Knockout mice reveal a role for protein tyrosine phosphatase H1 in cognition
Behavioral and Brain Functions volume 4, Article number: 36 (2008)
The present study has investigated the protein tyrosine phosphatase H1 (PTPH1) expression pattern in mouse brain and its impact on CNS functions.
We have previously described a PTPH1-KO mouse, generated by replacing the PTP catalytic and the PDZ domain with a LacZ neomycin cassette. PTPH1 expression pattern was evaluated by LacZ staining in the brain and PTPH1-KO and WT mice (n = 10 per gender per genotype) were also behaviorally tested for CNS functions.
In CNS, PTPH1 is expressed during development and in adulthood and mainly localized in hippocampus, thalamus, cortex and cerebellum neurons. The behavioral tests performed on the PTPH1-KO mice showed an impact on working memory in male mice and an impaired learning performance at rotarod in females.
These results demonstrate for the first time a neuronal expression of PTPH1 and its functionality at the level of cognition.
Tyrosine phosphorylation plays an important role in several signaling pathways regulating cell growth, differentiation, cell cycle, apoptosis and neuronal functions [1, 2]. The phosphorylation/dephosphorylation balance is controlled by protein tyrosine kinases and phosphatases. PTPs can be distinguished into four classes: 1) classical PTPs that can be subdivided into transmembrane, receptor-like enzymes, and the intracellular, nonreceptor PTPs, 2) dual-specificity PTPs (Ser and Tyr phosphatases), 3) low molecular weight PTP and 4) the Asp-based PTPs (Tyr/Ser phosphatase activity) .
Classical PTPs have been reported to play a key role in neural functions, from development to cognitive function. For example, RPTPs such as PTPδ, PTPσ, LAR, and especially PTPRO, are important players in axonal growth and guidance during development . Studies on PTPσ-KO (RPTP) mice have shown involvement of this PTP in the regulation of the developing hypothalamo-pituitary axis [5, 6] and in the development of the CNS architecture . PTPBL-KO (non receptor like PTP-NRPTP) mice display impaired motor nerve repair in a model of sciatic nerve crush lesion  and PTPMEG (NRPTP) interacts with key intracellular players leading to the stimulation of the channel activity of NMDA receptors .
In the present study we focused our attention on a NRPTP, PTPH1, and on its possible role on neural functions. Indeed PTPH1 has been shown to be expressed in the CNS  but little is currently known on its potential impact on CNS functions. PTPH1 (also called PTPN3) belongs to a sub-family of non receptor cytosolic PTPs characterized by the presence of a FERM domain (band 4.1, ezrin, radixin, moesin) at its N-terminus, responsible for the interaction with transmembrane proteins and/or phospholipids in the cell membrane [11–13]. In addition PTPH1 has a PDZ domain in the central part responsible for the interaction with other proteins, whereas the single catalytic domain is located at the C-terminus.
PTPH1 activity has been involved in a variety of cellular functions including TCR-signaling [14–16], cell cycle regulation [11, 16, 17], endoplasmic reticulum assembly , cardiac sodium channel modulation  and TNFα converting enzyme inhibition .
Recently, our group has demonstrated that PTPH1 dephosphorylates GHR in vitro and in cellular assays  and results in an increase of body weight in the functional PTPH1-knockout (KO) mice via modulation of IGF1 secretion  thus demonstrating its in vivo relevance.
PTPH1 has been shown in the rat to be highly expressed in thalamic nuclei as well as various cortical areas . However, no information is currently available on its impact on CNS functions. To address this question we have further characterized our PTPH1-KO mice line through behavioral and anatomical approaches. PTPH1 expression and localization was evaluated by LacZ staining in the brain and a behavioral test battery evaluated PTPH1 loss on CNS functions such as locomotor activity (open field), anxiety-like behavior (open field and elevated plus maze), motor ability, coordination and learning (accelerating rotarod), spatial working memory (Y maze) and nociceptive sensitivity (hot plate).
PTPH1-KO and wild type littermates (F2 generation, 87.5% C57Bl/6 – 12.5% 129S6SvEv) aged 3–4 months were used for behavioral phenotyping. Mice were individually caged and maintained in a 12:12 hours light: dark cycle (lights on at 7 am) at 21 ± 1°C with food and water available ad libitum. Protection of animals used in the experiment was in accordance with Directive 86/609/EEC, enforced by the Italian D.L. No. 116 of January 27, 1992. Physical facilities and equipment for accommodation and care of animals were in accordance with the provisions of EEC Council Directive 86/609. Tail snips from mice were collected and genotyped as previously reported .
PTPH1 KO design
PTPH1-KO mice were generated using the Velocigene technology , as described in details elsewhere . Briefly a mouse BAC containing the PTPH1 gene was modified: an in-frame LacZ reporter sequence and a neomycin-selectable marker replaced exons 15 to 22 encoding for the PDZ and the catalytic domain of PTPH1. BAC electroporation into embryonic stem cells was performed. F1 heterozygous mice were bred to generate F2 PTPH1-KO mice. Line breeding and animal care were performed in Charles River Italy and France.
LacZ staining procedure and immunohistochemistry
PTPH1-KO and WT mice, 12 months old, n = 2, male and females, were sacrificed by ip overdose of thiopental (5%), perfused with paraformaldehyde 4%, then washed in PBS and incubated overnight at 37°C in the solution containing the substrate for beta-galactosidase (beta-gal, encoded by the LacZ cassette) coupled to a NBT salt. The organs and the tissues in the sections display a green/blue staining where PTPH1 gene is normally expressed. After rinsing into PBS, organs were postfixed in PFA 4% for 1 hour, then incubated in 50% glycerol overnight at 4°C and finally maintained in 70% glycerol at room temperature. LacZ staining was observed through a low magnification microscope and described by an operator blind to the genotypes.
