Effects of essential amino acid deficiency: down-regulation of KCC2 and the GABAA receptor; disinhibition in the anterior piriform cortex


  • James W. Sharp,

    1. Department of Anatomy, Physiology and Cell Biology, School of Veterinary Medicine, University of California, Davis, California, USA
    Search for more papers by this author
  • Catherine M. Ross-Inta,

    1. Department of Anatomy, Physiology and Cell Biology, School of Veterinary Medicine, University of California, Davis, California, USA
    Current affiliation:
    1. Catherine M. Ross-Inta, Department of Comparative Pathology, School of Veterinary Medicine, UC Davis, Davis CA, USA
    Search for more papers by this author
  • Irène Baccelli,

    1. Department of Anatomy, Physiology and Cell Biology, School of Veterinary Medicine, University of California, Davis, California, USA
    Current affiliation:
    1. Irène Baccelli, HI-STEM - Heidelberg Institute for Stem Cell Technology, Experimental Medicine Gemeinnützige GmbH (HRB 705161), German Cancer Research (Deutsches Krebsforschungszentrum, DKFZ), Im Neuenheimer Feld 280 69120 Heidelberg, Germany
    Search for more papers by this author
  • John A. Payne,

    1. Physiology and Membrane Biology, School of Medicine, University of California, Davis, California, USA
    Search for more papers by this author
  • John B. Rudell,

    1. Department of Anatomy, Physiology and Cell Biology, School of Veterinary Medicine, University of California, Davis, California, USA
    Search for more papers by this author
  • Dorothy W. Gietzen

    Corresponding author
    1. Department of Anatomy, Physiology and Cell Biology, School of Veterinary Medicine, University of California, Davis, California, USA
    • Address correspondence and reprint requests to Dorothy W. Gietzen, Department of Anatomy, Physiology and Cell Biology, School of Veterinary Medicine, University of California, Davis, One Shields Ave, Davis, CA-95616, USA. E-mail: dwgietzen@ucdavis.edu

    Search for more papers by this author


The anterior piriform cortex (APC) is activated by, and is the brain area most sensitive to, essential (indispensable) amino acid (IAA) deficiency. The APC is required for the rapid (20 min) behavioral rejection of IAA deficient diets and increased foraging, both crucial adaptive functions supporting IAA homeostasis in omnivores. The biochemical mechanisms signaling IAA deficiency in the APC block initiation of translation in protein synthesis via uncharged tRNA and the general amino acid control kinase, general control nonderepressing kinase 2. Yet, how inhibition of protein synthesis activates the APC is unknown. The neuronal K+Cl cotransporter, neural potassium chloride co-transporter (KCC2), and GABAA receptors are essential inhibitory elements in the APC with short plasmalemmal half-lives that maintain control in this highly excitable circuitry. After a single IAA deficient meal both proteins were reduced (vs. basal diet controls) in western blots of APC (but not neocortex or cerebellum) and in immunohistochemistry of APC. Furthermore, electrophysiological analyses support loss of inhibitory elements such as the GABAA receptor in this model. As the crucial inhibitory function of the GABAA receptor depends on KCC2 and the Cl transmembrane gradient it establishes, these results suggest that loss of such inhibitory elements contributes to disinhibition of the APC in IAA deficiency.


The circuitry of the anterior piriform cortex (APC) is finely balanced between excitatory (glutamate, +) and inhibitory (GABA, −) transmission. GABAA receptors use Cl, requiring the neural potassium chloride co-transporter (KCC2). Both are rapidly turning-over proteins, dependent on protein synthesis for repletion. In IAA (indispensable amino acid) deficiency, within 20 min, blockade of protein synthesis prevents restoration of these inhibitors; they are diminished; disinhibition ensues. GCN2 = general control non-derepressing kinase 2, eIF2α = α-subunit of the eukaryotic initiation factor 2.

