The epidermal growth factor receptor (EGFR) pathway plays key roles in the development and maintenance of many tissue and organ systems . Recent reports have suggested that the EGFR pathway mediates two aspects of behavior, diurnal locomotor activity and suppression of locomotion in response to light (masking). Levels of the EGFR ligand transforming growth factor alpha (TGFA) fluctuate with a circadian rhythm within the suprachiasmatic nucleus (SCN) [2,3], which is located within the hypothalamus and is considered the primary anatomical circadian clock, and are associated with circadian time-dependent changes in gene expression ; similarly, EGFR ligands are expressed within cells of the retina, which modulates masking behavior . Both of these structures appear to input into the subparaventricular zone (SPZ), a hypothalamic region that is required for circadian rhythms  and that expresses high levels of EGFR . This anatomical network has been experimentally manipulated, with infusion of TGFA into the hamster hypothalamus reversibly suppressing locomotor activity [3,6]. However, exogenous administration of receptor ligands can lead to non-physiological responses, indicating what a protein can do, not necessarily its normal biological function .
Perhaps the most compelling evidence implicating the EGFR pathway in locomotor activity and masking was found in the behavior of mice homozygous for the Egfrwa2 allele, which produces a hypomorphic receptor with reduced kinase activity [8,9]. The Egfrwa2 allele is a valuable genetic reagent for dissecting biological functions of EGFR since homozygosity for Egfr null alleles results in lethality [1,10], while Egfrwa2 homozygous animals can survive to adulthood. Both abnormally high diurnal activity and strong masking defects were reported in four of five B6EiC3H-a/A-Egfrwa2/wa2 Wnt3avt/vt mice tested . Conflicting with these observations, a larger study using a greater number of mice from the same source (The Jackson Laboratory) and a range of illumination intensities did not detect any masking defects .
Previous analyses investigating alterations in locomotor and masking behaviors used mice carrying the Egfrwa2 allele on a mixed genetic background. In order to resolve the discrepancies implicating a critical role for EGFR in locomotor and masking behaviors, as well as to examine potential genetic background effects on these defects, we tested Egfrwa2 homozygotes on uniform congenic or hybrid backgrounds for locomotor activity and masking ability. The current results conclusively demonstrate that EGFR is not an essential mediator of locomotor or masking behaviors and suggests that other uncontrolled genetic or environmental parameters confounded previous experiments.
B6EiC3H-a/A-Egfrwa2/wa2 Wnt3avt/vt mice were obtained from The Jackson Laboratory (Bar Harbor, ME). The derivation of the Egfrwa2 congenic lines involved backcrossing the Egfrwa2 allele for greater than ten generations to C57BL/6J and 129S1/SvImJ genetic backgrounds. The removal of the linked Wnt3avt hypomorphic allele, maintained in cis with Egfrwa2 in the initial starting population, was verified by PCR-based genotyping. Mice were given at least one week to acclimate to their new environment prior to testing. Mice were provided PicoLab Mouse Diet 20 (LabDiet) ad libitum and autoclaved water.
Mice were housed singly and habituated several days in cages equipped with running wheels for activity measurements, within a light-tight box containing a cool-white fluorescent tube providing 200 to 300 lux or 400 to 500 lux light at cage level on a 12 h:12 h light-dark cycle. Diurnal activity was calculated as the percentage of total activity occurring in the light phase of the cycle. For the masking response analysis, either one or three hour light pulses were delivered on the third day starting at Zeitgeber Time (ZT) 14 (ZT0 = light on; ZT12 = lights off). Each animal was tested for masking to two or three light pulses with at least two days between pulses. Activity during the pulse was calculated as a percentage of the activity at the same time on the previous day. Locomotor activity was recorded and analyzed with ClockLab software (Actimetrics, Evanston, IL).
Results and discussion
Mice with normal light-mediated locomotor activity generally have one to two percent daytime activity, and masking of greater than 95%. An initial round of testing utilized commonly administered conditions, one-hour light pulses at 400–500 lux, to determine masking response. Of the six animals tested, none showed a masking defect, and only a single animal, an Egfrwa2/wa2 Wnt3avt/vt mouse on a mixed genetic background similar to those previously used , demonstrated higher than normal diurnal activity (data not shown).
A larger panel of mice was then tested using three-hour light pulses at 200–300 lux, conditions identical to those of the previous study reporting Egfrwa2-associated abnormalities . Surprisingly, the vast majority of Egfrwa2 homozygous mice exhibited both normal daytime activity and negative masking, as did all wild-type littermates and a C57BL/6J-Wnt3avt/vt mouse (Table 1). Indeed, no statistically significant differences were found when the data were grouped by Egfr status (one-way ANOVA: % daytime activity, p = 0.32; % masking, p = 0.27; total activity (rev), p = 0.49). Two Egfrwa2/wa2 mice, one male C57BL/6J-Egfrwa2/wa2 and one female B6.129 F1-Egfrwa2/wa2, exhibited both abnormal daytime activity and negative masking defects similar, but not as extreme, as those previously described . Unlike what happened in the previous study, the majority of the abnormally high daytime activity seen in the two affected mice was not sporadic, but rather occurred in a discrete time period anticipating the dark cycle and thus resulting in a consistent phase shift (Fig. 1A). Similar to the previous study, the negative masking in mice was delayed, with normal suppression of activity in the first hour of the three-hour light pulse (Fig. 1B). Thus, the current results on controlled genetic backgrounds show substantially less penetrance and expressivity than originally reported (18% versus 80% penetrance) .
