Upregulation of the Endogenous Peptide Antisecretory Factor Enhances Hippocampal Long-term Potentiation and Promotes Learning in Wistar Rats

—Antisecretory Factor (AF) is an endogenous peptide known for its powerful antisecretory and anti-inﬂammatory properties. We have previously shown that AF also acts as a neuromodulator of GABAergic synaptic transmission in rat hippocampus in a way that results in disinhibition of CA1 pyramidal neurons. Disinhibition is expected to facilitate the induction of long-term potentiation (LTP), and LTP is known to play a crucial role in learning and memory acquisition. In the present study we investigated the eﬀect of AF on LTP in CA3-CA1 synapses in rat hippocampus. In addition, endogenous AF plasma activity was upregulated by feeding the rats with specially processed cereals (SPC) and spatial learning and memory was studied in the Morris Water Maze (MWM). We found that LTP was signiﬁcantly enhanced in the presence of AF, both when added exogenously in vitro as well as when upregulated endogenously by SPC-feeding. In the presence of the GABA A -receptor antagonist picrotoxin (PTX) there was however no signiﬁcant enhancement of LTP. Moreover, rats fed with SPC demonstrated enhanced spatial learning and short-term memory, compared with control animals. These results show that the disinhibition of GABAergic transmission in the hippocampus by the endogenous peptide AF enhances LTP as well as spatial learning and memory. (cid:1) 2022 The Author(s). Published by Elsevier Ltd on behalf of IBRO. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).


INTRODUCTION
The Antisecretory Factor (AF) is a 41 kDa endogenous protein which act in vivo by regulating the transport of water and ions across the cellular membrane (Johansson et al., 1995;Lange and Lonnroth, 2001). AF presence has been demonstrated in all mammals' tissues and body fluids investigated so far (Johansson et al., 2009;Lange et al., 1999;Lange and Lonnroth, 2001). This protein has been ascribed to play a regulative, central role in the immune system due to its presence in the dendritic cells, in B-cells and in macrophages, but also in various forms of lymphoid tissues (Davidson and Hickey, 2004).
In the physiologically, normo-balanced organism most of the AF content is present in an inactive form. AF can, however, be stimulated to activity as well as to increased synthesis by e.g. bacterial toxins (Lange and Lonnroth, 2001), but also after oral intake of hydrothermally processed cereals commonly designated as SPC-FlakesÒ (specially processed cereals). This endogenous AF stimulation, achieved after intake of the processed cereals, is the result of a balanced content of sugars and amino acids in the feed and has been demonstrated to be beneficial for patients suffering from several inflammatory and/or hypersecretory conditions (Bjorck et al., 2000;Laurenius et al., 2003;Svensson et al., 2004;Finkel et al., 2004;Hanner et al., 2010). SPC-containing feed has also been demonstrated to induce AF in various animals including rats (Johansson et al., 2011).
The detailed AF mechanism of action remains to be described, but the nervous system seems to play a central role in close conjunction with binding proteins and membrane transport channels (Matson Dzebo et al., 2014;Nawrot-Pora z bka et al., 2015;Nicolas and Lievin-Le Moal, 2015;Bazzurro et al., 2018;Ilkhanizadeh et al., 2018). The conclusive results of AF action upon the various cellular activities is to bring the different constituent in the single cell as well as in the different tissues into a normo-regulative, physiological balance (Lonnroth et al., 2016;Johansson et al., 2018).
A high concentration of preformed, active AF can be demonstrated in the egg yolk of egg-laying hens fed an SPC-enriched diet (Goransson et al., 1993;Kaya et al., 2017). Oral intake of a spray dried form of this egg-yolk, designated SalovumÒ, has proven clinically effective in different pathophysiological conditions (Eriksson et al., 2003;Zaman et al., 2018;Gatzinsky et al., 2020). SPC-FlakesÒ and SalovumÒ have both been categorized as ''Food for specific medical purposes" by the Swedish and EU regulative, medical authorities. These products have been on the market since 1993, and so far no unwanted, medical side effects has been reported.
The AF protein has been sequenced, and the biologically, active site of the AF protein is located in the amino terminal part of the sequence (Johansson et al., 1997). Due to stability and potency, the amino acid sequence 36-51, designated AF-16, has been thoroughly tested in vivo as well as in vitro. AF-16 has been found capable of mediating anti-secretory as well as antiinflammatory effects in several different in vivo (Jennische et al., 2008;Al-Olama et al., 2011;Al-Olama et al., 2015;Lange et al., 2020) and in vitro (Rapallino et al., 2003) test systems. Conclusively, AF-16 was selected for our in vitro studies.