Abbreviations used

artificial CSF replete with essential amino acids


ACSFcomp with 0 threonine


anterior piriform cortex


area of the APC medial to the lateral olfactory tract


ventral-rostral area of the APC


activating transcription factor 4


averaged evoked potential


basal control diet


basal threonine-corrected diet


threonine basal devoid diet


calcium/calmodulin-dependent protein kinase type II




α-subunit of the eukaryotic initiation factor 2


excitatory postsynaptic potential


general control nonderepressing kinase 2




indispensable (essential) amino acid


neural potassium chloride co-transporter


lateral olfactory tract




N-methyl d glucamine


phosphate buffered saline with added goat serum


threonine basal complete diet

The chemosensor for indispensable (essential) amino acid (IAA) deficiency is found in the anterior piriform cortex (APC) (Leung and Rogers 1971; Gietzen and Rogers 2006). Animals reject an IAA deficient diet within 20 min (Koehnle et al. 2003) of exposure. Of the many brain areas with roles in eating behavior (Berthoud 2002), only the APC has been shown to be both necessary (Leung and Rogers 1971; Noda and Chikamori 1976) and sufficient (Rudell et al. 2011) for the detection of dietary IAA deficiency. Although the APC is part of the olfactory cortical system, as reviewed by Neville and Haberly (2004), the biochemical activation of the APC (but importantly, not taste or smell) is considered the primary sensory event (Gietzen et al. 2007; Gietzen and Aja 2012). Within the APC, a decrease in a single IAA (Koehnle et al. 2004) triggers the conserved general control non-derepressing kinase 2 (GCN2) signaling cascade (Hao et al. 2005), increasing phosphorylation of eukaryotic initiation factor 2 alpha (eIF2α-P), which blocks further initiation of mRNA translation and global protein synthesis (Wek et al. 2006). The neurons of the APC show molecular potentiation shortly after receiving an IAA deficient meal (Sharp et al. 2002, 2004); electrophysiological activation is seen in vitro in isolated APC slices exposed to IAA deficient media (Rudell et al. 2011). Still, how the APC converts blockade of the initiation of protein translation into potentiation of its neural output has not been determined. Reduced inhibition, because of loss of inhibitory proteins, should potentiate the APC circuits and provide an attractive basis for understanding this transduction.

The circuitry of the APC is under the control of GABAergic inhibitory interneurons (Neville and Haberly 2004). Blockade of GABAA receptors with bicuculline injected into the APC inhibits the behavioral rejection of an IAA deficient diet (Truong et al. 2002). This shows that GABA is involved in the adaptive feeding response to IAA deficient diets that is mediated by the APC. Fast GABAergic inhibition is mediated via GABAA receptors that are anion channels permeable primarily to chloride and hence, dependent upon the chloride electrochemical gradient largely maintained by the neuronal K+–Cl cotransporter (KCC2) (Payne et al. 1996, 2003; Rivera et al. 1999). Remarkably, shifts from hyperpolarizing to depolarizing (and often excitatory) GABAA-mediated responses have been observed in both physiological and pathophysiological conditions, including tetanic stimulation, neuronal trauma, and axotomy (Nabekura et al. 2002; Payne et al. 2003; Tominaga and Tominaga 2010). Such polarity shifts in GABAA-mediated responses appear to be largely because of alterations in functional KCC2 expression, leading to altered intracellular [Cl] and hence changes in the chloride electrochemical gradient. Plasmalemmal KCC2 has a rapid protein turnover rate (~ 10–20 min) that permits changes in KCC2 expression to have a potential role in various types of neuronal plasticity (Rivera et al. 2004). Given the role of GABAergic inhibition in the APC noted above, and its dynamic plasticity (Vithlani et al. 2011) we hypothesized that potentiation of the APC in response to IAA depletion might involve reduction of inhibitory elements, such as KCC2 and the GABAA receptor. For this report, we evaluated protein levels of KCC2 and the GABAA receptor as well as electrophysiological effects of GABAA inhibition.

Materials and methods

Animal use and care were according to National Institutes of Health guidelines and approved by the local Animal Use and Care Committee; male albino rats (Simonsen Laboratories, Gilroy, CA, USA or Harlan Laboratories, Hayward, CA, USA) weighed between 200 and 225 g. ARRIVE guidelines have been reviewed and observed. Housing, vivarium conditions and baseline feeding protocols were as described previously (Koehnle et al. 2004).

Experimental diets

The diets have been described in detail (Sharp et al. 2006), and were prepared in our facility. We used a control threonine basal diet [basal control diet (BAS)], an IAA deficient diet, threonine basal devoid diet (BTD) and another control diet: the threonine basal corrected diet [basal threonine-corrected diet (BTC)], which had 4 g of threonine/kg of diet with a corresponding subtraction of carbohydrate, rendering it fully replete for all IAAs. We have seen repeatedly that the taste and other sensory characteristics of the deficient and corrected diets are indistinguishable to the rat during the pre-absorptive period. The BAS was fed for at least a week prior to each experiment, when the test diet (BTD or BTC) was presented for the first time (a control group was given BAS) and tissues were taken at the noted times after this first diet presentation. For in vitro electrophysiology, rats were pre-fed a 6% casein diet, which is supplemented with methionine, and thus first limiting in threonine.

