通过UBB+1表达长期抑制转基因小鼠蛋白酶可使中枢呼吸控制功能失调——使人联想起阿尔茨默病中脑干神经病理

UBB⁺¹ overexpressing transgenic mice [UBB⁺¹ tg, line 3413, 008833C57Bl/6.Tg(CaMK2α-UBB) 3413, Jackson #008833] [11] were kept on a C57BL/6J background and heterozygous transgenic mice were analyzed further. Non-transgenic littermates were used as controls. To ensure comparability among individuals, all mice were kept under standard animal housing conditions [e.g., bedding, specified pathogen free (SPF) conditions, light/dark cycle, and climate control] with food and water ad libitum. Mice were analyzed according to the primary systemic phenotyping approach in the German Mouse Clinic at two different ages of approximately 3 and 12 months as described below and elsewhere [16]. Secondary phenotyping screens to extend our primary findings were carried out at additional ages as described in the manuscript. Respiratory functions are presented in detail because these were the most significant effects identified by the primary screening protocol.

Spontaneous breathing patterns of male and female mice (UBB⁺¹ tg mice and controls) were studied at the ages of 3 (n = 6 for all four groups), 12 (n = 6 for all four groups) and 18 months (n = 8 for male tg mice and controls; females were not analysed). For the analysis of lung function and of the ventilatory response to hypercapnic and hypoxic conditions, males at the age of 18 months were studied (n = 8). A detailed description of the whole body plethysmography and the assessment of the lung function are provided in the supplementary methods. Statistical analyses were performed using a commercially available statistics package (Statgraphics^®, Statistical Graphics Corporation, Rockville, MD, USA). Differences between groups were evaluated by Student’s t test. Statistical significance was assumed at p < 0.05. Data are presented as mean ± standard error of the mean (SEM).

Eight 3-month-old, four 6-month-old, five 15-month-old male and four 18-month-old UBB⁺¹ tg mice and three 3-month-old and three 15-month-old male control mice were given deep pentobarbital anesthesia (i.p.) and were perfused intracardially with phosphate-buffered saline (PBS) followed by PBS containing 4 % (w/v) paraformaldehyde (pH 7.4). Dissected brains were sectioned either coronally or sagittally in 50 μm thick sections with a Vibratome. Sections were stained with anti-UBB⁺¹ antibody (Ubi3 16/09/97, final dilution 1:1,000) [10]. As controls for Ubi3 immunoreactivity, the Ubi3 antiserum was adsorbed with ⁶His tag UBB⁺¹ and the pre-immune control serum was used (other controls were reported previously, [11]). In addition, an anti-14-3-3 zeta antibody (ζ, #1001) was generously supplied by A. Aitken (Edinburgh, UK) and tested on brain sections of UBB⁺¹ tg mice (final dilution 1:500). After the primary antibody incubation, the sections were incubated with a biotinylated donkey anti-rabbit antibody followed by avidin–biotin–peroxidase and 3,3′-diaminobenzidine (DAB) intensified by 0.2 % nickel ammonium sulphate (pH 7.6). Photographs were made using the dotSlide BX51 microscope (Olympus, Japan).

The primary phenotyping protocol in the German Mouse Clinic (GMC) included the measurement of more than 240 parameters in the areas of allergy, behavior, bone and cartilage, cardiovascular diseases, clinical chemistry, energy metabolism, eye development and vision, immunology, respiratory function, molecular phenotyping, neurology, nociception, and pathology [13, 15, 16]. As we expected that some mutant phenotypes might be associated with the aging process, UBB⁺¹ tg mice and control animals were analyzed for phenotypes in two cohorts at the ages of about 3 and 12 months. Mice of both age cohorts were analyzed in the primary screening protocol within 8 weeks of the modal age. Significant phenotypic differences between UBB⁺¹ tg and wild-type control mice that are presented here were found for respiratory function. In addition, we observed significantly altered parameters in metabolism, behavior and hippocampal gene expression, which are described in the supplementary information.

