Upper Airway Obstruction Elicited Energy Imbalance Leads to Growth Retardation that Persists after the Obstruction Removal

Upper airway obstruction can lead to growth retardation by unclear mechanisms. We explored the effect of upper airway obstruction in juvenile rats on whole-body energy balance, growth plate metabolism, and growth. We show that after seven weeks, obstructed animals’ ventilation during room air breathing increased, and animals grew less due to abnormal growth plate metabolism. Increased caloric intake in upper airway-obstructed animals did not meet increased energy expenditure associated with increased work of breathing. Decreased whole-body energy balance induced hindrance of bone elongation following obstruction removal, and array pathways regulating growth plate development and marrow adiposity. This is the first study to show that rapidly growing animals cannot consume enough calories to maintain their energy homeostasis, leading to an impediment in growth in the effort to save energy.

proliferating chondrocytes are regulated Sry-related transcription factor nine (Sox9) that has an important role in chondrogenesis differentiation 39 .
The mechanisms linking whole-body energy balance (i.e., caloric intake vs. energy expenditure) with EGP metabolism/architecture impairment that is AO-induced are poorly understood. We hypothesize that the increased work of breathing leads to abnormal whole-body energy balance and poor development of EGP and linear growth retardation. We used an integrative approach to explore in rapidly growing rats the effects of upper airway obstruction and its removal on whole-body energy balance, EGP metabolism, and linear growth.
In this study, we find that narrowing of the trachea diameter leads to increased energy expenditure and caloric intake, and was not sufficient to meet the energetic demand of breathing. Decreased energy balance leads to the impediment of metabolic processes involved in linear growth that persists after removal of the obstruction. Here, we show that rapidly growing animals cannot consume enough calories to maintain energy homeostasis. Deregulation of energy availability and lack of availability of circulating factors leads to an impediment in linear growth in order to save energy.
Decreased EGP width in AO and its partial improvement following OR (Table 1, Fig. 2C) were associated with reduced bone elongation ( Fig. 2A). Tibia length was 38.5 ± 0.55 (mm, n = 14), 31 ± 0.48 (mm) (p < 0.01, n = 19), and 36.3 ± 0.32 (mm) (p < 0.01, n = 15) for the control, AO, and OR groups, respectively. A three-dimensional   Table 1). The trabecular bone volume to total volume (BV/TV) ratio decreased by 34.8% and 15.6% (p < 0.01) in the AO and OR groups, respectively. Trabeculae number (Tb.N) decreased by 26.8% and 11.5% (p < 0.01) in the AO and OR groups, respectively, while trabecular separation (Tb.Sp) increased by 38.9% (p < 0.01) in both groups. Cortical BV/TV ratio decreased by 10% (p = 0.01) in both the AO and OR groups. Cortical bone mineral density (BMD) decreased by 8.6% and 7.3% (p = 0.01) in the AO and OR group, respectively. Safranin O showed decreased staining intensity of the primary spongiosa in the AO group and was only partially improved following OR (Fig. 2D). Serum TRAP 5b was undetected in all groups. Collagen II mRNA expression was reduced by 67% and 57% in the AO and OR groups, respectively (p < 0.05; Fig. 2E). Osteocalcin mRNA expression was reduced by 17% in the AO group (p < 0.05, Fig. 2F).
The mean number and range of EGP IGF 1 positive cells, as determined by immuno-histochemistry ( Both AO and OR groups had increased marrow adipose cell numbers (Fig. 5A,D). Marrow PPARγ protein increased by 236% (p < 0.01) and 184% (p < 0.01) in the AO and OR groups, respectively (Fig. 5B,C). The mean number and range of PPARγ positive cells were 3 (2-4 range), 17 (6-28 range), and 10 (7-13 range) in the control, AO, and OR groups, respectively (Fig. 5B). No significant change was found in the marrow adipose cell cross-sectional area (Fig. 5E).