Experimental Procedures for western blots and immunohistochemistry have been published previously (Sharp et al. 2006), with the exceptions detailed below. Methods for electrophysiology have been described in detail (Rudell et al. 2011). Details of these methods may be found in Supplementary Material.

Western blots

The brains were taken 20, 40 min, 2 h, or 3 h after a single meal of the appropriate experimental diet and dissected immediately (n = 12 rats per diet group). We used two areas of the ventral APC: APCvr, as defined by Ekstrand et al. (2001) and an area just deep to the lateral olfactory tract (LOT) that we termed APClot; for anatomical details see (Sharp et al. 2006; Rudell et al. 2011); tissues were taken within the active area of the APC, between 2.2 and 3.7 mm rostral to Bregma. Control brain areas were: cerebellum (CB) for non-cortical neural tissue, neocortex (NC) a well studied site for KCC2 studies (Deisz et al. 2011) and hippocampus (Hip; also known as archicortex) that shows GABAergic inhibition of its pyramidal cells (Rivera et al. 2004). Thus, three cortical types, neo- (NC), paleo- (APC), and archi- (Hip) were included.


Primary antibodies used for immunohistochemistry were the mouse monoclonal anti-GABAA receptor antibody to the alpha-1 subunit, β 2,3 chain (Chemicon, Temecula, CA, USA) and rabbit polyclonal KCC2 antibodies, which we have described in detail (Williams et al. 1999). The GABAA receptor antibody was diluted at 1 : 500 and the KCC2 antibody was diluted 1 : 2000 into 2% phosphate buffered saline with added goat serum. Immunofluorescence-based immunohistochemistry was employed to label the GABAA receptor and KCC2 as in Sharp et al. (2006) – see Supplementary Material.

In vitro electrophysiology in the APC slice

Excitatory post-synaptic potential (EPSP) and averaged field-evoked potentials (AEP) were recorded in the in vitro APC slice preparation (total studies n = 19) using standard extracellular electrophysiological techniques as described in detail previously (Rudell et al. 2011), see also Supplementary Material.

After a stable baseline recording was established with the slice bathed in an IAA-replete artificial CSF medium (ACSFcomp; Table S1) for 60–90 min, the experimental medium was introduced. The basic IAA deficient experiment was: ACSFcomp (control) versus ACSF containing all of the non-IAA and IAA except for threonine (ACSFdev). Recording continued for 60–90 min, at which point the medium was returned to ACSFcomp and responses recorded to allow the EPSP response to return to near baseline levels. In the positive control response to ACSFdev, demonstrating potentiation by IAA depletion, the increase in AEP amplitude is 110–120% of baseline, variability in this positive control condition is ~2%, given in previous reports (Rudell et al. 2011).

To study potential interactions between the devoid medium and the GABAA receptor, and because there was considerable receptor protein remaining after devoid diet ingestion, in both western and immunohistochemistry results (see above), bicuculline methiodide (Research Biochemicals Inc., Natick, MA, USA) was used at 30 μm, both in the control (n = 2) and devoid media (n = 4).

Because bicuculline can also inhibit the afterhyperpolarization mediated by potassium (Khawaled et al. 1999), additional media changes included removal of potassium (n = 2), and substituting N-methyl d glucamine (NMG; 3 mM; Sigma-Aldrich, St. Louis MO, USA) (n = 4).


Values for western blot proteins are given as band volume densities, determined by quantitative analysis of digitized western blot images using ImageQuant 5.1 (Molecular Dynamics, Piscataway, NJ, USA). Band volume density was the integrated intensity of all pixels excluding the background. Values were normalized to protein levels in the band of interest, determined using Coomassie Blue. Values for electrophysiological data are presented for the test responses as percent of the baseline potential, means ± SE. Data were analyzed by anova with F and P values calculated using Microsoft Office Excel 2003 (Redman, WA, USA) for the effects of diet (or medium) and brain region. Significance was set at < 0.05. After a significant F was found in the overall anova, pre-planned comparisons among means were determined by post hoc tests. If more than two means were compared, Tukey's multiple comparisons test was utilized. Where just two groups were compared a Student's t-test was used.