Spontaneous breathing patterns were measured at rest and at activity, i.e., during calm sitting and short periods of sleeping and while animals were exploring the test chamber, respectively. Data were analyzed for each gender separately. At 3 months of age, significant genotype-dependent respiration differences between control mice and UBB⁺¹ tg mice were detected for male mice during activity. In this group, the respiratory rate in UBB⁺¹ tg mice was 8 % higher mostly due to a shorter expiratory time (Fig. 1a; S4). In addition, a slightly increased relative duration of inspiration was also observed in male UBB⁺¹ tg mice compared to wild-type control mice (6 % increase, p < 0.001). The spontaneous breathing of UBB⁺¹ tg mice was even more affected at the age of 12 months when respiratory timing was significantly altered in male UBB⁺¹ tg mice during activity and at rest (Fig. 1b, c). The trend towards longer time of expiration and shorter relative duration of inspiration during aging was consistent among UBB⁺¹ tg males (Fig. 1b, c; Fig. S5) and females (Fig. S1, S5) in comparison to the respective wild-type control groups. At 12 months of age, male UBB⁺¹ tg mice exhibited significantly shorter relative duration of inspiration at rest (90 ± 3 %, Fig. 1b) and during activity (94 ± 3 %, Fig. 1c) compared to male wild-type control mice (set to 100 %). Since the inspiratory time and the mean inspiratory flow rate were not significantly affected in male UBB⁺¹ tg mice (Fig. S5), the observed shorter relative duration of inspiration values was primarily due to an elevated expiratory time. The mean expiratory flow rate was significantly lower at rest (83 ± 4 %) and activity (91 ± 4 %) in UBB⁺¹ expressing tg mice compared to the corresponding male control group (set to 100 %), which is most likely due to the differences in the expiratory time (Fig. 1b, c). Thus, overall the spontaneous breathing pattern of male transgenic UBB⁺¹ mice was affected by a statistically significant increase of the expiratory time over age. In contrast, the male wild-type control group had a tendency towards reduced expiratory time over age. Female transgenic mice at 12 months of age and male transgenic mice at 18 months of age showed similar trends, which became significant for the latter (p < 0.05; Fig. S1, S2). Absolute values for the spontaneous breathing parameters are provided in the supplementary figures S4, S5 and S6.

Fig. 1 UBB⁺¹ tg spontaneous breathing patterns measured by whole body plethysmography. Male UBB⁺¹ tg (black bars) and control mice (grey bars) at the age of a 3 months and b, c 12 months were analyzed (n = 6 for all groups). a Shows the respiratory rate (f), expiratory time (Te), and relative duration of inspiration (Ti/TT) for male mice under activity. b, c Show the mean expiratory flow rate (MEF), expiratory time (Te), and the relative duration of inspiration (Ti/TT) at the age of 12 months at rest and under activity, respectively. The mean values for the indicated parameters of the control group were set to 100 % and significance by Student’s t test is indicated by *p<0.05, **p<0.01, ***p < 0.001; error bars represent SEM

---

The altered spontaneous breathing pattern of the UBB⁺¹ transgenic mice is not a mutant phenotype that causes obvious or severe suffering of the mice. The UBB⁺¹ transgenic mice appear externally healthy and it is a mutant phenotype that is not apparent by gross inspection of the mice. However, within the primary lung screen of the German Mouse Clinic, which assesses the spontaneous breathing pattern by whole body plethysmography, we rarely observe such a consistent and clear mutant phenotype at least among mouse lines that externally appear rather normal.

To exclude the possibility that the observed differences in breathing patterns between transgenic and wild-type mice were merely caused by altered lung function per se, we tested various lung function parameters, including lung volume, mechanics, and gas exchange. None of the measured parameters were altered in UBB⁺¹ tg mice compared to the wild-type control mice (Fig. S3), suggesting that differences in breathing pattern are not due to lung dysfunction per se but rather due to changes in respiratory regulation.

The breathing pattern of vertebrates with lung respiration is tightly controlled by the partial pressures of carbon dioxide and oxygen in the blood. Peripheral and central chemoreceptors ensure that increased concentrations of blood carbon dioxide (hypercapnia) or decreased levels of blood oxygen (hypoxia) result in increased respiratory efforts. To test the functionality of these feedback loops and to further examine the relevance of the observed changes in the spontaneous breathing pattern, 18-month-old UBB⁺¹ tg and control mice were exposed to hypercapnic or hypoxic conditions.