We have previously shown that AF also have an effect on GABAergic transmission in the rat hippocampus in that it gives a disinhibition of CA1 pyramidal cells (Kim et al., 2005). This disinhibition was mediated by an enhancement of tonic GABAergic inhibition of the CA1 stratum radiatum interneurons (Strandberg et al., 2014). Such a disinhibition would be expected to facilitate the induction of LTP (Wigstrom and Gustafsson, 1983) and possibly learning.
In the present study we investigated how AF affects long-term potentiation (LTP) in CA3-CA1 synapses in the rat hippocampus. Furthermore, since LTP is also involved in learning and memory (Mansvelder et al., 2019;Nicoll, 2017), we also investigated how AF might affect spatial memory and learning. These experiments were performed by testing the rats with the Morris water maze (MWM) test.

EXPERIMENTAL PROCEDURES Animals
Male adult (P35-60) Wistar rats bred at the local animal facility (EBM Gothenburg) were used for all experiments. Room temperature and humidity were kept constant at 20°C and 55%, respectively. Rats were given ad libitum access to food and water. Daylight cycle was maintained artificially, lights off between 18:00 and 07:00 h and the experiments were executed during the light phase. Animals were housed in cages of three to six animals per cage in a colony room. The study was performed according to Swedish legislation on animal welfare with ethics permission given by the Ethics Committee for Animal Experiments, Gothenburg and in accordance with the guidelines of the European Communities Council Directive of 24 November 1986 (86/609/EEC).
Animals used for experiments in MWM were housed for a minimum of 21 days before experiments begun in order to acclimatize them to their new environment as well as human handling. There was a significant difference in average weight between the treatment groups used for experiments in the MWM and subset of electrophysiological experiments before feeding with SPC-flakes begun (average within the SPC-group = 91. 33 g, n = 13, in controls = 108.70, n = 12, p = 0.021), but not on the first acquisition day of the MWM (average weight in the SPC-group = 218.71 g, in controls = 227.11 g, p = 0.39), nor on the day after the MWM was finished (average weight in the SPC group = 245.51 g, in controls = 251.29 g, p = 0.58).
For electrophysiological experiments on hippocampal slices from rats fed with SPC, 4 out of 10 slices in the SPC group and 3 out of 10 slices in the control group were from animals that did not perform the MWM task. There were no significant differences in the amount of LTP between slices from animals that performed MWM compared to slices from animals that did not. Experiments on slices from rats that did perform the MWM task were performed 3-20 days after the MWM task was ended for the SPC group and 4-20 days after the MWM task was ended for the control group.
Endogenous AF activity was upregulated by feeding rats with a standard rodent diet containing 5% SPC for at least 21 days before testing in MWM and/or electrophysiological experiments begun. It has previously been readily demonstrated in humans, pigs, chickens, rats and cows that feeding with ordinary, commercially available feed supplemented with SPC in a concentration range of 5-15% and a 18-21 days long feeding period, induce a stimulated, high endogenous AF activity when determined in plasma/milk with ELISA (Lange and Lonnroth, 2001). The researcher was blinded to treatment status of each animal during all experiments and statistical analysis.

Electrophysiology
Animals were deeply anaesthetized by isoflurane before decapitated using a guillotine. The brain was quickly removed, the cerebral hemispheres separated and put in an ice-cold (3°C) slicing solution containing (in mM); glycerol 219; KCl 2.5; NaH 2 PO 4 1.2; CaCl 2 1.2; MgCl 2 7; NaHCO 3 26; glucose11 which was saturated with 95% oxygen and 5% carbon dioxide. By using a vibratome (Microtome HM 650 V, Thermo Fisher Scientific, Loughborough, UK) the hemispheres were cut into 400 lm thick transverse slices in the same solution.
A cut between the CA1 and CA3 was made to prevent spreading of spontaneous network activity and the hippocampal slices were then transferred to a storage chamber with an artificial cerebrospinal fluid (ACSF) containing (in mM); NaCl 119; NaH 2 PO 4 1; NaHCO 3 26; KCl 3; glucose 10; CaCl 2 2; MgCl 2 4, perfused with 95% oxygen and 5% carbon dioxide and holding a temperature of 25°C (the higher concentration of Mg 2+ and lower temperature, compared to recording conditions, were used to reduce neuronal activity thus reducing potential excitotoxicity in the slice). For experiments where slices were incubated in AF-16 the storage solution also contained 0.5 lg/ml (284 nM) AF-16 (50 ml of a stock solution of AF-16 dissolved in water was added to 100 ml of ACSF). The researcher was blinded to treatment status during experiments and analysis. All slices were incubated in the storage chamber for a minimum of one hour and a maximum of 8 h. A single slice was then put in a recording chamber containing an ACSF solution (containing in mM; NaCl 123; NaH 2 PO 4 1; NaHCO 3 26; KCl 3; CaCl 2 2; MgCl 2 1; glucose 10 saturated with 95% oxygen and 5% carbon dioxide. In a subset of experiments the solution also contained 100 mM picrotoxin (PTX) to block GABA A receptors. The solution was continuously circulating at a speed of about 2-3 mL/min and kept at a temperature of 30°C.