KCC2 protein levels in western blots of APC membrane preparations

In tissue taken 20 min after onset of a single IAA deficient (BTD) meal KCC2 protein expression was reduced in APClot (< 0.005) and APCvr (< 0.037) relative to control rats that ate the BAS diet (Fig. 1a). This reduction continued to be significant (< 0.005 and = 0.015, respectively) at 40 min (Fig. 1b), and also (= 0.002 and p = 0.013, respectively) at 2 h after onset of the meal, compared to either BAS or BTC diet groups (Fig. 1c). However, by 3 h (see 3 h data points in Fig. 1d) KCC2 expression in membrane preparations from the BTD-fed rats had returned almost to the levels of the BAS control group (> 0.05).

Figure 1.

Expression of the neural potassium chloride co-transporter (KCC2) in membrane preparations from brain areas. Western blot components of tissue taken at the indicated intervals after rats were offered the experimental diets, either Basal or threonine basal devoid diet (BTD). Areas are: from the top, anterior piriform cortex (APC) medial to the lateral olfactory tract (APClot), the ventral-rostral portion of the APC (APCvr), control areas: the CB, the hippocampus (Hip), and the primary motor area of the neocortex (NC). Western blot lanes show KCC2 expression, from individual rats fed the diet indicated at the bottom of the lane, for each of the above brain regions. Below the blots for each time point are histograms; bars indicate means ± SE; significance values are given in the text, approximated here as **< 0.001, *< 0.05. (a) Western blot components showing expression of KCC2 in tissue taken at 20 min. (b) Same as (a) for tissue taken at 40 min. (c) Expression of KCC2 in tissues taken 2 h after the onset of feeding the Basal, threonine devoid (BTD) or corrected (BTC) diets. Blots for 3 h are not shown; there were no differences due to diet in any of the brain areas at that time. In Fig. 1d KCC2 expression is given as the % of values for BTD/basal control diet (Basal) for the five brain areas at 20 and 40 min, 2 and 3 h. Significance of differences among brain areas is indicated as for histograms in Fig. 1a–c.

Of the five brain areas examined, only the two APC regions showed a significant (see significance levels given above) reduction of KCC2 protein expression after eating BTD (Fig. 1). The control brain areas, CB, Hip, and primary motor NC, were largely unaffected, as can be seen in the lower three bands in Fig. 1a and b, as well as the NC in Fig. 1c. Also, the APC appears to have low ambient levels of KCC2 expression relative to the other brain areas examined (compare APC bands, top and next, with CB, Hip, and NC below, as indicated at the left of each segment in Fig. 1). While the same amount of total protein from each brain area was electrophoresed in western blots, band volume densities for KCC2 averaged for each time point in the APClot and the APCvr of rats fed the control BAS diet, were approximately 40% of those measured in CB and Hip, and 42% of NC expression levels from the same rats for the first 2 h (Fig. 1d). This is consistent with the relative scarcity of inhibitory elements in the APCvr reported by Ekstrand et al. (2001). Two hours after a meal of the corrected diet, BTC, KCC2 expression was unchanged from basal control levels in all brain regions examined (compare Basal with BTC in Fig. 1c).

GABAA receptor protein in membranes of the APC

After feeding the test diets as above, expression levels of the alpha-1, β2,3 chain subunits of the GABAA receptor were examined in the APClot, APCvr, CB, and NC. Western blots of cell membrane preparations showed that GABAA receptor expression was reduced in the APClot and APCvr of rats after they ate BTD relative to control rats that consumed BAS (Fig. 2a–c). Twenty minutes after introduction of BTD, GABAA receptor expression was significantly (= 0.049) reduced in the APClot, but not in the APCvr (Fig. 2a). Thereafter, GABAA receptor expression was reduced at 40 min (= 0.018 and 0.011, respectively, Fig. 2b) and remained at a reduced level 3 h after a BTD meal in both APClot and APCvr (= 0.049 and 0.016, respectively, vs. BAS, Fig. 2c). Immunoadsorption controls for the GABAA receptor can be seen in Fig. 2d.

Figure 2.

Expression of the GABAA receptor alpha-1 subunit in membrane preparations from brain areas. Tissues were anterior piriform cortex (APClot), APCvr, CB, and neocortex (NC). Other conditions are the same as in Fig. 1; *< 0.05 (a) Western blot components in tissues taken 20 min after introduction of Basal or threonine basal devoid diet (BTD) diets. (b) Expression of the GABAA receptor in tissues taken at 40 min. (c) Expression of the GABAA receptor in tissues taken at 3 h. (d) Immunoadsorption controls. Lanes labeled 1X, 5X, and 10X ImmunoAbs, indicate one-, five- or ten-fold molar excess of the immunizing synthetic peptide added to the gel, alternating with the GABAA receptor protein.

