig. 5C) . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint

Discussion
In the current study we have investigated the impact of DIO on ICP and developed the first animal model for non-traumatic raised ICP demonstrating relevant sequalae mimicking the clinical IIH phenotype. Here we show that DIO rats develop raised ICP, headache related hypersensitivity and neuroretinal degeneration. This is critical, as recent data strongly suggests that obesity causes elevated ICP [3,11,18,19]. This is important considering the obesity pandemic, where the therapeutic approaches to management of elevated ICP are virtually absent and molecular mechanisms causing raised ICP remain unidentified. Our new animal model may change this.

Impact of obesity on ICP and CP alterations
A recent study demonstrated that in a mixed neurological patient cohort, BMI and percentage body fat both positively correlated with CSF opening pressure, although to a lesser degree than seen in IIH patients [5,20]. Here we demonstrate raised ICP in the DIO group, where the ICP was raised by 55% in the DIO rats, which associated with increased abdominal obesity in agreement with the clinical literature [5]. This study is congruent with our previous study where genetic obesity in rats caused 40% higher ICP [11]. We have recently shown that with normal weight gain in rats, there was a trend to an increase in ICP, supporting the concept of adiposity impacting ICP [13]. This data thus confirms excess adiposity alters CSF dynamics. Indeed previous data suggests that obesity increases CSF secretion in obese female rats without alterations in CSF drainage [21]. In agreement with this we demonstrate increased spectral power of non-respiratory slow waves after surgical recovery in DIO rats. This potentially indicates altered cerebral compliance and therefore cerebral pathology as seen in several human ICP disorders [22].
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint These data and those of others indicate that obesity has a direct effect on ICP, however no adipose derived factors have been explicitly demonstrated to raise ICP. However, glucocorticoids (GCs), hormones known to be increased tissue level activity in obesity, have been demonstrated to increase CSF secretion without altering CSF drainage suggesting that GC have the capacity to alter CSF dynamics [21,23]. Whether the observed increase in ICP in the obese animals is due to obesity, endocrine abnormalities or other neurochemical disturbances caused by obesity is yet unknown. Future investigations into obesity specific factors will help elucidate the molecular underpinnings of DIO mediated raised ICP. A technical limitation of the current study is not performing ICP recording during obesogenesis, thus preventing the observation of the temporal development of raised ICP. Although the use of female rats makes this work relevant to IIH, further studies assessing the effect of DIO on male rats will help complete our understanding.
In the current non-traumatic raised ICP model, there was only a trend to increased mRNA expression of Aqp1 and Slc12a2 at CP. This trend was also observed in protein expression, where there was tendency to increased ratio of glycosylated AQP1 to total AQP1. This is in contrast to our previous results where Zucker rats with elevated ICP had increased expression of AQP1 at CP [11]. This can be accounted for due to slower post-surgical weight gain in the DIO rats. The weight difference at 30 days was decreased to 3,4%. This may have affected the outcome in the molecular analysis and therefore requires further investigation. All together this suggests that the obesity-induced raised ICP may modulate CSF dynamics by altering the expression of the water channels but further studies are warranted here.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint

Obesity, raised ICP and headache development
Headache is associated in pathologies with raised ICP such as TBI, hydrocephalus and IIH [24]. Thus, headache is the key symptom in 75-94% of IIH patients at diagnosis [25][26][27]. In the present study, we show that DIO rats had cephalic cutaneous mechanical allodynia but not peripheral in the context of raised ICP, where the severity was closely associated with higher levels of abdominal obesity. The altered pain thresholds were cyclical after 5 weeks on the diet. Cephalic cutaneous allodynia is considered to be a measurement of central sensitization to pain and a marker of progression to migraine chronification in humans [28]. The present study adds novel data to the few experimental studies suggesting that obesity increases headache like behavior in rodents, where DIO increased facial pain behavior, and in contrast to our data, photophobic behavior in mice [29,30]. We did not find any difference in the light sensitivity between the DIO and control animals suggesting that these animals are not photophobic. Such difference could be accounted for different species between studies or use of fewer animals. Importantly, our data mirrors the human data, where obesity is associated with increased incidence of headache, where increased abdominal obesity confers both an increased risk of migraine diagnosis and increases cutaneous allodynia severity [28,31]. No link between ICP and cephalic thresholds was found, mimicking the clinical features of IIH, where ICP did not correlate with the reported headache severity [25]. In support of this, the vast majority of patients with IIH continue to experience disabling headaches after the other manifestations of the disorder including raised ICP has resolved [25].
In agreement with our behavioral data, we demonstrate increased expression of trigeminal ganglion Calca and Trpv1 in DIO rats. It is clear that CGRP plays a key role in headache disorders and inhibition of CGRP signaling has demonstrated clinical efficacy as migraine therapy [32]. Furthermore, . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint increasing attention has been given to the capsaicin-gated channel, transient receptor potential vanilloid subfamily member 1 (TRPV1), suggested to play a role in migraine and related sensitization [33].
Thus, the increased expression of these genes suggests increased molecular sensitivity of the TG to nociceptive stimuli. This cutaneous allodynia was present in the absence upregulated pro-inflammatory genes. The direct cause of the molecular alterations and therefore cutaneous allodynia remains uncertain. The raised ICP through direct mechanical stimulation of the dura mater and TG, through raised ICP, could increase sensitivity. Additionally, obesity related factors could be involved. Further work to elucidate the underpinnings of the cutaneous allodynia is required.

Neuroretinal changes
Raised ICP is known to cause papilledema and visual decline. In this study, the OCT data showed thickening of the peripapillary RNFL in DIO rats with a positive correlation to ICP, suggesting axonal swelling. Across both the DIO and the control animals, there was also a clear correlation between RNFL thickness and the total body weight and the abdominal fat percentage. This indicates abdominal obesity is a potential disease aggravating factor. Moreover, this highlights that ICP is likely involved in the retinal phenotype present in this experimental model. Histological analysis demonstrates that after a prolonged period of DIO and elevated ICP there was a significant thinning of RNFBs entering the optic disc. This is in accordance with clinical findings, where obesity is found to correlate with both raised ICP and RNFL thinning [34,35]. To our knowledge this is the first animal model to demonstrate a link between obesity induced ICP elevation and retinal nerve fiber degeneration. The link between ICP and neuroretinal degeneration has previously been demonstrated in mouse models of sustained and robust elevated ICP which caused papilledema with neuroretinal and optic nerve degeneration [36,37]. No signs of acute papilledema were detected in the present study, which may be explained by the mild ICP . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint increase that occurs in this non-traumatic model. Even in a rat hydrocephalus with markedly increased ICP (> 15 mmHg) only modest signs of acute papilledema were detected. The mechanisms underlying the RNFL swelling and subsequent degeneration remain unclear, future studies to determine if the observed phenomena are a consequence raised ICP or another obesity related process is required. The presence of similar fasting glucose and IOP levels suggest that diabetic retinopathy and glaucomatous optic neuropathy are not influencing factors in this model.

Conclusions
This study demonstrates for the first time the neuroophthalmological manifestations of DIO with raised ICP and the clinically relevant sequalae of headache like behavior and retinal degeneration in rodents.
These findings contribute to the understanding of the impact of obesity on brain health and the neurobiology of raised ICP disorders, especially IIH. This non-traumatic raised ICP model provides tools that can be utilized in the future evaluation of underlying molecular mechanisms and the development of novel ICP reducing agents to determine their efficacy. . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made . CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint Table 1. Characteristics of rats in each study: Weight, abdominal fat mass, abdominal fat % and lean mass in control and DIO rats in the headache, OCT and ICP studies. All tests are unpaired student's T-test. *=P<0.05, **=P<0.01, ***=P<0.001 and ****=P<0.0001). Data presented as mean±SD.
. CC-BY 4.0 International license available under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprint this version posted August 6, 2021. ; https://doi.org/10.1101/2021.08.05.455266 doi: bioRxiv preprint Table 2. OCT and histology data: There was no significant difference between right and left eye in any of the segmented layers of the peripapillary OCT scans in any of the groups. The mean RNFL from both eyes was significant thicker in the DIO animals. In the end of the study histology demonstrated