Under hypoxic conditions, the most prominent response in all mice of the transgenic as well as the control group was observed for the tidal volume (Fig. 2b), which increased about 1.5-fold with 8 % O₂ in the respiratory air in comparison to normal air (about 21 % O₂). When hypoxic conditions were combined with hypercapnic conditions (3 % CO₂ in the respiratory air), the tidal volume further increased almost twofold. The combined challenge of hypercapnic and hypoxic conditions also increased the relative duration of inspiration by about 170 % (Fig. 2d), indicating that it took the animals more time to inhale the increased tidal volume. All values returned to baseline conditions within a 15-min recovery period. Even though UBB⁺¹ tg animals and control mice showed a similar increase in tidal volume under hypoxic conditions at 8 % O₂, they showed significant differences in two other parameters under this condition. Thus, the respiratory rate in UBB⁺¹ tg mice declined to about 75 % at 8 % O₂ (206.7 ± 8.3 min⁻¹) whereas this parameter was only slightly changed in control animals (about 90 %; 246.8 ± 8.4 min⁻¹; p < 0.001; Fig. 2a). Secondary to this, the minute ventilation was also significantly lower in UBB⁺¹ tg mice (Fig. 2c; p < 0.01). Interestingly, when adding the hypercapnic challenge of 3 % CO₂ in the breathing air to the hypoxic challenge, the breathing patterns of UBB⁺¹ tg and controls were not significantly different (Fig. 2a, c). In addition, we examined breathing patterns of mice exposed to 3, 5 and 8 % CO₂ in the air (at normal oxygen conditions). Under these conditions, we did not observe differences between UBB⁺¹ tg animals and wild-type mice (Fig. S7). These findings support the notion that the hypoxic response is primarily affected in UBB⁺¹transgenic mice. In contrast, the feedback loop for CO₂ levels is apparently not affected in UBB⁺¹ transgenic mice.

Fig. 2 Ventilatory responses to hypoxic and hypercapnic conditions are affected by UBB⁺¹ expression. Male mice at the age of 18 months (n = 8 for each group) were exposed to 10 and 8 % O₂ (hypoxia) followed by a combined exposure of 8 % O₂+3 % CO₂ (hypoxia and hypercapnia) each for 7 min and a recovery period of 15 min. a Shows the respiratory rates (f), b shows the tidal volume (TV), c shows the minute ventilation (MV), and d shows the relative duration of inspiration time (Ti/TT). Values were calculated relative to baseline levels (set to 100 %) and significance by Student’s t test is indicated by ****p<0.001, ***p<0.01; error bars represent SEM. The response to hypercapnic conditions is shown in Supplementary Figure S7

---

The expression of UBB⁺¹ in brain sections of UBB⁺¹ tg mice was examined using the UBB⁺¹-specific Ubi3 antibody (bleeding 16-9-97) and a more sensitive procedure (avidin–biotin method) than previously used [11]. In the present study, hippocampal fibers in the CA region were positive, especially in the CA2 region (Fig. 3a, b). The intense fiber staining in the region where the CA3 borders the CA2 appears to reflect an increased fiber density. The incubation of mouse brain tissue with ⁶His tag UBB⁺¹ and the pre-immune control serum resulted in an absence of immunopositive staining (data not shown). There was no staining in the brains of wild-type control mice.

Fig. 3 Immunohistochemical expression of UBB⁺¹ in the brain of tg mice. UBB⁺¹ staining of a, b hippocampus and c–h brainstem of male UBB⁺¹ tg mice at the age of 15 months (hippocampus) and 3 months (brainstem) with the Ubi3 antibody (sagittal section, a/b). Hippocampal cells are immunoreactive for UBB⁺¹. The fibers widespread in the hippocampus are also stained (see boxed area in a), especially in the region where the CA3 borders the CA2 (asterisk). Another region of more intense UBB⁺¹ staining reflects the stratum lacunosum (arrowhead), apparently with crossing fibers of the perforant pathway. Sagittal (c,d) and coronal (e–h) sections of UBB⁺¹ tg mice were incubated with Ubi3. c Note the expression of UBB⁺¹ in the dorsal vagal complex (nucleus tractus of the solitarius, area postrema, dorsal motor nucleus of vagus nerve). d Region of dorsal vagal complex as indicated by the boxed area in c at a higher magnification. e Coronal section shows again the expression of UBB⁺¹ in the dorsal vagal complex and f shows this region (boxed area in e) at a higher magnification. Note in e UBB⁺¹ immunostaining in the spinal 5 nucleus. g Note the expression of UBB⁺¹ in neurons of the different subnuclei of the parabrachial nucleus (boxed area). h Photograph of parabrachial nucleus at a higher magnification. Nomenclature based upon [14]. Abbreviations see paper. Scales: a = 200 μm, b = 20 μm, c = 1 mm, d = 100 μm, e = 500 μm,f = 100 μm, g = 1 mm, h = 100 μm