LacZ staining was also performed on CNS sections. Mice (n = 3, 12 months old) were sacrificed by ip injection of an overdose of thiopental (5%), perfused with PBS and PFA 4%. Brains were removed and postfixed overnight at 4°C in PFA 4%, then placed overnight at 4°C in 15% and finally in 30% sucrose buffer. The brains were then included in O.C.T. (Tissue-Tek) and sections were cut on slides with a cryostat at 20 μm thickness. The slides were incubated in LacZ staining solution (see above) overnight at 37°C, washed thrice in PBS (5 min each) and either counterstained with H&E (Merck KGaA)  or co-expressed with NeuN immunostaining. Briefly, sections were incubated for 3 hours at room temperature in blocking solution (Vectastain Kit), washed in TBS (Tris-buffered saline), incubated overnight at 4°C with a solution containing the primary antibody mouse anti-mouse Neuronal Nuclei (Chemicon MAB377, 1/1000). The staining was revealed by ABC kit secondary antibody (mouse Vectastain Kit), and DAB (Sigma). After dehydration, sections were transferred onto coverslips. LacZ staining and co-expression with NeuN-immunoreactivity (NeuN-ir) was observed by microscopy and described by an operator blind to the genotypes.
Semiquantitative RT-PCR for beta galactosidase gene
Semiquantitative RT-PCR for PTPH1 and beta-gal gene expression was performed in different brain areas of PTPH1-KO and WT mice in order to confirm the presence of beta-gal expression in the KO tissues, replacing PTPH1 PDZ and catalytic domain. Brains from KO and WT mice (n = 5, 6 months old) were freshly removed and rinsed in HBSS. Hippocampus, cerebellum, cortex, striatum, midbrain and olfactory bulbs were dissected. Total RNA was extracted using Trizol Reagent (Invitrogen) and cleaned-up by RNAeasy columns from Qiagen. 5 μg of total RNA were used to perform the RT-PCR reaction (SuperScript II RT kit, Invitrogen). The primer sequences for LacZ amplification were the following: LacZ – forward 5'-GAT GTA CGT GCC CTG GAA CT/reverse 5'-GGT CCC ACA CTT CAG CAT TT. In order to load equally the reaction mixes, a 300 bp fragment of Histone 2A was amplified as a house keeping gene with the following primers: H2Az forward – 5' CGT ATT CAT CGA CAC CTG AAA; H2Az reverse – 5' CTG TTG TCC TTT CTT CCC GAT.
Behavioral phenotyping test battery
Neurological functions of PTPH1-WT and KO mice (males and females, 11 weeks-old, n = 10 per gender per genotype) were assessed through a behavioral test battery.
The sequence of the test battery was chosen from the least invasive to the most ones. The schedule of the testing sessions included one week of recovery from one test to the next, as reported in Table 1.
After one hour of adaptation in the testing room, each mouse was placed in an open field chamber (50 cm2 wide with white floor and walls) (ViewPoint Life Sci. Inc.) to test locomotor activity and anxiety-like behaviors. Locomotion was recorded for one hour by a video camera and analyzed automatically by VideoTRACK® software (ViewPoint Life Sci. Inc.). Locomotor activity was evaluated by calculating the total path length traveled, whereas the relative time spent in the center was taken as indicative of anxiety-like behavior . The tests were performed in two sessions with equivalent group representation.
Elevated plus maze
After one hour of adaptation in the testing room, anxiety-like behavior was tested for each mouse by EPM within one session. The apparatus consists of four arms (29.5 cm long and 5 cm wide each). Two arms are open whereas the 2 others are limited by 2 black walls (20 cm high). The number of entries of each mouse in the open and closed arms was recorded by a video camera during a period of 5 minutes and analyzed by the SMART Video-Tracking Software (ViewPoint Life Sci. Inc.). The total number of entries into the arms is an index of locomotion, whereas the percentage of time spent and percentage of entries in the closed arms is an index of anxiety-like behaviors .
Motor ability, coordination and learning were evaluated by using an Accelerated Rotarod apparatus for mice (Cat. # 7650 by Jones and Roberts, distr. by Basile Instr., Italy). The apparatus was placed within the animal colony room and was cleaned after each trial. Mice were tested for their abilities to maintain a balance on a rotating bar, which accelerated from 4 to 40 rpm/min in a 5 min trial. Latency to fall off was measured within one session and all mice underwent four trials (one every 30 min) [26–28]. The differences at the rotarod performances in WT and KO were assessed by a single set of trials [27, 28]. This set-up allows a major focus on the early phases of motor learning, involving a strong activation of prefrontal cortex and of the associative areas of basal ganglia and cerebellum [29, 30].
After one hour of adaptation in the testing room, mice were tested on a Y maze apparatus (40 cm long/8 cm wide arms with transparent walls) to investigate spatial working memory . The number and the sequence of the arm entries for each mouse were recorded during 5 minutes. The locomotion index was calculated as the overall number of arm entries, whereas the working memory index was calculated as following: number of exact alternations (entries into three different arms consecutively)/possible alternations (i.e. the number of arms entered minus 2) × 100.
Thermal sensitivity was assessed by a hot plate apparatus for mice (Cat. # 7280 by Biol. Research Apparatus, distr. by Basile Instr., Italy) and lasted a maximum of 45 seconds, time at which damages could occur . The apparatus was placed in the animal colony room and all the mice were tested within one session. Animals were placed on a surface heated at 52.5°C and the latency (seconds) to shake or lick the paw was recorded by the operator.
Statistical comparisons were performed by unpaired two-tailed T-test (p < 0.05) and two-way ANOVA (p < 0.05) followed by post-hoc test as necessary. In the accelerated rotarod, two-way ANOVA with repeated measures followed by T-test was used. Results are expressed as mean ± SEM.
LacZ staining in whole mount
In PTPH1-KO adult animals, LacZ staining in the brain was observed in the cerebellum, hippocampus and in the thalamic nuclei. In addition a strong staining was observed in the cerebral cortex, tenia tecta and septum (Figures 1a, 1b).
LacZ staining in sections and RT-PCR results
No gross cytoarchitectural brain differences were observed by simple visual observation at the microscope in the cortex, hippocampus and thalamus in PTPH1-KO mice compared to WT littermates.
LacZ staining was performed on frozen brain sections to confirm and to describe the expression of PTPH1 at the brain structural level (Table 2).
In cortical regions, LacZ was expressed in the external pyramidal (III) and internal granular layer (IVA) of the cerebral cortex (Figures 2a, 2b), in the retrosplenial cortex (Figures. 3a, 3b 3c, 4a and 4b) and indusium griseum (Figures 3a, 3b). In the cerebellum, in spite of a strong staining in the whole mount (Figure 2), only a faint LacZ signal was observed in sections (Figures 5a, 5b) in particular in the granule cells, close to the nuclei. The RT-PCR on cortical and cerebellar extracts confirmed the presence of LacZ expression in these brain areas (Figure 6).