Similar to the results for expression of KCC2 in tissues other than the APC, GABAA receptor expression was unaffected by BTD in the control brain regions, CB or NC (> 0.05) (Fig. 2a–c). In control tissue, parallel to our findings with KCC2, ambient expression of the GABAA receptor was significantly (< 0.05) less in APC than in CB or NC (Fig. 2a–c). Also western blot band volume densities in tissue from rats fed only BAS and averaged for the three time points in the APClot and APCvr were 48% of that of analogous averages for CB and 54% for NC, again showing the paucity of GABA's inhibitory capability in APC, as expected (Ekstrand et al. 2001).

To illustrate that not all proteins are down-regulated in the APC after ingestion of an IAA deficient meal, in Figure S1 we show western bands for the proteins eIF2α, eIF2α-P and double-stranded RNA-activated protein kinase, another eIF2α kinase (Gietzen et al. 2004). Also shown are calcium/calmodulin-dependent protein kinase type II, which is phosphorylated rather than degraded in IAA deficiency (Sharp et al. 2004), and ATF4, which is responsive to general control non-derepressing kinase 2 activation and has a short half-life (Novoa et al. 2001). These bands, each taken from APC tissue collected at 20 min after introduction of a basal (B) or IAA deficient (D) meal, show the phosphorylation of eIF2α and the stability of the four proteins during IAA deficiency. See Supplementary Material, Figure S1

Localization of KCC2 and the GABAA receptor

To evaluate the tissue distribution of KCC2 in the APC, after quantitation of the proteins in the western blots, confocal images were made using previously published techniques (Sharp et al. 2006). In tissue from rats fed BAS diet, KCC2 expression in the APC appeared to be most abundant in APC layer I. There appeared to be a gradient of KCC2 expression with the highest levels near the LOT and the pial surface, and progressively less expression in the deeper tissue (Fig. 3a).

Figure 3.

Confocal microscopy images of neural potassium chloride co-transporter (KCC2) in the anterior piriform cortex (APC). Plates b and d are magnifications of plates a and c, respectively, and correspond to the areas outlined in white. White horizontal lines = 100 μm. APC layers are numbered I, II, and III and follow the designations of Haberly and Price (Haberly and Price 1978a). (a) KCC2 expression (white) can be seen in APC layer I, progressively less in layers II and III, and none in the LOT. (b) Higher magnification of (a). KCC2 appears to be concentrated at the plasma membrane (large white arrow) of large neuronal somata and in apical dendrites, which extend from layer II through layer I toward the LOT (dark arrow). (c, d) 20 min after a single threonine basal devoid diet (BTD) meal reduction of KCC2 expression is evident. Decreased KCC2 expression is particularly evident in APC layer I (Compare layer I in a and c).

Inspection of the confocal microscope images (Fig. 3c and d) shows that 20 min following a BTD meal KCC2 protein expression was clearly reduced in all three APC layers, consistent with the western blot data. Decreased KCC2 immunoreactivity was most apparent in APC layer I.

GABAA receptor expression in the APC of rats, 20 min after eating the control BAS diet, also appeared to be most abundant in APC layers I and II (Fig. 4a, b). GABAA receptor expression was visually less abundant in APC cortical layer III. Visual examination of confocal images showed that GABAA receptor expression was decreased 20 min after a BTD meal (Fig. 4c, d), again consistent with the quantitation seen in the western blot data. Diminished GABAA receptor expression was most evident in APC layer I.

Figure 4.

Confocal microscopy images of the GABAA receptor in anterior piriform cortex (APC). Conditions are the same as for Fig. 3. (a, b) Labeling is seen with the highest levels in layers I and II, with much less labeling in layer III and deeper structures, and none in the LOT. (c, d) Twenty minutes after introduction of threonine basal devoid diet (BTD), decreased expression of the GABAA receptor is particularly evident in APC layer I, although residual labeling persisted, as discussed.

Electrophysiology in the APC slice

Results of electrophysiology performed in the in vitro APC slice preparation can be seen in Fig 5. When comparing the responses to ACSFdev we normalized each trial to its own control determined in ACSFComp. In the presence of bicuculline, the increased amplitude of the AEP was slightly higher (about 4% over that seen in the control), but did not differ from the devoid response in the control medium (usually near 20%). To control for possible effects of bicuculline blocking the neural afterhyperpolarization (Khawaled et al. 1999), which is because of potassium, we removed potassium from the medium. Zero potassium with or without NMG resulted in a clear loss of the response to ACSFdev medium (F3,15 = 4.98; = 0.013).