Field recordings were performed in the stratum radiatum of the hippocampal CA1 region. Electrical stimulation of CA3-CA1 Schaffer collateral afferents was enabled by visually placing two stimulation electrodes (made of tungsten) in the middle of the stratum radiatum (CA1) on either side of the recording electrode. Field excitatory postsynaptic potentials (fEPSPs) were recorded in the stratum radiatum by a borosilicate glass micropipette filled with a 1 M NaCl solution and connected to a AgCl electrode. Recordings were made with a sampling frequency of 10 kHz and filtered at 1 kHz using a MultiClamp 700B amplifier and a Digidata 1440A data acquisition system (Molecular Devices, Sunnyvale, CA, USA). fEPSP magnitude was measured as the initial slope (linear regression over the first 0.8 ms) of the fEPSP rising phase. The magnitude of the presynaptic volley was measured with linear regression of the negative slope of the initial positivenegative deflection. Experiments in which the volley changed >10% from the level of the baseline just prior to the first tetanization were excluded. For input-output measurements fEPSPs were evoked at five different stimulation strengths (12 stimuli per stimulation strength) and the magnitudes of the fEPSPs and presynaptic volleys were measured for the averaged potentials for each stimulation strength. For each experiment the magnitudes of the fEPSPs were plotted against the magnitudes of the corresponding presynaptic volleys and a linear regression was calculated. The slope of the linear regression line was taken as the input-output measurement for that experiment. Paired-pulse ratio (PPR) measurements were done with an interstimulus interval of 50 ms and these measurements were made on averaged fEPSPs (in the same manner as for inputoutput measurements).
Stimulation intensity was set by manual inspection of the fEPSP and the intensity never over-reached the point at which population spike activity appeared. In all experiments a test frequency of 0.1 Hz was used. Before tetanizations a stable baseline of 15 min was established (experiments were excluded if fEPSP size changed >10%). LTP was then induced, first by a weak tetanization consisting of a train of 20 impulses at 100 Hz, one stimulation electrode at a time. This was followed by 30 min of test stimulation. After this LTP was induced again by a second strong tetanization consisting of five trains of 20 impulses at 100 Hz given every 20 s, simultaneously in both stimulating electrodes to enhance cooperativity, again followed by 30 min of test stimulation. The amount of potentiation was measured as the change in fEPSP size (average of 20 fEPSPs) at the end of each 30 minutes test stimulation compared to immediately before the first tetanization.
All data are presented as means ± standard error of the mean (SEM) unless stated otherwise. Statistical comparisons were made using nonparametric tests, Mann-Whitney U test or Wilcoxon signed ranks test, and results were considered statistically significant when p < 0.05.

Spatial learning and memory
Spatial learning and memory were tested in the MWM, a 1.6 m in diameter circular pool and a circular platform with a diameter of 10 cm were used. The water was held at 25 ± 1°C and made opaque by water soluble white paint (Hobbylack matt 250 ml white (#202101), Panduro, Gothenburg, Sweden). The platform was submerged 1 cm below the water surface. The pool was placed in a room with external cues that were kept in constant positions during the whole experiment. Data recordings were made by using the Water Maze Viewer Plug-in (BIOBSERVE, GE).