In subcortical regions, LacZ was detected in the anterior ventral, mediodorsal, ventrolateral, anteromedial and central lateral thalamic nuclear groups (Figure 7a). In more caudal thalamic areas, LacZ was again detectable in the posterior thalamic nuclear group (Po), and to a lesser extent in posteromedial, in posterolateral and in reticular thalamic nuclei and also in the dorsal lateral geniculate nuclei (Figures 7b, 7c, 7d and 7e). In the tenia tecta, LacZ staining visible in the whole mount preparation was confirmed (Figures 2, 8a, 8b). The RT-PCR on substriatal regions including the thalamus, the midbrain and the pontine areas confirmed the presence of LacZ expression in some of these brain areas (Figure 6). To exclude any potential impact of LacZ blood signal contamination in brain areas, RT-PCR on 5 to 20 μl of whole blood was carried out and did not reveal any significant signal [Additional files 2, 3].
In the hippocampus, LacZ expression was observed in the cytoplasm of a few pyramidal cells and through the fibers of the oriens and radiatum layer in a rostral caudal spread (Figure 9a). In rostral sections, LacZ was expressed in the septohippocampal nuclei (Figure 8a). In more caudal sections LacZ was present in the CA1 and CA3, and in a lesser extent in the CA2 (Figures 9b, 9c and 9d). In the CA3 LacZ was strongly expressed in the oriens and pyramidal cell layer (Figure 9d), but its intensity was reduced in the radiatum and oriens compared to CA1 (Figure 9b). No staining was detected in the lacunosum-molecular layer in CA1, CA2 and CA3 (Figures 9b, 9c and 9d). The dentate gyrus showed a strong positive LacZ signal in the molecular layer, but not in the hilus (Figure 9e).
As previously demonstrated, PTPH1-KO mice were healthy, reproduced normally and did not show any phenotypic traits distinguishing them from their WT littermates by simple visual observations [22, 33]. An increased body weight has been detected in PTPH1-KO mice compared to WT littermates, more pronounced in male mice and probably due to an enhanced GHR sensitivity, that leads to increased IGF-1 mRNA and protein expression in liver and plasma, respectively .
In EPM, open field test and hot plate tests (anxiety-related behavior and thermal pain sensitivity), PTPH1-KO male and female mice did not show any significant differences in comparison with their WT littermates (data not shown).
In the accelerated rotarod and Y-maze test, significant differences were observed between PTPH1-KOs and WTs based on gender and genotype factors. In the accelerating rotarod test PTPH1-WT mice did not show any gender differences (P2WAY = 0.5824 (WT gender vs WT activity); PAUC = 0.3218 (PTPH1-WT male vs female)) (Figure 10a). PTPH1-KO male mice displayed an overall significant better performance compared to their matched female littermates (P2WAY = 0.007 (KO gender vs KO activity) (Figure 10b). Post-hoc T-test analyses showed that the difference was significant at the second trial of the test (P0 = 0.109; P30 = 0.015; P60 = 0.067; P90 = 0.835), and the area under the curve for PTPH1-KO male mice was significantly higher (by 50%) compared to the matched values of the female littermates (P = 0.0194) (Figure 10c).
Considering this gender effect, the follow up analysis was carried out in males or females assessing genotype effects on activity. PTPH1-KO female mice performed significantly worse compared to their matched WT littermates, starting from the second trial and onwards (P0 = 0.171; P30 = 0.002; P60 = 0.028; P90 = 0.025) (Figure 10d). No significant differences were observed in PTPH1-KO male mice compared to their matched WT littermates (P0 = 0.92; P30 = 0.363; P60 = 0.222; P90 = 0.135) (Figure 10e).
In the Y-maze test, no differences were detected between PTPH1-KO and WT female mice either in working memory (Pfemale = 0.972) or in locomotion indices (Pfemale = 0.73; Figures 11a, 11b). On the other hand, PTPH1 KO male mice displayed a significantly higher working memory index (percentage of exact alteration; Pmale = 0.041) but similar locomotion activity (total arm entries) (Pmale = 0.348) compared to their matched WT littermates (Figures 11a, 11b).
PTPs are key factors in multiple signaling pathways, leading to modulated functional activities in various cell types [34, 35]. Among all PTP forms, PTPH1 has been shown in vitro to modulate cardiac sodium channel Nav1.5 , that it is also known to be expressed in the axons of cerebral cortex, cerebellum, thalamus and brain stem . Moreover, PTPH1 contains a domain with high sequence homology with the members of the band 4.1 superfamily protein, FERM. This domain mediates the linkage of actin filaments to the plasma membrane , and therefore may be involved in cytoskeleton-membrane interactions, crucial for axon functionality. To further understand the potential role of PTPH1 in neural functions in vivo, we first investigated its expression pattern in embryonic and adult PTPH1-KO mice CNS by LacZ staining, and second its role in CNS functions by behavioral phenotype characterization.
In rat embryonic stage Es19, PTPH1 expression through FISH analyses has already been shown in the dorsal thalamic nuclei, which give rise to the thalamo-cortical connections in adulthood . Thus, it has been suggested to play a role in the maintenance of these connections in adults. We replicated these data in PTPH1-KO mice at Es14 and Es16 embryonic stages. PTPH1 is expressed in the hypothalamic area and but also in the dorsal root ganglia of the spinal cord, excluding the spinal cord itself [Additional file 1] . Moreover, at postnatal P1, PTPH1 expression is also present in peripheral organs such as muscles and intestines as in the adults . On the other hand, the CNS expression at P1 appears weaker than in the adults suggesting a pattern of PTPH1 expression corresponding to specific developmental stages of the CNS as well as peripheral organs (data not shown). These changes in expression may play a role in various developmental functions that need to be further understood.
In PTPH1-KO adults, LacZ is expressed in different CNS areas such as cerebral and retrosplenial cortices (Figures 1, 2, 3 and 4), hippocampus (Figure 9), thalamus (preferentially ventral thalamus) (Figure 7), cerebellum (Figure 5) and in the region of the tenia tecta (Figures 1, 8). This data confirms previously observed expression patterns in the rat brain by Sahin et al.  and extends the observation to other brain regions. We, furthermore, demonstrate that PTPH1 is expressed within the cytoplasm and close to the cell membrane of neurons in most of the brain area investigated (Figures 4a, 4b). It is known that the FERM domain is indeed necessary for PTPH1 localization close to the plasma membrane in Jurkat T cells  and it could be responsible for the punctate expression pattern of PTPH1 in the cytosol of the neurons (Figure 4b) . This supports the concept that PTPH1 may be involved in cytoskeleton-membrane interaction within extended neuronal population in the CNS, potentially playing a role in various neuronal functions.
