It has recently been suggested that the low-density lipoprotein receptor-related protein 5 (LRP5) regulates bone mass by suppressing secretion of serotonin from duodenal enterochromaffin cells. In mice with targeted expression of a high bone mass–causing (HBM-causing) LRP5 mutation and in humans with HBM LRP5 mutations, circulating serotonin levels have been reported to be lower than in controls whereas individuals with loss-of-function mutations in LRP5 have high blood serotonin. In contrast, others have reported that conditionally activating a knock-in allele of an HBM-causing LRP5 mutation in several tissues, or genetic deletion of LRP5 in mice has no effect on serum serotonin levels. To further explore the possible association between HBM-causing LRP5 mutations and circulating serotonin, levels of the hormone were measured in the platelet poor plasma (PPP), serum, and platelet pellet (PP) of 16 affected individuals from 2 kindreds with HBM-causing LRP5 mutations (G171V and N198S) and 16 age-matched controls. When analyzed by HPLC, there were no differences in levels of serotonin in PPP and PP between affected individuals and age-matched controls. Similarly, when analyzed by ELISA, there were no differences in PPP or PP between these two groups. By ELISA, serum levels of serotonin were higher in the affected individuals when compared to age-matched controls. A subgroup analysis of only the G171V subjects (n = 14) demonstrated that there were no differences in PPP and PP serotonin between affected individuals and controls when analyzed by HPLC. PP serotonin was lower in the affected individuals when measured by ELISA but serum serotonin levels were not different. We conclude that there is no change in PPP serotonin in individuals with HBM-causing mutations in LRP5. © 2014 American Society for Bone and Mineral Research.
Low-density lipoprotein receptor-related protein 5 (LRP5), which is a co-receptor for Wnt proteins, is known to have an important role in skeletal metabolism. Loss-of-function mutations in LRP5 lead to osteoporosis pseudoglioma (OPPG), which is a disorder characterized by low bone mass, low bone formation rates, and blindness due to continued embryonic eye vascularization. There are also mutations in LRP5 that lead to high bone mass.[3-5] It has recently been suggested that LRP5 regulates bone mass in part by modulating circulating levels of serotonin, but other experimental data are at odds with this hypothesis.[7, 8] Thus, whether serotonin is a key downstream effector mediating the effects of LRP5 in bone remains controversial.
In 2008, Yadav and colleagues reported evidence that serotonin produced by duodenal enterochromaffin cells is an endocrine regulator of bone mass. They found that gut-specific expression of a cDNA for a high bone mass–causing (HBM-causing) mutant LRP5 led to a reduction in serum serotonin by inhibiting tryptophan hydroxylase 1 (Tph1) expression, the rate-limiting enzyme in serotonin synthesis. In 2 individuals with an HBM-causing LRP5 mutation, serotonin levels in platelet poor plasma (PPP) were 50% lower than in age-matched controls. In addition, 3 subjects with OPPG had increased serum serotonin levels compared to age-matched controls. In mice, serotonin was shown to directly suppress osteoblast function. On the basis of these data, these investigators propose that signaling through LRP5 modulates the enterochromaffin cell's production of serotonin, which acts as an endocrine hormone to control bone formation.
In 2010, Frost and colleagues reported that levels of PPP serotonin were significantly lower in a Danish family with a different HBM-causing mutation in LRP5 (T253I). In 2011, Frost and colleagues also showed that serum serotonin levels were lower in 19 subjects with an HBM-causing mutation in LRP5 (T253I) compared to 19 age-matched and sex matched controls. Further support for the hypothesis that peripheral serotonin regulates bone mass was provided by a cross-sectional study in 275 women in whom serum serotonin levels were inversely correlated with bone mass.
In contrast to the data summarized above, Cui and colleagues have provided evidence that serotonin does not play a role in regulating the effects of LRP5 on bone but instead that LRP5 acts locally in the skeleton. In mice, conditionally activating a knock-in mutant allele of LRP5 in the appendicular skeleton increased bone mass only in the limbs but not in the spine. Furthermore, intestine-specific activation of HBM-causing LRP5 mutations had no effect on bone mass. They also did not observe any differences in serum serotonin levels among HBM LRP5 knock-in, knockout, or wild-type mice. Recently, Chang and colleagues found that serum levels of serotonin were no different in LRP5–/– mice when compared to controls.
In view of these conflicting data and because a variety of methodologies have been used to measure circulating serotonin, as well as the difficulties in measuring this hormone, we revisited the relationships between bone mass and serotonin levels in PPP, serum, and platelet pellet (PP). We analyzed serotonin levels in affected individuals from two kindreds with HBM-causing LRP5 mutations as well as in age-matched controls using both ELISA and high-performance liquid chromatography (HPLC) methodologies.
Subjects and Methods
Primary and secondary outcomes
Our primary outcome was to determine if there was a statistically significant difference between PPP serotonin levels in affected and control groups when measured by ELISA or HPLC. Our secondary outcomes were to determine if there were statistically significant differences between these groups in serum serotonin and PP serotonin by ELISA or in PP by HPLC.