Rats (SPC-group n = 13 and control-group n = 12) were given an acclimatization training session in the water maze for 30 seconds (s) during which the platform was removed the day before acquisition begun. During the following four days rats had one swim session consisting of twelve swims every day, first and last trials being pre-respective post-probe trials (the platform removed). On the fifth day a single 60 s probe trial was performed. The platform position was held constant within each acquisition day but was changed between acquisition days. Starting position was randomized and varied so that rats were placed in the pool from one of four positions (SW, SE, NE, NW). The rat was always placed in the pool with its nose facing the pool wall and was then allowed to seek the platform for 60 s. If the rat managed to locate the platform within the time limit, it was allowed to rest on the platform for 20 s, if not, the rat was placed upon the platform and got to rest there for 20 s as well. The rat was then put in a plastic cage for 20 s during in which time the pool was cleaned from stools and odorants and the swim procedure was then repeated. After each swim session the rat was put back in its home cage. The MWM paradigm was originally developed by Morris (1984) and adapted by Baldi et al. (2005) to assess short-and longterm memory on each acquisition day. Since the aim of the present study was to assess a potential improvement of cognitive performance in healthy rats, the cognitive demand was increased by varying the platform position between days using the matching-to-place task (Steele and Morris, 1999;Wass et al., 2008). Spatial learning was assessed by recording mean time to find the platform (escape latency) on each of the acquisition days whereas spatial short-term memory was defined as time spent in the target quadrant during the post-probe test and finally, long lasting reference memory was defined as the time spent in the target quadrant on the pre-probe test. To assess development of search strategy, comparisons of mean time to find platform could be assessed across all four test days. Thus, taken together using this MWM paradigm, spatial learning, development of search strategy and spatial short-and long-term (i.e. reference) memory could all be investigated.
All data are presented as means ± SEM. Significance was defined as p 0.05 using two-tailed levels. Repeated measures two-way ANOVAs with treatment as betweensubjects factor and acquisition day/swim number/probetest day as within-subjects factor was used to analyze mean escape latency/time in target quadrant over the four acquisition days. One-way ANOVAs were used to compare escape latency and time in target quadrant between treatment groups on the different acquisition days. Post hoc pair wise comparisons were Bonferroni corrected.
T-tests were used to compare mean time to find platform on acquisition Day 1 on swim one, and on mean swim speed and average thigmotaxis between treatment groups.

AF-16 enhances LTP in hippocampal CA3-CA1 synapses
AF-16 reduces GABAergic inhibition onto CA1 pyramidal cells both when applied to hippocampal slices in vitro as well as when endogenous production of AF is upregulated (Kim et al., 2005). This disinhibition could potentially modulate hippocampal synaptic plasticity and in this study we hypothesize that AF enhances LTP and by doing so also enhances learning. To investigate the effect of AF on LTP we performed field recordings from CA3-CA1 glutamate synapses in rat hippocampal slices that had been incubated in AF-16 for 1 h prior to start of the recording and control slices without added AF-16. A disinhibition of the CA3-CA1 synapses could give rise to a lowering of the induction threshold for LTP as well as an increase in the amount of LTP. Therefor we used two induction protocols in series to induce LTP, a weak tetanization and a strong tetanization (see Experimental procedures). The amount of LTP measured 30 min after the given tetanizations was enhanced in slices incubated in AF-16 both after the weak tetanization (AF-16: 113 ± 2% of baseline, n = 8, compared to control: 103 ± 2% of baseline, n = 7, p = 0.0037, Mann-Whitney U test) as well as after the strong tetanization (AF-16: 133 ± 4% of baseline, n = 8, compared to control: 111 ± 3% of baseline, n = 7, p = 0.0037, Mann-Whitney U test) (Fig. 1). After the weak tetanization there was no significant potentiation compared to baseline in control slices (p = 0.091, Wilcoxon signed ranks test), whereas the potentiation in AF-16-incubated slices was significant (p = 0.012). There was no significant difference in fiber volley magnitude between the two groups (after weak tetanization, AF-16: 98 ± 2% of baseline, n = 8, compared to control: 100 ± 2% of baseline, n = 7, p = 0.19, Mann-Whitney U test, and after strong tetanization, AF-16: 98 ± 1% of baseline, n = 8, compared to control 97 ± 1% of baseline, n = 7, p = 0.78, Mann-Whitney U test).

AF-16 does not enhance LTP when GABA A receptors are blocked by PTX
To test the effect of AF on LTP when blocking GABAergic transmission, we also performed LTP measurements with the same protocol as previously but in the presence of the GABA A receptor channel blocker PTX. In the presence of PTX there was no significant difference in the amount of LTP neither after a weak tetanization (AF-16: 118 ± 5% of baseline, n = 11, compared to control: 114 ± 3% of baseline, n = 10, p = 0.92, Mann Whitney U test) nor after a strong tetanization (AF-16: 172 ± 6%, n = 11, compared to control: 161 ± 6%, n = 10, p = 0.25, Mann-Whitney U test) (Fig. 4). There was no significant difference in fiber volley magnitude between the two groups (after weak tetanization, AF-16: 99 ± 1% of baseline, n = 11, compared to control: 100 ± 2% of baseline, n = 10, p = 0.86, Mann-Whitney U test, and after strong tetanization, AF-16: 99 ± 2% of baseline, n = 11, compared to control 97 ± 1% of baseline, n = 10, p = 0.28, Mann-Whitney U test).