Indeed the neural expression of PTPH1 in CA1, CA3 and DG of the hippocampus (Figures 9a, 9b, 9c, 9d and 9e), in the retrosplenial cortex (Figures 3, 4) and in a series of thalamic nuclei (Figures 7a, 7b, 7c, 7d and 7e) suggests an involvement of PTPH1 in the modulation of the memory circuit. Both hippocampus and retrosplenial cortex are key regions in the spatial working memory functions [40–46]. Moreover, several thalamic nuclei have also been shown to be important in the memory process [47, 48]. For example, a strong loss of dorsomedial and ventral posterior thalamic neurons is associated with severe cognitive and memory disabilities in patients affected by traumatic brain injury . Lesions in the lateral thalamus may lead to important working memory defects in rodents . The anterior thalamic nuclei project via the retrosplenial cortices to the hippocampus [51, 52], thus underlying the importance of both these circuits and of PTPH1 in the mnemonic process.
Another interesting PTPH1-positive area is the indusium griseum (Figures 3a, 3b) whose role in the adult brain is not clear. It is thought to be part of the limbic system, receiving afferents from the entorhinal and pyriform cortex and projecting to the septohippocampal nuclei, olfactory tubercle (presumably the tenia tecta) and the medial frontal cortex [53, 54]. The expression of PTPH1 in these specific regions suggests a potential role in the processing/integration of memory and sensory information to the SHi and likely the cortex.
Indeed PTPH1 expression is also detectable in the pyramidal neurons in layer III and IVA of the cerebral cortex of the mouse (Figures 2a, 2b), in agreement with Sahin's findings in the rat brain. The middle layers (III and IV) of the cerebral cortex are key sites for thalamic inputs [55, 56] especially for VPM and VPL, primary thalamic nuclei for somato-sensory information integration . Furthermore a strong cortico-cortical communication has been assessed between these two layers , thus suggesting a role for PTPH1 as key regulator in the transmission of the thalamo-cortical and cortico-cortical information.
The cerebellar cortex is also positive for PTPH1 expression, in particular in the cytoplasm of granule cells (Figure 5b). The cerebellum is known to be the main structure for motor learning functions. In particular, the cerebellar cortex seems to be involved in the early learning phases of motor activities [59, 60] that include also a strong activation of other areas such as prefrontal cortex and basal ganglia [29, 30]. PTPH1 expression in the granule cells seems to indicate a potential involvement in the processing of afferent information to the purkinje cells, since it is known that afferents fibers to the cerebellar cortex will project in part through the granule cell layer.
PTPH1 expression pattern observed in our analysis points out a potential involvement of this phosphatase in numerous CNS processing functions such as locomotion, sensorial integration, learning and memory. In this study, the behavioral phenotyping of the PTPH1-KO mice allowed us to test these hypotheses in vivo. Indeed, as already demonstrated by our group  and also by others , PTPH1-KO mice are healthy and do not display any phenotype, distinguishing them from their matched WT littermates, detectable by simple visual observation. Therefore PTPH1-WT and KO mice underwent a battery composed by five behavioral tests, from the least to the most invasive (Table 1), with the tolerable limitation of the handling bias.
Behavioral testing revealing locomotor dysfunctions, such as open-field, EPM and Y-maze did not highlight differences between the two genotypes (Figure 11b), suggesting that PTPH1 does not play a critical role in the integration of locomotor information.
Anxiety-like behaviors measured by open-field (as path in the center) and EPM (as time spent in the open arms), exploiting rodents natural aversion to open space, did not show any differences between the two genotypes (data not shown), leading to the conclusion that PTPH1 may not be involved in the integration of thalamo-limbic information, key paths for anxiety behavior processing. Similar conclusions can be drawn from the lack of difference between the genotypes regarding integration of nociceptive information, based on hot plate test.
In the behavioral test, that partly depends on working memory performances (Y maze), PTPH1-KO male mice showed a slightly better short-term memory than their WT littermates (Figure 11a). Thus, PTPH1 may be involved in the integration of memory information. This was further strengthened by results obtained with a test assessing learning and coordination, the rotarod. Contrary to other behaviors where little differences have been observed, learning and coordination capacities in PTPH1-KO female mice are significantly impaired (Figures 10b, 10c). The low rotarod performance on the early trials, compensated by the last trial, is suggestive of a delay in learning acquisition (Figures 10b, 10d).
As reported in Pilecka et al., our PTPH1-KO mice express the non-catalytic part of PTPH1 in frame with the enzymatically active part of LacZ gene. LacZ is widely used as a reporter for promoter activity in KO mice and all those mice express a modified protein, whose full function is not known. So far it was never reported a function of LacZ alone in cognition and we consider quite unlikely that this is the case in our mice. Thus, it is very likely that the behavioral phenotype we detect in our mice is linked to the deletion of the catalytic domain of PTPH1.
The impairment in learning and coordination of PTPH1-KO female mice may be resulting from the involvement of PTPH1 in the GH signaling pathway . Indeed our group has already shown that PTPH1-KO mice display higher GHR response in vivo and consequently a higher expression of its down-stream effector hormone, the IGF1 in liver and plasma . GHR is highly expressed in most areas of the CNS, in particular in the choroid plexus, hippocampus, putamen, thalamus and hypothalamus. Similarly IGF1 and IGF1-receptors are localized predominantly in hippocampus, but also in amygdala, cerebellum and cortex . Although IGF1 is considered a neuroprotective hormone, it can be produced in the CNS, it is primarily synthesized in the liver and can cross the blood-brain barrier [62–65]. The GH-IGF1 axis is also known to influence cognitive functions due to several neuroprotective effects on the hippocampus . Furthermore it has been recently pointed out that old conditional liver-IGF1-KO mice display impaired spatial learning and memory . The presence of PTPH1 in key CNS regions, as well as the consequent deregulation of the GH-IGF1 axis in KO mice, strengthens the concept that the PTPH1 network (CNS and downstream peripheral effectors) may be involved in cognitive functions.