---

The present immunohistochemical staining procedure also enabled a superior anatomic localization of UBB⁺¹ in other brain regions as compared to our previously published results [11]. In particular, in the brainstem of UBB⁺¹ tg mice labeling with the Ubi3 antibody was prominent in regions associated with known functions in respiratory modulation (Fig. 3c–h). Both in the dorsal respiratory group (Fig. 3c–f) and the pneumotaxic center (Fig. 3g, h), UBB⁺¹ immunoreactive nuclei were found. In UBB⁺¹ tg animals, UBB⁺¹ was present in specific areas of the dorsal vagal complex, such as the area postrema (AP), the nucleus of the tractus solitarius (NTS), and the dorsal motor nucleus of the vagus nerve (Fig. 3c–f). Strongly positive UBB⁺¹ immunoreactive neurons were also present in the locus coeruleus (LC) and subnuclei of the parabrachial nucleus (PBN) (Fig. 3g, h), especially in the external lateral (PBNel) and dorsal lateral (PBNdl) subnuclei, which receive visceral inputs from the caudal NTS. The rodent external lateral nucleus is involved in respiratory function and in the modulation of respiratory responses to hypoxia and hypercapnia [35, 36]. In particular, the central respiratory control of the response to hypoxia was affected in UBB⁺¹ tg mice (Fig. 2). These findings are consistent with the hypothesis that highly elevated UBB⁺¹expression might have mild toxic effects that could interfere with the normal neuronal function [42]. Mice were analyzed for UBB⁺¹ expression at the ages of 3, 6, 15 and 18 months, but no differences among the four ages were found with respect to UBB⁺¹ immunohistochemical staining, which is in agreement with previously published data [11].

The key findings of the respiratory function analysis were related to altered breathing patterns particularly in male UBB⁺¹ tg mice, which affected the respiratory rate and respiratory timing. Respiratory timing was predominantly affected by a reduction of the relative duration of inspiration (Ti/TT) in combination with an increased expiratory time (Te). An affected breathing pattern per se is a rather unspecific finding, which could be due to several factors, (1) altered central breathing control, (2) impaired sensitivity of peripheral chemoreceptors, (3) altered respiratory muscle function, or (4) affected respiratory function per se. To address these different issues, we tested UBB⁺¹ tg mice for respiratory lung function (lung volumes, mechanics, and gas exchange) and the ventilatory response to hypercapnic and hypoxic conditions, respectively. With regard to lung function, the parameters were not different in UBB⁺¹ tg and wild-type mice. These findings indicated that the increased expiratory time is not related to an increased airway resistance that the gas exchange function of the respiratory system per se (4) is not affected by UBB⁺¹ transgenic overexpression and that the alterations in breathing pattern are not caused by impaired respiratory muscle function (3). This conclusion is further supported by the hypercapnic and hypoxic challenge responses. A similar increase up to a factor of 2 in tidal volume in controls and UBB⁺¹ tg mice was observed indicating that UBB⁺¹tg animals have normal thoracic muscle strength. To further distinguish between central (1) and peripheral (2) chemoreceptor sensitivity, mice were exposed to hypercapnic and/or hypoxic conditions and the ventilatory response was measured. The results show that the hypercapnic response was not affected in UBB⁺¹ tg mice, while the hypoxic response was. UBB⁺¹ tg mice showed declining respiratory rates with increasing levels of hypoxia, being 16.3 % lower than in controls at 8 % O₂. As a result, minute ventilation was also significantly lower in UBB⁺¹ tg mice. Interestingly, breathing differences between UBB⁺¹ tg and control mice were virtually absent when the hypercapnic challenge of 3 % CO₂ was added to the hypoxic challenge. This supports the conclusion that primarily the hypoxic feedback loop is affected in UBB⁺¹ tg mice. Hence, these experimental results suggest that an altered central breathing control (1) is the most likely explanation for the observed phenotype.