Spatial learning improved over test days and was improved by SPC treatment
The potential effects of SPC treatment on spatial learning and memory were investigated in the water maze task (Morris, 1984) as adapted by Baldi et al. (2005) with changes in platform position between days to increase cognitive demand (Steele and Morris, 1999;Wass et al., 2008).
Importantly, on day one, there was no difference in time to find platform between treatment groups on acquisition swim number one. Thus, there were no baseline differences between treatment groups in finding the platform before any acquisition training had occurred. In addition, no significant difference between the groups was detected regarding swim speed on the different acquisition days (Fig. 5A) and average thigmotaxis over the four acquisition days.
There was a significant main effect of acquisition day: F(3, 69) = 25.89, p < 0.001 on escape latency over the four acquisition days such that learning occurred in both treatment groups (Fig. 5B). Post hoc Bonferroni corrected analysis found significant differences in mean acquisition latency on Day 1 compared to; Day 2 (p < 0.05); Day 3 (p < 0.001); Day 4 (p < 0.001) as well as between Day 2; and Day 3 (p = 0.05): and Day 4 (p < 0.001), such that escape latency decreased over days, indexing spatial learning. There was a main effect of treatment: F(1, 23) = 4.19, p = 0.05 such that SPC treated animals displayed a shorter escape latency compared to controls. Repeated measures oneway ANOVA, between subject comparisons, found significant shorter escape latency for SPC treated animals on Day 1 (p = 0.04) and Day 4 (p = 0.03) (Fig. 5B). No significant interaction effect between treatment and day was identified.
To further investigate the effect of SPC-treatment on spatial learning, time to find platform over the 10 swims was compared between the groups on each day. On Day 1, there was a significant main effect of treatment F (1, 23) = 4.59, p = 0.04 such that SPC treated animals displayed a shorter mean time to find the platform compared with controls, between subject comparisons found significantly shorter latencies in SPC treated animals on swim number 4 (p = 0.01) and swim number 8 (p = 0.003). A significant main effect of swim number F(5.95, 136.90) = 9.33, p < 0.001 on escape latency and a trend towards an interaction effect between swim number and treatment, F(5.95, 136.90) = 2.11, p = 0.06 was identified (Fig. 6). Significant improvement in mean time to find platform during Day 1 was detected between swim number 1 and; 7 (p < 0.01); 8 and 9; 10 (p < 0.05), between swim number 2 and; 5, 9 and 10 (p < 0.01); 7 (p < 0.001); 8 (p < 0.05), between swim number 3 and; 5, 8, 9 (p < 0.05); 7 (p = 0.001); 10 (p = 0.01), as well as between swim number 4 and 7 (p > 0.05).
On day 2 and 3 there was a significant effect of swim number on swim latency; F(4.41, 101.49) = 16.64, p < 0.001, F(5.03,115.72) = 14.08, p < 0.001, respectively, but no effect of treatment, nor any interaction between the two.
On Day 4, there was a significant main effect of swim number F(3.66, 84.14) = 8.69, p < 0.001 and treatment Field recordings in the CA1 stratum radiatum in rat hippocampal slices treated with AF-16 (n = 8) and controls (n = 7). After establishing a stable baseline, a weak tetanization (a train of 20 impulses at 100 Hz delivered to the two different inputs separately, indicated by thin arrow) was given and followed by 30 minutes of testpulse stimulation (0.1 Hz). Thereafter a strong tetanization (five trains of 20 impulses at 100 Hz delivered every 20 seconds simultaneously to the two inputs, indicated by thick arrow) was given, also followed by 30 minutes of test-pulse stimulation. Values are plotted as means ± standard error of the mean (SEM). Insets in the top show representative traces (average of 20 records) for time points indicated in the graph, scale bars represent 300 mV and 10 ms. (B) Box-and-whiskers plots depicting the amount of LTP 30 minutes after the weak tetanization and the strong tetanization in slices treated with AF-16 (black boxes) and control slices (grey boxes) respectively. In this and the following graphs the boxes depict median as well as 25th and 75th quartiles, the whiskers depict the 5th and 95th percentiles and individual values are shown as open circles. **p < 0.01.