The behavioral tests assessing working memory and specifically learning revealed not only a genotype effect but also a gender effect, as mentioned above. Sex hormones are known to modulate the somatotropic system [68, 69]. In humans, testosterone has an important effect on GH axis, in part by its aromatization to estradiol. Administration of estrogens, or aromatized androgen, modulates GH axis neuroregulation [69, 70]. In particular, chronic E2 administration has been shown to reduce GH-induced IGF1 increased expression in liver and plasma via a negative feedback mechanism, while acute E2 administration leads to the expected GH-induced IGF1 release . Furthermore, it has been reported that estrogens play not only regulatory functions on neuroendocrine systems but can also have stimulatory or inhibitory impacts on the inter-connectivity of the hippocampal structure depending on the gender [72–75], meaning that the same stimulus can have opposite effects in male vs female mice. Thus, the cognitive behavioral differences observed in our KO mice are underlying the potential impact of the PTPH1 network on neuroendocrine regulation as well as on cellular architecture within specific brain regions.
In conclusion, we have demonstrated that PTPH1 is expressed in neural populations present in adult brain areas mainly involved in locomotor and cognitive functions. The behavioral assessments have allowed us to reveal PTPH1 functionality especially within cognitive domains. Better understanding the interplay between various phosphatases regulating CNS functions, which now includes PTPH1, will be key in the future to unravel some of the complexity of CNS signaling pathways necessary for information processing.
protein tyrosine phosphatase H1
central nervous system
protein tyrosine kinases
protein tyrosine phosphatases
receptor-like protein tyrosine phosphatases
4.1, Ezrin, Radixin, Moesin
TNFα converting enzyme
growth hormone receptor
insulin-like growth factor 1
bacterial artificial chromosome
phosphate buffered saline
Hank's balanced salts solution
Elevated plus maze
area under the curve
mediodorsal lateral thalamic nuclei
anteroventral thalamic nuclei
ventromedial thalamic nuclei
ventral posteromedial thalamic nuclei
ventrolateral thalamic nuclei
ventral posterolateral thalamic nuclei
laterodorsal thalamic nuclei
posterior thalamic nuclei
reticular thalamic nuclei
ventral posteromedial parvicel thalamic nuclei
lateral geniculate nucleus
parafascicular thalamic nuclei
Stoker AW: Protein tyrosine phosphatases and signalling. J Endocrinol. 2005, 185: 19-33. 10.1677/joe.1.06069.
Andersen JN, Jansen PG, Echwald SM, Mortensen OH, Fukada T, Del Vecchio R, Tonks NK, Moller NPH: A genomic perspective on protein tyrosine phosphatases: gene structure, pseudogenes, and genetic disease linkage. FASEB J. 2004, 18: 8-30. 10.1096/fj.02-1212rev.
Alonso A, Sasin J, Bottini N, Friedberg I, Friedberg I, Osterman A, Godzik A, Hunter T, Dixon J, Mustelin T: Protein tyrosine phosphatases in the human genome. Cell. 2004, 117: 699-711. 10.1016/j.cell.2004.05.018.
Stepanek L, Stoker AW, Stoeckli E, Bixby JL: Receptor tyrosine phosphatases guide vertebrate motor axons during development. J Neurosci. 2005, 25: 3813-3823. 10.1523/JNEUROSCI.4531-04.2005.
Wallace MJ, Batt J, Fladd CA, Henderson JT, Skarnes W, Rotin D: Neuronal defects and posterior pituitary hypoplasia in mice lacking the receptor tyrosine phosphatase PTPsigma. Nat Genet. 1999, 21: 334-338. 10.1038/6866.
Elchebly M, Wagner J, Kennedy TE, Lanctot C, Michaliszyn E, Itie A, Drouin J, Tremblay ML: Neuroendocrine dysplasia in mice lacking protein tyrosine phosphatase sigma. Nat Genet. 1999, 21: 330-333. 10.1038/6859.
Meathrel K, Adamek T, Batt J, Rotin D, Doering LC: Protein tyrosine phosphatase sigma-deficient mice show aberrant cytoarchitecture and structural abnormalities in the central nervous system. J Neurosci Res. 2002, 70: 24-35. 10.1002/jnr.10382.
Wansink DG, Peters W, Schaafsma I, Sutmuller RP, Oerlemans F, Adema GJ, Wieringa B, van der Zee CE, Hendriks W: Mild impairment of motor nerve repair in mice lacking PTP-BL tyrosine phosphatase activity. Physiol Genomics. 2004, 19: 50-60. 10.1152/physiolgenomics.00079.2004.
Hironaka K, Umemori H, Tezuka T, Mishina M, Yamamoto T: The protein-tyrosine phosphatase PTPMEG interacts with glutamate receptor delta 2 and epsilon subunits. J Biol Chem. 2000, 275: 16167-16173. 10.1074/jbc.M909302199.
Sahin M, Slaugenhaupt SA, Gusella JF, Hockfield S: Expression of PTPH1, a rat protein tyrosine phosphatase, is restricted to the derivatives of a specific diencephalic segment. Proc Natl Acad Sci U S A. 1995, 92: 7859-7863. 10.1073/pnas.92.17.7859.
Zhang SH, Eckberg WR, Yang Q, Samatar AA, Tonks NK: Biochemical characterization of a human band 4.1-related protein-tyrosine phosphatase, PTPH1. J Biol Chem. 1995, 270: 20067-20072. 10.1074/jbc.270.34.20067.
Zhang SH, Kobayashi R, Graves PR, Piwnica-Worms H, Tonks NK: Serine phosphorylation-dependent association of the band 4.1-related protein-tyrosine phosphatase PTPH1 with 14-3-3beta protein. J Biol Chem. 1997, 272: 27281-27287. 10.1074/jbc.272.43.27281.
Takeuchi K, Kawashima A, Nagafuchi A, Tsukita S: Structural diversity of band 4.1 superfamily members. J Cell Sci. 1994, 107 ( Pt 7): 1921-1928.
Gjorloff-Wingren A, Saxena M, Han S, Wang X, Alonso A, Renedo M, Oh P, Williams S, Schnitzer J, Mustelin T: Subcellular localization of intracellular protein tyrosine phosphatases in T cells. Eur J Immunol. 2000, 30: 2412-2421. 10.1002/1521-4141(2000)30:8<2412::AID-IMMU2412>3.0.CO;2-J.
Han S, Williams S, Mustelin T: Cytoskeletal protein tyrosine phosphatase PTPH1 reduces T cell antigen receptor signaling. Eur J Immunol. 2000, 30: 1318-1325. 10.1002/(SICI)1521-4141(200005)30:5<1318::AID-IMMU1318>3.0.CO;2-G.