Discussion
The effects of AO and its removal on energy balance and EGP development/metabolism were explored from weaning to adulthood. Increased caloric intake in the AO group was not sufficient to meet the energy demand of the increased work of breathing; thus, energy balance (i.e., caloric intake vs. energy expenditure) was reduced close to 60%. This study is the first to show that decreased energy balance led to an impediment in bone elongation and bone development. The EGP growth impediment was associated with an alteration in the array of metabolic processes involved in development, including IGF 1, Sox9, AMPK, and OX1R. Despite the normalization of energy balance in OR, only partial improvement of EGP growth and metabolism were observed. Bone requires a substantial portion of the available fuel and nutrients to generate ATP for proper devlopment and growth 40 . Increased adiposity in the bone marrow was associated with upregulation of PPARγ and loss of bone mass. Our findings indicate that AO leads to abnormal function of the stroma for hematopoietic cell differentiation. Here, we show that rapidly growing animals cannot consume enough calories to maintain energy homeostasis; this deregulation of energy availability and lack of availability of circulating factors leads to impediment in EGP growth/architecture in order to save energy. www.nature.com/scientificreports www.nature.com/scientificreports/ Upper airway obstruction and energy expenditure. Untreated OSA may lead to growth retardation [1][2][3][4]7,8,18,[41][42][43] , and bone mass loss in adults 5 by mechanisms that are poorly understood. To the best of our knowledge, this study is the first to show that deregulation of energy homeostasis plays an important role in AO-induced growth retardation. Increased energy expenditure in AO is due to upregulation of ventilation in order to maintain respiratory homeostasis 30,44,45 and the extra energy needed for increased additional wakefulness 30 . Orexin participates importantly in maintaining respiratory homeostasis in AO, but it is also responsible for partial sleep loss 29 . Decreased EGP width and bone growth in AO were associated with reduced mRNA expression of type II collagen, the main component of cartilage 46 and osteocalcin 47 , respectively. Interestingly, serum osteoclast marker TRAP 5b was undetected in our study. Insufficient sleep over a long duration can lead to poor bone health that is associated with increased TRAP 5b and decreased osteoblast activity 31 . AO animals increased their caloric intake due to elevation of gut-derived ghrelin, an array of hypothalamic mediator factors, and orexin 30 . Orexin and ghrelin-containing neurons could influence each other, and thereby regulate feeding www.nature.com/scientificreports www.nature.com/scientificreports/ behavior 48 , while short sleep per se also can stimulate gut-derived ghrelin and feeding 49 . Although caloric intake increased, the slow body weight gain strongly indicates the higher energetic cost of airway obstruction, and AO animals could not consume enough calories to meet the additional energy requirements to maintain energy homeostasis -a condition that is not associated with malabsorption of calories 15,50 . Tissue composition of water, protein, and fat in the vital organs of AO are largely spared, and the slow body weight gain was mostly related to reduced adiposity tissue 15 . In contrast, food restriction resulting in weight loss is associated with severe organ weight losses 51 .
Earlier studies in children demonstrated that the z score for weight correlates with sleeping energy expenditure in OSA because of the increased work of breathing during sleep 1 . AO in rats, on the other hand, was both inspiratory and expiratory, and may be representative of upper airway resistance patients' that is not exclusively sleep related (i.e., patients with increased nasal resistance, subglottic or tracheal stenosis, retrognathia, or macroglossia). In OSA, airway obstruction during sleep is intermittent with opportunity for recovery during the day 2,43 . The amount of oxygen consumed by the respiratory muscles is about 1% to 2% of the basal VO 2 52 . Increased resistive breathing and a variety of other pulmonary diseases can substantially increase the energetic cost of breathing. Regulation of energy balance depends on factors such as basal metabolic rate, thermic effect of food, and caloric intake 11,53,54 . OSA may also elevate energy expenditure by increasing sympathetic activity [11][12][13] . An increased firing rate of the sympathetic nerves to brown adipose tissue through β-adrenergic receptors increases energy expenditure to generate heat. However, this possibility is unlikely since AO animals do not respond to administration of norepinephrine to generate heat due to decreased brown adipose uncoupling protein 1, leptin level, and lack of available fuel for heat generation 30 . Growth plate metabolism/architecture. AO-induced growth retardation, and despite the normalization of energy balance in OR, only partial improvement in tibial length and EGP width was observed. The skeleton in highly metabolic active tissues requires substantial amounts of energy, particularly during periods of rapid growth and physical activity 40,55,56 . Bone formation requires available fuel and communicates metabolic demands to other metabolically active tissues via circulating factors 40 . Further studies are needed to explore the effects of   www.nature.com/scientificreports www.nature.com/scientificreports/ shorter obstruction duration and longer recovery period on growth velocity to determine if growth retardation is largely irreversible. Growth retardation was associated with a large suppression of an array of metabolic processes that are involved in EGP and hematopoietic cell differentiation. The IGF system plays an important role in early longitudinal growth by acting both as an endocrine and as an autocrine/paracrine close to the site of synthesis 20 . IGF1 with IGFBP2, one of the main EGP and bone IGFBPs, stimulate AMPK activation and osteoblast differentiation, and AMPK knockout mice have reduced bone mass 37 . OR normalized EGP IGF 1 cell number; however, IGFBP 2 mRNA and AMPK were not normalized. Moreover, pharmacologically restoring local EGP IGF 1 level only partially restored growth in AO animals 14,22 . This indicates that other pathways are involved in addition to the GH/IGF1 axis in this growth retardation. In mice, OX1R could regulate ghrelin content locally in the bone 33 . In our study, elevation of OX1R in the EGP was associated with partial improvement in EGP architecture following OR. Several signaling pathways of ghrelin play a role in chondrogenesis and osteoblastogenesis 33,57,58 . Moreover, OX1R is suppressed during osteoblast differentiation and elevated during adipocyte differentiation 33 . In this study, reduction of growth pate GHSR1α was associated with abnormal EGP architecture and growth retardation. Interestingly, EGP OX1R did not improve in the OR group, while GHSR1α was similar to the control. This finding may suggest that in EGP, GHSR1α and OX1R are independently regulated. Further studies are needed to explore the effects of sleep and caloric restriction on EGP OX1R. We found increased adipose cell number and marrow PPARγ, an essential element in adipocyte differentiation processes in many tissues 59 . Orexin activates PPARγ in the marrow and is associated with bone mass loss and increased adipogenesis 33 . Sox9 is important in chondrogenesis differentiation 60 . Summary. In this study, we showed that rapidly growing animals cannot consume enough calories to maintain energy homeostasis. This deregulation of energy availability is associated with a considerable impediment in EGP metabolism/architecture and abnormal growth in order to save energy. Thus, surgical intervention per se may not be sufficient to prevent growth retardation, and endocrine correction with one or more availability circulating factors may be essential.