F(1, 23) = 5.13, p = 0.03, such that SPC treated animals had shorter swim latencies. A trend towards an interaction effect between swim number and treatment, F(3.66, 84.14) = 2.32, p = 0.07, was found (Fig. 7). Between subject comparisons found that SPC treated animals had significantly shorter latencies on swim 1 (p = 0.03) compared with placebo treated animals. Significant improvement in mean time to find platform during Day 4 was detected between swim number 1 and; 4 and 8 (p < 0.05); 5 and 9 (p < 0.01) as well as swim number 2 and 8 and 9 (p < 0.05).

Post-probe performance improved over test days and was enhanced by SPC treatment on day 1
A trend towards a main effect of the treatment; F(1, 21) = 3.51, p = 0.08 and a significant effect of day; F (3, 63) = 13.77, p < 0.001 on time spend in the target quadrant during the post-probe test and a significant interaction between day and treatment F(3, 63) = 3.34, p = 0.03, was identified (Fig. 8A). SPC treated animals spent more time in the target quadrant, compared with placebo animals on Day 1 (p = 0.03) and a trending effect on Day 3 (p = 0.06). Post hoc pairwise comparisons identified significant differences between Days 1 and 4 (p = 0.01), 2 and 4 (p = 0.01) and between 3 and 4 (p < 0.001), such that animals spent more time in the target quadrant on Day 4 compared to all other days (Fig. 8A).

Pre-probe performance improved over test days but was not affected by SPC treatment
On the pre-probe test, a main effect of day F(3, 69) = 17.02, p < 0.001, but no effect of treatment nor any interaction between treatment and day, were found. Effect of day was such that animals spent more time in the previous target quadrant on day 5 (p = 0.002) compared with day 2, on day 3 compared to day 4 (p < 0.001) and on day 5 compared to day 4 (p < 0.001), indicating formation of long-term memory in both treatment groups (Fig. 8B).

DISCUSSION
We have previously shown that the endogenous peptide AF-16 apart from its antisecretory and anti-inflammatory effects also modulates GABAergic transmission in the rat hippocampus in such a way that it produces disinhibition of CA1 pyramidal neurons (Kim et al., 2005;Strandberg et al., 2014). Such disinhibition could potentially alter the threshold for induction of, and/or the amount of LTP (Wigstrom and Gustafsson, 1983) and therefore potentially also enhance hippocampus dependent learning (Nicoll, 2017;Mansvelder et al., 2019). In Fig. 2. Synaptic properties of CA3-CA1 synapses in rats treated with specially processed cereals (SPC). (A) Box-and-whiskers plots depicting input-output relationship measured as the slope of the linear regression of fEPSP magnitudes plotted against presynaptic volley magnitudes in hippocampal slices from SPC-fed rats (n = 7) and controls (n = 9). Examples of fEPSPs elicited at different stimulations strengths are shown at the top (average of 12 records at each stimulation strength), scale bars represent 300 mV and 10 ms, and below are the corresponding input-output plots of fEPSP slope (mV/ms) plotted against presynaptic volley slope (mV/ms) and the linear regression curve depicted as a dotted line. (B) Box-andwhiskers plots depicting paired-pulse ratio (PPR) in hippocampal slices from SPC-fed rats (n = 7) and controls (n = 9). Examples of fEPSPs elicited at 50 ms interstimulus interval (average of 12 records) are shown at the top, scale bars represent 300 mV and 20 ms. n.s. not significant.
this study we investigated the effects of AF-16 on LTP in CA3-CA1 synapses and the effect of enhancing endogenous production of AF on spatial learning. We found that AF enhances LTP at hippocampal CA3-CA1 synapses, promotes spatial learning and enhances spatial short-term memory, but had no effect on spatial long-term memory.
AF applied in vitro, as well as when upregulated endogenously by SPC-feeding, enhanced both the small and largely decaying LTP after a weak tetanization as well as LTP after a strong tetanization.