Sozio MS, Mathis MA, Young JA, Walchli S, Pitcher LA, Wrage PC, Bartok B, Campbell A, Watts JD, Aebersold R, Van Huijsduijnen RH, van Oers NS: PTPH1 is a predominant protein-tyrosine phosphatase capable of interacting with and dephosphorylating the T cell receptor zeta subunit. J Biol Chem. 2004, 279: 7760-7769. 10.1074/jbc.M309994200.
Zhang SH, Liu J, Kobayashi R, Tonks NK: Identification of the cell cycle regulator VCP (p97/CDC48) as a substrate of the band 4.1-related protein-tyrosine phosphatase PTPH1. J Biol Chem. 1999, 274: 17806-17812. 10.1074/jbc.274.25.17806.
Lavoie C, Chevet E, Roy L, Tonks NK, Fazel A, Posner BI, Paiement J, Bergeron JJ: Tyrosine phosphorylation of p97 regulates transitional endoplasmic reticulum assembly in vitro. Proc Natl Acad Sci U S A. 2000, 97: 13637-13642. 10.1073/pnas.240278097.
Jespersen T, Gavillet B, van Bemmelen MX, Cordonier S, Thomas MA, Staub O, Abriel H: Cardiac sodium channel Na(v)1.5 interacts with and is regulated by the protein tyrosine phosphatase PTPH1. Biochem Biophys Res Commun. 2006, 348: 1455-1462. 10.1016/j.bbrc.2006.08.014.
Zheng Y, Schlondorff J, Blobel CP: Evidence for regulation of the tumor necrosis factor alpha-convertase (TACE) by protein-tyrosine phosphatase PTPH1. J Biol Chem. 2002, 277: 42463-42470. 10.1074/jbc.M207459200.
Pasquali C, Curchod ML, Walchli S, Espanel X, Guerrier M, Arigoni F, Strous G, Van Huijsduijnen RH: Identification of protein tyrosine phosphatases with specificity for the ligand-activated growth hormone receptor. Mol Endocrinol. 2003, 17: 2228-2239. 10.1210/me.2003-0011.
Pilecka I, Patrignani C, Pescini R, Curchod ML, Perrin D, Xue Y, Yasenchak J, Clark A, Magnone MC, Zaratin P, Valenzuela D, Rommel C, Van Huijsduijnen RH: Protein-tyrosine Phosphatase H1 Controls Growth Hormone Receptor Signaling and Systemic Growth. J Biol Chem. 2007, 282: 35405-35415. 10.1074/jbc.M705814200.
Valenzuela DM, Murphy AJ, Frendewey D, Gale NW, Economides AN, Auerbach W, Poueymirou WT, Adams NC, Rojas J, Yasenchak J, Chernomorsky R, Boucher M, Elsasser AL, Esau L, Zheng J, Griffiths JA, Wang X, Su H, Xue Y, Dominguez MG, Noguera I, Torres R, Macdonald LE, Stewart AF, DeChiara TM, Yancopoulos GD: High-throughput engineering of the mouse genome coupled with high-resolution expression analysis. Nat Biotechnol. 2003, 21: 652-659. 10.1038/nbt822.
Ramboz S, Oosting R, Amara DA, Kung HF, Blier P, Mendelsohn M, Mann JJ, Brunner D, Hen R: Serotonin receptor 1A knockout: an animal model of anxiety-related disorder. Proc Natl Acad Sci U S A. 1998, 95: 14476-14481. 10.1073/pnas.95.24.14476.
Lister RG: The use of a plus-maze to measure anxiety in the mouse. Psychopharmacology (Berl). 1987, 92: 180-185.
Jones BJ, Roberts DJ: The quantiative measurement of motor inco-ordination in naive mice using an acelerating rotarod. J Pharm Pharmacol. 1968, 20: 302-304.
Pettibone DJ, Hess JF, Hey PJ, Jacobson MA, Leviten M, Lis EV, Mallorga PJ, Pascarella DM, Snyder MA, Williams JB, Zeng Z: The effects of deleting the mouse neurotensin receptor NTR1 on central and peripheral responses to neurotensin. J Pharmacol Exp Ther. 2002, 300: 305-313. 10.1124/jpet.300.1.305.
Bontekoe CJ, McIlwain KL, Nieuwenhuizen IM, Yuva-Paylor LA, Nellis A, Willemsen R, Fang Z, Kirkpatrick L, Bakker CE, McAninch R, Cheng NC, Merriweather M, Hoogeveen AT, Nelson D, Paylor R, Oostra BA: Knockout mouse model for Fxr2: a model for mental retardation. Hum Mol Genet. 2002, 11: 487-498. 10.1093/hmg/11.5.487.
Hikosaka O, Nakamura K, Sakai K, Nakahara H: Central mechanisms of motor skill learning. Curr Opin Neurobiol. 2002, 12: 217-222. 10.1016/S0959-4388(02)00307-0.
Rustay NR, Wahlsten D, Crabbe JC: Influence of task parameters on rotarod performance and sensitivity to ethanol in mice. Behav Brain Res. 2003, 141: 237-249. 10.1016/S0166-4328(02)00376-5.
Holcomb LA, Gordon MN, Jantzen P, Hsiao K, Duff K, Morgan D: Behavioral changes in transgenic mice expressing both amyloid precursor protein and presenilin-1 mutations: lack of association with amyloid deposits. Behav Genet. 1999, 29: 177-185. 10.1023/A:1021691918517.
Mogil JS, Richards SP, O'Toole LA, Helms ML, Mitchell SR, Belknap JK: Genetic sensitivity to hot-plate nociception in DBA/2J and C57BL/6J inbred mouse strains: possible sex-specific mediation by delta2-opioid receptors. Pain. 1997, 70: 267-277. 10.1016/S0304-3959(97)03333-2.
Bauler TJ, Hughes ED, Arimura Y, Mustelin T, Saunders TL, King PD: Normal TCR signal transduction in mice that lack catalytically active PTPN3 protein tyrosine phosphatase. J Immunol. 2007, 178: 3680-3687.
Bialy L, Waldmann H: Inhibitors of protein tyrosine phosphatases: next-generation drugs?. Angew Chem Int Ed Engl. 2005, 44: 3814-3839. 10.1002/anie.200461517.