|Mouse number||Strain||Genotype||Total activitya||Daytime activityb||Maskingc|
The discrepancy between previous results using the B6EiC3H mixed genetic background and our current results with congenic mice is particularly surprising since most abnormal phenotypes increase in severity with inbreeding, this being particularly striking for phenotypes associated with Egfr [1,12]. However, use of the B6EiC3H-a/A-Egfrwa2 Wnt3avt stock to study effects of the Egfrwa2 allele is immediately problematic since Egfrwa2 homozygotes are also homozygous for the linked Wnt3avt mutation, a hypomorphic allele of Wnt3a known for producing the vestigial tail phenotype . Since Wnt3a-deficient mice exhibit defects in the hippocampus and central nervous system , the Wnt3avt allele is possibly responsible for the activity defects. Additionally, the non-inbred B6EiC3H background, maintained through a cross-outcross mating scheme, harboring the Egfrwa2 and Wnt3avt alleles in cis segregates known and unknown mutations from the C57BL/6JEi and C3H/HeSnJ strains . Consequently, defects like those previously reported for locomotor activity cannot be attributed to Egfr because an appropriate control does not exist for the mixed B6EiC3H background.
Wildtype inbred mice from the C57BL/6JEi and C3H/HeSnJ strains have different masking thresholds and vary in diurnal locomotor activity in a manner that is likely multigenic [16,17]. One strong candidate for such a modifier mutation is retinal degeneration (Pdebrd1), which is carried by the C3H/HeSnJ strain but not C57BL/6EiJ, and causes progressive and selective degeneration of photoreceptor cells [18,19]; the Pdebrd1 mutation alone has a highly significant effect on masking behavior . Melatonin production is also vastly different between the two strains contributing to the B6EiC3H mixed background, with C3H mice exhibiting high, rhythmic melatonin levels, and C57BL/6 mice exhibiting low to undetectable melatonin levels caused by mutation of at least one gene (N-acetyltransferase 2) related to melatonin production [20,21]. The previous Egfrwa2 homozygous cohort that had a high penetrance of activity defects may have fortuitously contained a high frequency of these or other genetic modifiers given the small number of individuals tested.
Previous studies utilizing direct infusion of ligand to hyperactivate EGFR in the brain have shown that abnormally high EGFR signaling is capable of perturbing light-associated locomotor activity control [3,6]. However, such an approach can produce non-physiological responses that are artifactual or neomorphic in nature rather than being representative of normal biology . Thus, the suggestion that locomotor activity is strongly dependent on normal levels of EGFR signaling is not supported, especially since extensive testing of B6EiC3H-a/A Egfrwa2/wa2 Wnt3avt/vt mice across a wide range of lighting conditions failed to detect any differences in masking response . Using appropriately controlled genetic conditions, our results conclusively demonstrate that EGFR activity is not required to produce and is not a major mediator of abnormal activity phenotypes, and that other environmental, genetic, or stochastic effects are required to reveal abnormal phenotypes. A recent report revealed significant intrastrain and intraindividual fluctuations of EGFR ligand levels in the SCN of inbred mice . Thus we cannot eliminate the possibility that homozygosity for Egfrwa2 may cause individuals to be sensitized to phenotypically express abnormal locomotor activities. For example, Egfrwa2 homozygotes have variable eye defects  that could combine with other factors to contribute to light-associated activity abnormalities. In fact one of the Egfrwa2/wa2 individuals in our study supports the presence of intraindividual variability, exhibiting an abnormal masking response in one trial followed by a near-normal response in a subsequent trial (data not shown).
The mice used to originally implicate EGFR as a major mediator of locomotor activity carried numerous other mutations and allelic variants that could contribute to the observed results. Since the animals were not properly controlled for genetic background, numerous other interpretations (such as unequal genetic background distribution in control and test mice) most likely contributed to the erroneous results. Our data using genetically uniform congenic lines indicate that EGFR is not a required mediator of the locomotor or masking behaviors, though it may modulate the activity of other pathways involved in control of locomotor activity. Further investigation is required to properly elucidate the molecular pathways and factors mediating locomotor activity.
The author(s) declare that they have no competing interests.
RBW participated in all experiments, in the analysis and discussion of the results, and in the writing of the manuscript. CLT participated in all experiments, in the analysis and discussion of the results, and in the writing of the manuscript. DL participated in all experiments and in the analysis and discussion of the results. RWM participated in all experiments and in the analysis and discussion of the results. AS participated in the conceptualization of the experiments and in the analysis and discussion of the results. DWT participated in the conceptualization of the experiments, in the analysis and discussion of the results, and in the writing of the manuscript. All authors read and approved the final manuscript.