Figure 5.

Results of electrophysiological recordings using the in vitro anterior piriform cortex (APC) slice preparation of (Rudell et al. 2011) with modifications of the media as described in the text. Experimental treatments were: control (showing the expected ~20% increase in response to devoid [Rudell et al. 2011]), or blockade of the GABAA receptor with bicuculline, absence or substitution of potassium (K+) with a non-permeant substance N-methyl d glucamine (NMG), (a) Effect of changing the medium to devoid (indicated by horizontal line) in the presence of bicuculline. Data are amplitude (μV) of the evoked potential over time (waveforms were recorded every 3 min) (b) comparisons of ACSFcomp (Comp) and ACSFdev (Dev) in control or bicuculline-containing media, means ± SE. There were no significant differences between control and bicuculline in this study. (c) Effects of changing from control media to potassium free media (at wave form five) and from K-free comp to K-free devoid (at about wave form 11). Axes are amplitude versus waveform as in (a) Clearly the increase usually seen with the devoid medium was not seen in K-free medium. (d) Comparison of the mean ± SE for μV amplitudes in ACSFcomp (Comp) versus ACSFcomp-K (Comp-K) or ACSFdev-K (Dev-K). The stars ‘*’ indicate significance (< 0.05).


The initial responses to IAA deficiency in the APC lead to increased phosphorylation of eIF2α (Hao et al. 2005), which blocks general protein synthesis. This should prevent restoration of proteins subject to rapid turnover, resulting in a rapid and sustained reduction of inhibitory function, as we saw first with expression of KCC2 and then GABAA receptor protein in our western blots. Such a reduction of inhibitory capacity suggests a straightforward explanation for disinhibition and potentiation of the APC in its responses to IAA depletion.

The APC is both necessary and sufficient as the chemosensor for IAA deficiency

The APC, a primary olfactory cortex, is a paleocortex with a three layer structure (Haberly and Price 1978b; Fontanini and Bower 2006); glutamatergic pyramidal cells provide the primary output (Jung et al. 1990). The APC is highly excitable and has long been known to show regular oscillatory electrical activity (Freeman 1957); stimuli arrive from the LOT with each breath (Fontanini and Bower 2006), as well as from intracortical fibers (Poo and Isaacson 2011). The APCvr has also been called the ‘Area Tempestas’ because of its high sensitivity to seizurogenic drugs such as GABA antagonists (Doherty et al. 2000; Ekstrand et al. 2001). GABA provides the main inhibitory control over this excitatory circuit (Manns et al. 2003).

While the APC has long been known to be necessary for the behavioral responses to IAA deficient diets (Leung and Rogers 1971), potentiation of the isolated APC slice and increased eIF2α-P in IAA deficiency have also been shown in the isolated brain slice preparation (Rudell et al. 2011), indicating that the APC is also sufficient for sensing IAA depletion. A relatively small number of APC neurons respond to IAA deficiency with robust co-labeling for both phosphorylated extracellular signal-related protein kinase 1/2 (ERK1/2-P) and eIF2α-P (Sharp et al. 2006); these few neurons are the presumptive chemosensory cells. Once IAA-deficiency is detected, in vitro molecular, neural (Sharp et al. 2002; Rudell et al. 2011), and in vivo electrophysiological (Hasan et al. 1998) potentiation spreads throughout the APC.

Loss of KCC2 and the GABAA receptor

Two factors may be critically important for the spread of potentiation in the APC. First, the unique cellular architecture of the APC appears to foster the spread of neuronal potentiation. The APC is characterized by a dense recurrent-collateral innervation from the large glutamatergic pyramidal neurons of APC layers II and III (Haberly 2001; Franks et al. 2011). Once activated, these neurons activate neighboring neurons, resulting in potentiation, which is kept under control by GABAergic inhibition. Thus, it is important that the APCvr is also characterized by very low numbers of GABAergic ‘cartridge’ endings on initial axonal segments (Ekstrand et al. 2001), so the ease of the spread of potentiation in the APC could result from the small number of inhibitory neuronal elements found there (Doherty et al. 2000; Ekstrand et al. 2001), and see Fig. 2 where one can compare basal levels of GABAA receptor protein in the APC versus control brain areas, CB and NC. Diminished expression of KCC2 and the GABAA receptor should further weaken inhibitory control over the spread of APC neuronal activity in response to acute dietary IAA deficiency.