These results demonstrate that AF facilitates the induction of LTP and enhances the magnitude of LTP. Such an enhancement of LTP is fully consistent with a disinhibitory effect of AF (Wigstrom and Gustafsson, 1983;Hanse and Gustafsson, 1992). We obtained direct support for this conclusion by showing that AF did not facilitate LTP in the presence of the GABA A receptor blocker PTX. Moreover, we have previously shown that AF causes disinhibition of CA1 pyramidal cells (Kim et al., 2005). Intriguingly, this disinhibition is likely mediated by an AF-induced upregulation of tonic GABAergic signaling in inhibitory interneurons (Strandberg et al., 2014). Thus, we propose a scenario for facilitating LTP by an up-regulation of tonic GABAergic signaling on inhibitory interneurons. Such a scenario has previously been described in the amygdala where knocking out the GABA A d subunit, which mediates tonic inhibition of interneurons in the lateral amygdala, decreased disinhibition onto projection neurons and thereby inhibited LTP in these neurons, and by doing so impaired fear learning (Liu et al., 2017). Thus, our results from the hippocampus add to the concept that tonic inhibition of inhibitory interneurons provides a powerful modulatory mechanism regulating learning. The cellular mechanisms by which AF promotes the increased tonic GABAergic signaling specifically on inhibitory interneurons are not known. It is, however, interesting that this facilitation of LTP can be induced solely by a specific combination of carbohydrates and amino acids. Indeed, we found similar facilitation of LTP in SPC-fed animals as with acute treatment of hippocampal slices with the AF peptide. Based on previous studies (Rapallino et al., 2003;Kim et   Field recordings in the CA1 stratum radiatum in rat hippocampal slices from SPC-fed rats (n = 10) and controls (n = 10). Stimulation protocol was the same as described in Fig. 1. Values are plotted as means ± SEM. Insets in the top show representative traces (average of 20 records) for time points indicated in the graph, scale bars represent 300 mV and 10 ms. (B) Box-and-whiskers plots depicting the amount of LTP 30 minutes after the weak tetanization and the strong tetanization in slices from SPC-fed rats (black boxes) and controls (grey boxes) respectively. *p < 0.05. Fig. 4. LTP induced in slices treated with AF-16 in the presence of the GABAA-receptor antagonist picrotoxin (PTX). (A) Field recordings in the CA1 stratum radiatum in rat hippocampal slices treated with AF-16 (n = 11) and controls (n = 10). Stimulation protocol was the same as described in Fig. 1. Values are plotted as means ± SEM. Insets in the top show representative traces (average of 20 records) for time points indicated in the graph, scale bars represent 300 mV and 10 ms. (B) Box-andwhiskers plots depicting the amount of LTP 30 minutes after the weak tetanization and the strong tetanization in slices treated with AF-16 (black boxes) and control slices (grey boxes) respectively. n. s. not significant. of AF of 0.5 mg/ml (284 nM) for the treatment of the hippocampal slices. Also based on previous studies we used food containing 5% SPC to induce endogenous upregulation of AF (Johansson et al., 2013). There are, however, no studies of the exact plasma concentrations of AF induced by 5% SPC. Despite this uncertainty regarding the in vivo concentration of AF in SPC-fed animals we propose, based on the very good match between the results from SPC-fed animals and from slices treated with exogenous AF, that both these disparate manners of increasing AF produce maximal facilitating effect of AF on LTP.
We have measured the amount of LTP 30 minutes after its induction, which might seem a rather early time given that LTP could last for, at least, hours in vitro. Thus, our conclusion that AF facilitates the induction of LTP (via disinhibition) cannot be extended to what sometimes is referred to as late LTP. Different induction protocols are typically used to induce LTP of different durations. For example, repeated strong tetanizations are often used to induced long-lasting LTP, whereas brief weaker tetanization is used to induce more transient LTP. In this study, we used one such weak induction and one stronger induction of LTP, both of which were facilitated by AF.
When endogenously upregulating AF and thereby enhancing LTP in the hippocampus one might expect alterations in long-term synaptic strength. An increased synaptic strength would give a greater synaptic response upon a given presynaptic activation and this should increase the input-output relationship. We did not, however, observe any significant difference in input-output relationship between SPC-fed rats and controls. This could be because of the sample size being too small and that we were not able to detect a small difference, but maybe more likely due to a homeostatic synaptic scaling, a process in where the   6. Spatial learning Day 1 was enhanced by SPC. Day 1 swim latency (in seconds), during swim 1 to 10 (bars represent ± SEM), to find platform in the MWM, effect of test day and treatment. Repeated measures two-way ANOVA identified significant effect of swim number ***p < 0.001 and treatment *p 0.05, pair wise comparisons identified significant differences between treatment groups on swim number 4 **p = 0.01 and swim number 8 **p = 0.003. Fig. 7. Spatial learning Day 4 was enhanced by SPC. Day 4 swim latency (in seconds), during swim 1 to 10 (bars represent ± SEM), to find platform in the MWM, effect of test day and treatment. Repeated measures two-way ANOVA identified significant effect of swim number ***p < 0.001 and treatment *p = 0.03, pair wise comparisons identified significant differences between treatment groups on swim number 1 *p = 0.03. whole synapse population is downscaled during sleep to ''make room" for synaptic plasticity to proceed day after day (Tononi and Cirelli, 2020).