Fourteen subjects with the G171V HBM-causing LRP5 mutation were recruited from our previously reported kindred. Two subjects with an N198S mutation in LRP5 associated with HBM were also studied. PPP serotonin levels for 2 individuals with HBM-causing LRP5 mutations were previously reported and were resampled for the current study. We also recruited 16 healthy age-matched individuals not from these two kindreds. These volunteers were not taking any antidepressants. The study complied with the World Medical Association Declaration of Helsinki–Ethical Principles for Medical Research Involving Human Subjects. The study was approved by the Yale Human Research Protection Program Institutional Review Board (IRB). All subjects gave written consent to participate in the study.
Sample collection and preparation
The majority of the subjects had their blood collected in the morning after an overnight fast. Two subjects were not fasting (1 subject with the G171V mutation and 1 age-matched control). Blood for serotonin measurements in PPP and PP was collected in tubes containing EDTA (5.4 mg per 3 mL of blood) and ascorbic acid at a final concentration of 7.5 mM and then immediately placed on ice. Blood for serum serotonin measurements was collected in untreated tubes and allowed to clot on ice for 30 to 60 minutes.
The blood collected in EDTA was processed within 3 hours of collection using the following protocol. The blood was centrifuged at 100g for 15 minutes at room temperature. The supernatant, which is the platelet-rich-plasma, was removed and transferred to a new polypropylene tube. This was chilled on ice for 10 minutes before repeat centrifugation for 6 minutes at 4°C at 14,500g. This separated the specimen into a PP and PPP. The PPP was aliquotted into Eppendorf tubes. The PP was resuspended in 1 mL of saline for analysis. Specimens were stored at –80°C until analyzed. Samples were analyzed within 3 months of collection.
An ELISA (Serotonin ELISA; Immuno-Biological Laboratories, Minneapolis, MN, USA) was used to measure serotonin levels in the PP and the serum. The intraassay and interassay coefficients of variation (CV ± SE) were 9.1% ± 2.4% and 25.9% ± 5.5%, respectively. The CVs for this assay were generated in the Yale Mineral Metabolism Laboratory.
An ultrasensitive ELISA (Serotonin Research ELISA; Labor Diagnostika Nord GmbH & Co. KG, Nordhorn, Germany) was used to measure serotonin levels in the PPP. The intraassay and interassay coefficients of variation (CV ± SE) were 15.2% ± 0.3% and 26.8% ± 9.7%, respectively. The CVs for this assay were also generated in the Yale Mineral Metabolism Laboratory.
HPLC with fluorimetric detection was used to measure serotonin levels in the PPP and PP. Samples were applied to a C18 reverse-phase column (AppliChrom Application & Chromatography, Germany) and eluted at 20°C in 10 mM potassium phosphate buffer, pH 5.0, containing 5% methanol at a flow rate of 2 mL/min. Fluorescence of 5HTP and 5-HT was excited at 295 nm and measured at 345 nm.
Five subjects in the kindred bearing the G171V LRP5 mutation had not been previously genotyped. Therefore, PCR amplification of genomic DNA and Sanger sequencing was used to confirm the presence of the mutation in these 5 individuals. Primers located in intron 2 and intron 3 were designed using Primer 3 software. The forward primer used was 5′-GGCAGGAATACCTGAAACCA-3′. The reverse primer used was 5′-TGACGCTGTTCCAAGTTCTG-3′.
Bone mineral density measurement
Subjects were asked to provide their most recent dual-energy X-ray absorptiometry (DXA) bone mineral density measurements of the lumbar spine (L1–L4) and femoral neck. If a DXA study had not been performed previously, subjects were scanned at the Yale Bone Center using an Hologic 4500C machine densitometer.
Student's two-tailed t test was used for all statistical analyses.
The baseline characteristics of the two study groups are shown in Table 1. The mean lumbar spine (L1–L4) T-scores in the affected and age-matched control groups were +5.11 ± 0.68 and +0.34 ± 0.71, respectively (mean ± SEM). The mean lumbar spine (L1–L4) Z-scores in these same groups were +5.89 ± 0.43 and +1.79 ± 0.62, respectively. The mean femoral neck T-scores were +4.36 ± 0.71 and –1.00 ± 0.44, respectively. The mean femoral neck Z-scores were +5.11 ± 0.60 and +0.32 ± 0.34, respectively. The mean body mass index (BMI) values in these subjects were 28.2 ± 2.2 and 26.0 ± 1.4, respectively (mean ± SEM). The affected group included 14 subjects with the G171V mutation and 2 subjects with an N198S mutation.