The importance of KCC2 in control of the APC circuitry

Brain trauma, hypoxia-ischemia, hypoglycemia, and seizures are all associated with a dramatic loss of KCC2 protein (to 20–40% of control levels) e.g., (Payne et al. 2003; Rivera et al. 2004). This loss is rapid and associated with the calcium activated protease, calpain (Chung and Payne 2009; Puskarjov et al. 2012). We have previously shown that calcium is elevated in the APC with IAA deficiency (Sharp et al. 2004), so calpain should have the necessary calcium for activation in the APC. Also, during such pathological events GABAA receptors mediate a depolarizing, often excitatory, Cl efflux in the affected brain regions (Payne et al. 2003). This is accompanied by diminished ion gradients, including an increase in intracellular Cl and extracellular K+. There is good evidence that KCC2 protein is down-regulated in animal models after epileptiform activity, which leads to impaired neuronal Cl extrusion (Rivera et al. 2002, 2004). Optical imaging techniques demonstrate a two-phase intraneuronal Cl increase in hippocampal CA1b neurons due, in part, to impaired or down-regulated Cl efflux mechanisms largely mediated by KCC2 (Galeffi et al. 2004). Axonal injury in dorsal motor neurons is also accompanied by a large reduction in KCC2 transcript (Nabekura et al. 2002). It is significant that no change in sodium-potassium-chloride (NKCC1) co-transporter transcript or protein was observed in these studies, suggesting that both the increased intraneuronal concentration of Cl and the possible depolarizing action of GABAA receptors are primarily related to down-regulated KCC2 protein expression (Payne et al. 2003).

Down-regulation of KCC2 is dependent on neural activity and increased Ca2+ (Fiumelli et al. 2005); both are present in the APC, as shown by the following. First, increased CaMKII-P, indicating increased intracellular Ca2+, in the APC occurs at 20 min after onset of an IAA deficient meal (Sharp et al. 2004); we have also seen increased Ca2+ in IAA deficient slices in vitro (unpublished findings). This down-regulation is not because of internalization of plasmalemmal KCC2, because total cellular protein labeled for KCC2 is rapidly (15–20 min) lost because of activation of the calcium-activated protease, calpain (Chung and Payne 2009; Puskarjov et al. 2012). Second, increased neural activity occurs with each breath (Fontanini and Bower 2006). Thus, the conditions required for down-regulation of KCC2 are in place.

GABA and the inhibitory control of neural excitation in the APC

Decreased expression of the GABAA receptor subunit was less pronounced and occurred later than that of KCC2 in the present results, suggesting that the primary effect is likely at the level of KCC2, as reported previously (Payne et al. 2003; Deisz et al. 2011). As noted above, the loss of KCC2 has a profound effect on GABAergic inhibition. To assess this inhibitory circuitry further we looked at the effects of inhibiting the GABAA receptor. In electrophysiological studies using bicuculline, the classical GABAA receptor antagonist, the increased amplitude of the EPSP in response to IAA depletion was similar in control and bicuculline-containing media. This is consistent with diminished receptor levels in the devoid medium, because the primary neurons of the APC circuitry would be subjected to less GABAergic inhibition in both cases. Alternatively, the persistent observation of the devoid response in the presence of bicuculline could be because of bicuculline-insensitive GABA receptors that also are Cl ionophores, such as the GABAC receptor (Zhang et al. 2001), now termed the GABAAρ subunit (Alexander et al. 2011). This again supports our idea that the primary effect is at the level of KCC2.

Bicuculline can also inhibit the afterhyperpolarization mediated by potassium (Khawaled et al. 1999). However, when we eliminated potassium from the medium, the evoked potential was inhibited, rather than enhanced, as it was with bicuculline. So bicuculline was most likely acting selectively at the remaining GABAA receptors. Along with these observations, we have previously seen a down-regulation of mRNA for another GABAA receptor subunit (β-3) in APC tissue taken 2 h after introduction of an IAA-imbalanced diet (Truong et al. 2002). Also, as noted earlier, bicuculline injections into the APC blocked the anorectic response to the IAA deficient diet only (Truong et al. 2002), i.e., it was selective for the diet and did not cause malaise. Taken together, these findings show that both KCC2 and GABA's inhibitory roles in the normal APC are compromised in IAA deficiency.