Tonks NK: Protein tyrosine phosphatases: from genes, to function, to disease. Nat Rev Mol Cell Biol. 2006, 7: 833-846. 10.1038/nrm2039.
Ling Wu KN: Localization of Nav1.5 sodium channel protein in the mouse brain. Neuroreport. 2002, 13:
Arpin M, Algrain M, Louvard D: Membrane-actin microfilament connections: an increasing diversity of players related to band 4.1. Curr Opin Cell Biol. 1994, 6: 136-141. 10.1016/0955-0674(94)90127-9.
Vollmer JY, Clerc RG: Homeobox genes in the developing mouse brain. J Neurochem. 1998, 71: 1-19.
Huang X, Cheng HJ, Tessier-Lavigne M, Jin Y: MAX-1, a novel PH/MyTH4/FERM domain cytoplasmic protein implicated in netrin-mediated axon repulsion. Neuron. 2002, 34: 563-576. 10.1016/S0896-6273(02)00672-4.
Cooper BG, Mizumori SJ: Temporary inactivation of the retrosplenial cortex causes a transient reorganization of spatial coding in the hippocampus. J Neurosci. 2001, 21: 3986-4001.
Miro-Bernie N, Ichinohe N, Perez-Clausell J, Rockland KS: Zinc-rich transient vertical modules in the rat retrosplenial cortex during postnatal development. Neuroscience. 2006, 138: 523-535. 10.1016/j.neuroscience.2005.11.049.
Vann SD, Aggleton JP: Testing the importance of the retrosplenial guidance system: effects of different sized retrosplenial cortex lesions on heading direction and spatial working memory. Behav Brain Res. 2004, 155: 97-108. 10.1016/j.bbr.2004.04.005.
Jinno S, Klausberger T, Marton LF, Dalezios Y, Roberts JD, Fuentealba P, Bushong EA, Henze D, Buzsaki G, Somogyi P: Neuronal diversity in GABAergic long-range projections from the hippocampus. J Neurosci. 2007, 27: 8790-8804. 10.1523/JNEUROSCI.1847-07.2007.
Otani S, Connor JA, Levy WB: Long-term potentiation and evidence for novel synaptic association in CA1 stratum oriens of rat hippocampus. Learn Mem. 1995, 2: 101-106. 10.1101/lm.2.2.101.
Gaskin S, White NM: Cooperation and competition between the dorsal hippocampus and lateral amygdala in spatial discrimination learning. Hippocampus. 2006, 16: 577-585. 10.1002/hipo.20187.
White NM, Gaskin S: Dorsal hippocampus function in learning and expressing a spatial discrimination. Learn Mem. 2006, 13: 119-122. 10.1101/lm.138806.
Wolff M, Gibb SJ, Dalrymple-Alford JC: Beyond spatial memory: the anterior thalamus and memory for the temporal order of a sequence of odor cues. J Neurosci. 2006, 26: 2907-2913. 10.1523/JNEUROSCI.5481-05.2006.
Komura Y, Tamura R, Uwano T, Nishijo H, Kaga K, Ono T: Retrospective and prospective coding for predicted reward in the sensory thalamus. Nature. 2001, 412: 546-549. 10.1038/35087595.
Maxwell WL, Pennington K, MacKinnon MA, Smith DH, McIntosh TK, Wilson JT, Graham DI: Differential responses in three thalamic nuclei in moderately disabled, severely disabled and vegetative patients after blunt head injury. Brain. 2004, 127: 2470-2478. 10.1093/brain/awh294.
Mitchell KJ, Johnson MK, Raye CL, Greene EJ: Prefrontal cortex activity associated with source monitoring in a working memory task. J Cogn Neurosci. 2004, 16: 921-934. 10.1162/0898929041502724.
Warburton EC, Baird A, Morgan A, Muir JL, Aggleton JP: The conjoint importance of the hippocampus and anterior thalamic nuclei for allocentric spatial learning: evidence from a disconnection study in the rat. J Neurosci. 2001, 21: 7323-7330.
Van Groen T, Wyss JM: Projections from the anterodorsal and anteroventral nucleus of the thalamus to the limbic cortex in the rat. J Comp Neurol. 1995, 358: 584-604. 10.1002/cne.903580411.
Conde F, Maire-Lepoivre E, Audinat E, Crepel F: Afferent connections of the medial frontal cortex of the rat. II. Cortical and subcortical afferents. J Comp Neurol. 1995, 352: 567-593. 10.1002/cne.903520407.
Kunzle H: The hippocampal continuation (indusium griseum): its connectivity in the hedgehog tenrec and its status within the hippocampal formation of higher vertebrates. Anatomy and Embryology. 2004, 208: 183-213. 10.1007/s00429-004-0384-3.
McLaughlin DF, Juliano SL: Disruption of layer 4 development alters laminar processing in ferret somatosensory cortex. Cereb Cortex. 2005, 15: 1791-1803. 10.1093/cercor/bhi056.
Noctor SC, Palmer SL, McLaughlin DF, Juliano SL: Disruption of layers 3 and 4 during development results in altered thalamocortical projections in ferret somatosensory cortex. J Neurosci. 2001, 21: 3184-3195.
Jones EG: The thalamic matrix and thalamocortical synchrony. Trends Neurosci. 2001, 24: 595-601. 10.1016/S0166-2236(00)01922-6.
Thomson AM, Bannister AP: Interlaminar connections in the neocortex. Cereb Cortex. 2003, 13: 5-14. 10.1093/cercor/13.1.5.
McCormick DA, Thompson RF: Neuronal responses of the rabbit cerebellum during acquisition and performance of a classically conditioned nictitating membrane-eyelid response. J Neurosci. 1984, 4: 2811-2822.
Boyden ES, Katoh A, Raymond JL: Cerebellum-dependent learning: the role of multiple plasticity mechanisms. Annu Rev Neurosci. 2004, 27: 581-609. 10.1146/annurev.neuro.27.070203.144238.
Schneider HJ, Pagotto U, Stalla GK: Central effects of the somatotropic system. Eur J Endocrinol. 2003, 149: 377-392. 10.1530/eje.0.1490377.
Aberg ND, Brywe KG, Isgaard J: Aspects of growth hormone and insulin-like growth factor-I related to neuroprotection, regeneration, and functional plasticity in the adult brain. ScientificWorldJournal. 2006, 6: 53-80. 10.1100/tsw.2006.22.