In support of this hypothesis, we observed strong UBB⁺¹ tg expression in various brainstem nuclei which are involved in the regulation of respiratory function. For example, the nucleus of the solitary tract (NTS), which is part of the dorsal respiratory group, was UBB⁺¹ immunoreactive in all the different subnuclei. The caudal part of this nucleus receives afferent information from baroreceptors, mechanoreceptors and from central and peripheral chemoreceptors. Furthermore, it has been shown that the AP together with NTS project to the PBN [20], which comprises a second respiratory center positive for UBB⁺¹ expression. Specifically, the external lateral and dorsal lateral PBN subnuclei were strongly UBB⁺¹ immunoreactive. It has been reported that the external lateral PBN subnucleus in rodents has an important role in the respiratory response to hypoxic and hypercapnic conditions [35, 36]. This is of particular interest because UBB⁺¹ tg mice showed differences in their response to hypoxia compared to the controls. Interestingly, it has previously been shown that the locus coeruleus area and the PBN are among the early targets of the AD-related cytoskeletal pathology, which suggests that parts of the brainstem respiratory centers are affected during AD progression [4, 33]. UBB⁺¹ tg mice are not the only mouse model with altered respiration. For example, Tau-P301L transgenic mice show a breathing phenotype, which can be linked to Tau (Mapt) expression in defined nuclei of the brainstem such as the Kölliker-Fuse nucleus [8, 29]. Apart from respiratory dysfunction, swallowing impairments are common in AD patients and frequently lead to aspiration pneumonia [23]. The three phases of swallowing (oral, pharyngeal and esophageal) are regulated by central pattern generators in the brainstem and one of them is the NTS [27]. This suggests that there might be a functional link between the respiratory changes exhibited by the UBB⁺¹ tg mice and the respiratory and swallowing dysfunctions in AD patients, because similar anatomical regions are affected as we have shown in our examination of the distribution of UBB⁺¹ in the brains of AD patients. Intriguingly, it was also recently shown that the minute volume in AD patients under peak exercise conditions is reduced in comparison to a non-demented control group [3]. Our data suggest that this change in respiratory response in AD patients could also be due to a dysfunction of central breathing control.

Our comprehensive phenotyping approach revealed a respiratory phenotype in UBB⁺¹ tg mice, which prompted us to perform an extensive immunohistochemical UBB⁺¹ analysis of the brainstem of these mice. In several areas involved in respiratory regulation such as the lateral parabrachial nuclei [35, 36] and the dorsal vagal complex, we detected intense UBB⁺¹ immunoreactivity. Our data suggest that also in other AD mouse models respiratory function analysis needs to be addressed.

As UBB⁺¹ tg mice are an animal model for early AD, we subsequently analyzed UBB⁺¹ expression in AD patients and healthy individuals. In the human brainstem, UBB⁺¹ immunoreactivity and misfolded Tau were seen in similar areas as in UBB⁺¹ tg mice and they may well contribute to the selective vulnerability of the brainstem to AD [31]. A role of the lateral parabrachial nucleus in mediating the feedback control of inspiratory drive has been reported [35]. The dorsal respiratory group communicates with this area and shows UBB⁺¹ immunoreactivity as well. It could well be that breathing dysfunction may play a major role in cognitive, behavioral and emotional impairments in early onset AD [22]. Also, a disproportional contribution of bronchopneumonia as a cause of death in AD patients is consistent with our findings that indicate breathing dysfunction as a contributing factor. With respect to respiratory function in aged patients and during AD per se, it has been noted that increased cardiorespiratory fitness in early-stage AD is associated with reduced brain atrophy as compared with non-demented individuals [6]. Moreover, it was recently reported that aerobic exercise has positive effects on mild cognitive impairment (e.g., [1]), underscoring the need to enhance respiratory function. Together with these published observations, our data might suggest that the examination of the respiratory system could be an important parameter for monitoring the progression of AD. It remains to be analyzed whether other cognitive disorders also display UBB⁺¹ expression and neuronal pathology in central breathing control regions.