Time course

In mechanistic studies, rapid plasmalemmal KCC2 degradation has been shown, within 15–20 min, because of activation of the calcium-activated protease, calpain (Chung and Payne 2009; Puskarjov et al. 2012) while KCC2 mRNA remains at control levels for at least 30 min (Rivera et al. 2004). The persistent loss of these proteins over 2–3 h is consistent with inhibition of translation in the presence of eIF2α-P (Gietzen et al. 2004; Hao et al. 2005; Wek et al. 2006). By the time one sees dephosphorylation of eIF2α by its constitutive phosphatase (Novoa et al. 2001) allowing the re-initiation of translation, i.e., between 2 and 3 h, KCC2 was restored in our western blots (Fig. 1d). In line with this timing, tRNA charging was restored, and even reversed in APC tissue taken 2 h after an IAA devoid meal (Magrum et al. 2002), suggesting that recovery might be because of provision of IAA substrates from protein degradation by this time.

These findings suggest a potential next link in the mechanism of how IAA deficiency excites the APC, leading to rejection of a deficient meal and foraging for a better food source, behavioral responses that are adaptive for survival (Gietzen and Aja 2012). However, excessive excitation can be harmful in the brain, if not kept under inhibitory control, as potentiation of the positive-feedback loop can occur in the highly excitable recurrent circuitry of the APC.

Implications for seizure disorder

GABA in the APC has long been associated with seizure disorders (Doherty et al. 2000). We have previously shown increased seizure susceptibility in rats fed IAA deficient diets, using the GABAA receptor antagonist, picrotoxin (Gietzen et al. 1996). Seizures related to IAA disproportion are known to be associated with genetic defects in amino acid metabolism, e.g., loss of control of branched-chain amino acid catabolism results in seizures (Harris et al. 2005) and defective production of serine, associated with low serine levels in the cerebrospinal fluid and plasma is linked to seizures (de Koning et al. 1999). Brain specimens surgically removed from epilepsy patients show involvement of GABA (Palma et al. 2006). Many epileptic patients have difficulty controlling their seizures on virtually any treatment protocol. Others, normally under control, have breakthrough-seizures of unknown etiology. IAA deficiency-mediated down-regulation of inhibitory elements, such as KCC2 and GABAA receptor proteins (present results), widespread molecular potentiation of the APC (Sharp et al. 2002, 2004, 2006), and the unique ability of the APC to generate seizures (Doherty et al. 2000) suggest that a single meal deficient in an IAA may increase the potential for seizures in susceptible patients.


Our results show that the protein expression of the KCC2 co-transporter and the GABAA receptor is reduced in the APC membrane preparations and in confocal micrographs of APC tissue within 40 min of ingestion of an IAA deficient meal. These losses are likely maintained in the absence of mRNA translation because eIF2α is phosphorylated, blocking initiation of translation in the APC under the same conditions (Gietzen et al. 2004; Hao et al. 2005). The loss of GABAA receptors was supported by electrophysiological results. Other inhibitory elements likely are be involved in this complex and important system, as we have reviewed previously (Gietzen et al. 1998; Gietzen and Magrum 2001), including any of the several GABAA subunits not studied here. Nonetheless, the present results provide clear examples of the loss of inhibition (disinhibition) that should contribute to the mechanisms of the potentiation of the APC in response to IAA deficiency (Gietzen et al. 1996; Hasan et al. 1998). Furthermore, these findings offer the novel and provocative suggestion that people who suffer from seizure disorders might be advised of a potential seizure risk from IAA-imbalanced meals.


This research was supported by a Research grant from the National Institutes of Health (NS043210 to D.W.G.). I.B. was an intern of the Institute National Research Agronomique, France. J.A.P. was supported by NIH NS03296 for the production of the KCC2 antibodies. D.W.G. has served as a consultant for Ajinomoto North America Inc., but has no other potential conflicts of interest. The remaining authors declare no conflicts of interest.

Author contributions

J. W. S. designed, performed, and interpreted experiments and wrote most of the manuscript. I. B. performed the experiments with the corrected diet. C.M.R-I. performed western blot experiments and contributed to writing and editing the manuscript. J.A.P. provided the KCC2 antibodies and contributed to editing the manuscript. J. B. R. performed the electrophysiological experiments, the associated statistical analysis and contributed to editing the manuscript. D.W.G. contributed to the design and interpretation of the experiments, and the writing and editing of the manuscript.