Edmondson SR, Werther GA, Russell A, LeRoith D, Roberts CT, Beck F: Localization of growth hormone receptor/binding protein messenger ribonucleic acid (mRNA) during rat fetal development: relationship to insulin-like growth factor-I mRNA. Endocrinology. 1995, 136: 4602-4609. 10.1210/en.136.10.4602.
Carro E, Nunez A, Busiguina S, Torres-Aleman I: Circulating insulin-like growth factor I mediates effects of exercise on the brain. J Neurosci. 2000, 20: 2926-2933.
Frago LM, Paneda C, Dickson SL, Hewson AK, Argente J, Chowen JA: Growth hormone (GH) and GH-releasing peptide-6 increase brain insulin-like growth factor-I expression and activate intracellular signaling pathways involved in neuroprotection. Endocrinology. 2002, 143: 4113-4122. 10.1210/en.2002-220261.
Le Greves M, Le Greves P, Nyberg F: Age-related effects of IGF-1 on the NMDA-, GH- and IGF-1-receptor mRNA transcripts in the rat hippocampus. Brain Res Bull. 2005, 65: 369-374. 10.1016/j.brainresbull.2005.01.012.
Svensson J, Diez M, Engel J, Wass C, Tivesten A, Jansson JO, Isaksson O, Archer T, Hokfelt T, Ohlsson C: Endocrine, liver-derived IGF-I is of importance for spatial learning and memory in old mice. J Endocrinol. 2006, 189: 617-627. 10.1677/joe.1.06631.
Bondanelli M, Ambrosio MR, Margutti A, Franceschetti P, Zatelli MC, degli Uberti EC: Activation of the somatotropic axis by testosterone in adult men: evidence for a role of hypothalamic growth hormone-releasing hormone. Neuroendocrinology. 2003, 77: 380-387. 10.1159/000071310.
Veldhuis JD, Metzger DL, Martha PM, Mauras N, Kerrigan JR, Keenan B, Rogol AD, Pincus SM: Estrogen and testosterone, but not a nonaromatizable androgen, direct network integration of the hypothalamo-somatotrope (growth hormone)-insulin-like growth factor I axis in the human: evidence from pubertal pathophysiology and sex-steroid hormone replacement. J Clin Endocrinol Metab. 1997, 82: 3414-3420. 10.1210/jc.82.10.3414.
Giustina A, Veldhuis JD: Pathophysiology of the neuroregulation of growth hormone secretion in experimental animals and the human. Endocr Rev. 1998, 19: 717-797. 10.1210/er.19.6.717.
Murphy LJ, Friesen HG: Differential effects of estrogen and growth hormone on uterine and hepatic insulin-like growth factor I gene expression in the ovariectomized hypophysectomized rat. Endocrinology. 1988, 122: 325-332.
Donahue CP, Kosik KS, Shors TJ: Growth hormone is produced within the hippocampus where it responds to age, sex, and stress. Proc Natl Acad Sci U S A. 2006, 103: 6031-6036. 10.1073/pnas.0507776103.
Shors TJ, Chua C, Falduto J: Sex differences and opposite effects of stress on dendritic spine density in the male versus female hippocampus. J Neurosci. 2001, 21: 6292-6297.
Shors TJ, Falduto J, Leuner B: The opposite effects of stress on dendritic spines in male vs. female rats are NMDA receptor-dependent. Eur J Neurosci. 2004, 19: 145-150. 10.1046/j.1460-9568.2003.03065.x.
Wood GE, Beylin AV, Shors TJ: The contribution of adrenal and reproductive hormones to the opposing effects of stress on trace conditioning in males versus females. Behav Neurosci. 2001, 115: 175-187. 10.1037/0735-7044.115.1.175.
K.B.J F, G P: The mouse brain in stereotaxic coordinates. Academic Press. 1997
This work was fully supported by MerckSerono International S.A.. We would like to thank Rob Hooft van Huijsduijnen, Andrea Graziani and Linda Chaabane for their scientific support and for kindly reviewing the manuscript. A special thank to Sonia Carboni and Gabriele Dati for their helpful comments and advices on immunohistochemistry and to Niels Adams for his support on LacZ staining data.
The present work is part of CP’s PhD program at the University of Eastern Piedmont, in close collaboration with MerckSerono International S.A.. MCM, PT, VM, BG, PFZ are employed by MerckSerono International S.A., which is involved in the discovery and the commercialization of therapeutics for the prevention and treatment of human diseases.
The study was devised by CP and MCM and carried out by CP. PT was responsible for the genotyping of all the adult animals that have been used in this study. MA performed the LacZ staining experiment on adult mice. VM and BG have been deeply involved in the first editing of the manuscript and all the authors contributed to modifications in subsequent drafts. PFZ has been involved in critically revising the manuscript and has given the final approval of the version to be published. All the authors read and approved the final version of the manuscript.
Electronic supplementary material
Additional file 1: LacZ staining on PTPH1-WT and KO embryos. PTPH1-WT embryos do not show any staining either at embryological stage 14 (Es14) or at Es16. PTPH1-KO embryos display a positive LacZ staining in the hypothalamic area and but also in the dorsal root ganglia of the spinal cord, excluding the spinal cord itself. (PPT 13 MB)
Additional file 2: Semiquantitative RT-PCR for beta galactosidase gene in blood samples. a: white and red blood cells count in PTPH1-WT and KO mice; no major differences were found in the hematological composition in WT and KO mice. b: Beta-gal mRNA signal was present in PTPH1-KO hippocampus and cortex and not in the WTs; histone H2A gene was used as positive control. c: RT-PCR for beta-gal/H2A on 4 increasing amounts of whole blood (WB): 5, 10, 15 and 20 μl. No signal for beta-gal or H2A was detectable using 5 and 10 μl of WB, due to the low amount of total RNA; a faint signal for H2A was detectable on 15 and 20 μl of WB and a faint band for beta-gal was present only in KO mice, representing the maximum blood contamination in the whole mouse brain. Thus, blood contamination is minimum and it cannot interfere with the main source of signal. (PPT 212 KB)
Authors’ original submitted files for images
Below are the links to the authors’ original submitted files for images.
About this article
Cite this article
Patrignani, C., Magnone, M.C., Tavano, P. et al. Knockout mice reveal a role for protein tyrosine phosphatase H1 in cognition. Behav Brain Funct 4, 36 (2008). https://doi.org/10.1186/1744-9081-4-36