Animals. This study was approved by the Ben-Gurion University of the Negev Animal Use and Care
Committee protocol number IL-40-07-2018. All protocols comply with the American Physiological Society Guidelines. Male 22-day-old Sprague-Dawley rats (48-55 gr) were used. Animals were kept on a 12-12 light-dark cycle with lights on 09:00 at 23 ± 1.0 °C. Animals were given food (3272 Kcal/kg) and water ad libitum.
Surgery. The technique used for sham surgery and to induce AO in juvenile rats was as previously described 14,15,17,47 . During the surgical procedures, the mortality rates in the AO and OR group were less than 10%, and an additional 5% mortality was observed 2-5 days after surgery. Data were collected on days 45-49 post surgery. Following surgery, prophylactic enrofloxacin 5 mg/ml (s.c.) and water containing ibuprofen (0.1 mg ml −1 ) were given for three days 14,15,17,30 . experimental schedule. In the current study, we used a previously described experimental schedule ( Supplementary Fig. S3) 17,30 . On day 14, the AO group was randomized, and OR of the silicon band was performed on half of the animals. Measurements of respiration, energy metabolism, and the effect of propranolol energy metabolism were performed on days 45-48. On day 49, animals were sacrificed, serum was collected and tissue was extracted after death and stored at −80 °C. Respiratory and energy metabolism. Respiratory activity at room air was recorded by whole body plethysmography (Buxco, DSI, St. Paul, MN, USA). ∆Pes was measured from a saline-filled catheter placed in the lower third of the esophagus and connected to a pressure transducer 22,30,44,45 . Seven animals were used to analyze esophageal pressure. Metabolic activity was measured using a behavioral phenotyping system (Sable Instruments, Las Vegas, NV, USA), as previously described 30,61 . Animals were allowed a 24 h acclimation period followed by a 48 h sampling duration. Energy expenditure was calculated as VO 2 × (3.815 + 1.232 × respiratory quotient), and was normalized to effective body mass. Resting energy expenditure was calculated as the mean value for a 30-min period with lowest energy expenditure. The respiratory quotient was calculated as the ratio of CO 2 produced by O 2 consumed by the body.
Histology. Five-μm thick longitudinal sections were cut and collected on Superfrost ™ Plus slides for histology staining of Safranin O 15 , hematoxylin, and eosin 22 . Total growth-plate width and proliferative and hypertrophic zone widths were measured as previously described (n = 8 in each group) 15,29 . Trachea histology photomicrographs were obtained by light microscope, with the internal border of the trachea outlined and cross-sectional area and diameter calculated for each animal, as previously described 17,30 . immunohistochemistry. Immunohistochemistry staining was performed using a protocol described previously by our laboratory (n = 6 in each group) 15 . Anti-rabbit OX1R, anti-mouse IGF-1 (Abcam, Cambridge, MA, USA), anti-mouse PPARγ and anti-mouse Sox9 (Santa Cruz Biotechnology, Santa Cruz, CA, USA) were used for immunohistochemistry staining. For image processing, Cellsens Entry software (MATIMOP, Tel Aviv, Israel) was used. All of the experiments and observations were repeated at least three times.
Quantitative real-time PCR. Assays were performed with power SYBR green PCR master mix (Applied Biosystems) as previously described 15,29 using the ABI Prism 7300 Sequence Detection System (Applied Biosystems).