Enhancing LTP in the hippocampus should facilitate learning and memory and we found that SPC treated rats indeed displayed a more efficient learning in the MWM. As such, all animals did improve their search strategy to find the platform over days (Fig. 5B), and there was a significant effect of treatment such that SPC treated animals demonstrated an enhanced learning on days 1 and 4, i.e. they had a shorter mean time to find the platform, see Figs. 6 and 7. The reason SPC treated animals located the platform faster than control animals during the initial swims on day 4, is hypothesized to be due to having acquired a more efficient search strategy, rather than remembering the previous platform position, as the platform is altered between days (Fig. 7). Now, since rats do learn to navigate a stationary platform quite readily, the platform positions were altered between days in order to not hit a floor effect for learning after the first acquisition day. This probably gave way for an enhancement in spatial learning by SPC in otherwise healthy rats, compared to control animals. In addition, all animals did improve their performance in the post-and pre-probe tests, i.e. an index of short-and long-term memory respectively, such that the time they spent in the target quadrant improved over days. However, there was only a treatment effect on post-probe on day 1, such that SPC treated animals spent a significantly longer time in the target quadrant, indicating enhanced short-term memory, compared to control animals (Fig. 8A). Enhancement of short-term memory in SPC treated animals would be expected as they also had an enhanced learning curve on day 1. More surprising was the lack of treatment effect on the pre-probe, i.e. long-term memory task.
A study investigating the effect of AF-16 (given intranasally) on brain edema and cognitive function after diffuse traumatic brain injury, in rats, found improved performance in the MWM in rats with brain injury that had been given AF-16 compared to rats with brain injury given a scrambled peptide (Clausen et al., 2017). However, they did not see any difference in control rats without brain injury given AF-16 compared to a scrambled peptide. This discrepancy with our findings could possibly be explained by differences in the MWM protocols as training occurred with the platform in a constant position during the four training days and only one post-probe trial, three days after the last training day. Thus, there are several major differences in the MWM paradigms between our and the Clausen et al. (2017) study. In the present study, platform position was altered between days in order to make the task more cognitively demanding. Healthy rats quickly learn to navigate spatially in the MWM and thus it is unlikely that improving cognitive functioning by increasing AF-16 would have been detected. As such, in the paper by Clausen et al. (2017) controls rats, as well as AF-16only treated rats, improved their spatial learning equally and substantially between day 1 (mean latency to platform around 50 seconds (S)) and day 2 (mean latency around 15 s), and after that, time to find platform was more or less stationary around 10 s day 3 and day 4, thus there was not much room for improvement of AF-16 on spatial learning. Also, the probe test (assessing long term memory) was carried out 72 h after the last training session in Clausen et al.'s (2017). study. In our study, probe testing, indexing long term memory, was carried out in the morning, i.e. within 24 h, after training session. However, neither study found an effect of AF-16 on long-term spatial memory in healthy animals. Given that AF-16 in fact seems to enhance LTP in the CA1 hippocampal region, as found herein, it was hypothesized that SPC treatment would have enhanced long term spatial memory in the MWM. However, and again, healthy rats seem to acquire a strategy for remembering platform position from the previous day, as we did find a significant improvement in time spent in target quadrant on the pre-probe test Day 4 compared to all other test days, see Fig. 8B. Thus, it is possible that a ceiling effect was reached in healthy control rats, why no such improvement would have been detected had there been one. It is also possible that our protocol is not optimal for assessing long-term spatial memory, or there is no such effect of AF-16 on spatial memory.
The disinhibitory effect of AF described here, mediated by decreasing interneuron excitability through During the post-probe test, a significant effect of day ***p < 0.001, but no effect of treatment, but a significant interaction between day and treatment p = 0.03, was found. Pair wise comparisons found effect of treatment on Day 1 *p = 0.03. (B) During the pre-probe test, a significant effect of day ***p < 0.001, but no effect of treatment was found.
an interneuron-specific upregulation of tonic GABAergic inhibition (Kim et al., 2005;Strandberg et al., 2014), adds another mechanistic scenario to previously described mechanisms of disinhibition in the hippocampus which include various examples of inhibition of interneurons by other interneurons in the cortical network (Letzkus et al., 2015;Artinian and Lacaille, 2018). This AF-mediated disinhibition, which enhanced hippocampal LTP and learning as well, can readily be achieved by adding SPC to the food. This is quite intriguing since SPC-FlakesÒ actually are available for humans as a ''Food for specific medical purposes" and AF-16 has been clinically tested according to the regulatives of a phase I study and might develop into a certified medical drug in the future. Whether enhancing AF activity in humans have any effect on LTP and learning is however unclear, but it could potentially be beneficial in neuro-rehabilitation and/or neurodegenerative diseases.