|Variable||Affected (n = 16)||Age-matched controls (n = 16)|
|Age (years)||59.3 ± 5.4||60.4 ± 5.3|
|BMI (kg/m2)||28.2 ± 2.2||26.0 ± 1.4|
|Femoral neck T-score||+4.36 ± 0.71||–1.00 ± 0.44|
|Femoral neck Z-score||+5.11 ± 0.60||+0.32 ± 0.34|
|L1–L4 T-score||+5.11 ± 0.68||+0.34 ± 0.71|
|L1–L4 Z-score||+5.89 ± 0.43||+1.79 ± 0.62|
The mean serotonin level of PPP in the affected individuals was not different from that in the age-matched control group when measured either by ELISA (10.8 ± 1.5 versus 14.9 ± 2.6 ng/mL; affected versus control; p = 0.19; Fig. 1A) or HPLC (7.8 ± 1.1 versus 7.5 ± 1.1 ng/mL; affected versus control; p = 0.87; Fig. 1D). Serotonin measurements in PP were also not significantly different in the affected and control groups when measured either by ELISA (882.8 ± 219.0 versus 1349.0 ± 222.1 ng/mL; affected versus control; p = 0.15; Fig. 1B) or by HPLC (964.1 ± 152.7 versus 945.1 ± 123.0 ng/mL, affected versus control; p = 0.92; Fig. 1E). By ELISA, mean serum serotonin levels were significantly higher in the affected group than in the controls (232.0 ± 35.6 versus 136.9 ± 21.8 ng/mL; affected versus control; p = 0.03; Fig. 1C).
The kindred with the N198S mutation has been previously reported. We were only able to recruit 2 subjects from that kindred and we therefore reanalyzed the data excluding these 2 subjects. When measured by ELISA, PPP serotonin and serum serotonin were not different in individuals with the G171V mutation as compared to age-matched controls (Fig. 2A, C). By ELISA, PP serotonin was lower in affected individuals than age-matched controls (646.1 ± 156.8 versus 1427.0 ± 243.4 ng/mL; affected versus control; p = 0.01; Fig. 2B). When analyzed by HPLC, there were no statistically significant differences in PPP or PP serotonin levels in the two groups (Fig. 2D, E). Because there were only 2 subjects with the N198S mutation, a comparison to the subjects with the G171V mutation could not be performed. The PPP serotonin values in the 2 subjects previously reported by Yadav and colleagues were found to be 7.0 ng/mL and 10.8 ng/mL by HPLC and 17.2 ng/mL and 18.9 ng/mL by ELISA in the current study. Yadav and colleagues reported values of approximately 1.5 ng/mL and 1.3 ng/mL, respectively, for these 2 subjects.
We also performed an analysis of the subjects with the G171V mutation restricted to only those subjects who were fasting (excluding 1 G171V subject and 1 age-matched control subject). We found that by HPLC, there were no statistically significant differences between the groups when measuring PPP or PP. By ELISA, again there was no difference in PPP serotonin or serum serotonin. PP serotonin levels in the G171V subjects were lower than age-matched controls (mean ± SEM; 658.4 ± 168.9 versus 1477.0 ± 257.2 ng/mL; affected versus control; p = 0.01), which is the same result found in the analysis including all G171V subjects and all controls.
LRP5 is known to have an important role in regulating bone mass.[1-3] Apart from the dramatic but rare skeletal phenotypes associated with point mutations in the first beta propeller loop or loss-of-function mutations, this gene has been identified as one of only a handful consistently identified as contributing to the hereditability of bone mass in genomewide association studies (GWASs).[2, 4, 13] Therefore, the mechanism by which LRP5 acts to regulate skeletal metabolism remains an area of considerable investigative and clinical interest.
Evidence summarized in the Introduction suggesting that serotonin produced by duodenal enterochromaffin cells is an endocrine mediator of LRP5's actions in bone has generated great interest because, if correct, it would provide an attractive therapeutic target for treating low bone mass. As noted, based on animal and some human studies, the proposal has been advanced that circulating serotonin negatively regulates bone formation.[6, 10-12] Therefore, inhibiting the peripheral production of serotonin could be of considerable therapeutic benefit. In particular, orally active inhibitors of serotonin synthesis, which do not cross the blood-brain barrier, have already been developed.[14, 15] Inose and colleagues showed that an oral small molecule inhibitor of tryptophan hydroxylase 1 (Tph1) prevented the low bone mass phenotype of LRP5 knockout mice and reversed ovariectomy-induced bone loss in mice by stimulating bone formation without affecting resorption. In a phase 2 study involving patients with the non-constipating form of irritable bowel syndrome (IBS), LX1031, the oral locally acting, small-molecule inhibitor of tryptophan hydroxylase relieved symptoms and increased stool consistency. A phase 2b study using a second generation of this locally acting TPH inhibitor, LX1033, is ongoing (http://clinicaltrials.gov;Identifier:NCT01494233). The availability and apparent safety of these drugs certainly makes it feasible to manipulate serotonin levels in patients with low bone mass.
Measurement of serotonin in plasma is difficult for a number of reasons. First among them is that serotonin circulates in very low concentrations, with the vast majority present in platelets. Considerable care must be taken to avoid activating platelets when measuring serotonin in extracellular fluids.[16, 17] These two difficulties have led to considerable variation in results and may in part have contributed to the current uncertainty about the levels of circulating serotonin in high bone mass murine models of LRP5.[6, 7] As summarized in the Introduction, there have been four previous reports describing a relationship between LRP5 function and circulating levels of serotonin in humans, which are consistent with the conclusion that circulating serotonin regulates bone mass.[6, 10-12] However, as also noted in the Introduction, studies in experimental animals by Cui and colleagues and Chang and colleagues have led to the conclusion that circulating serotonin does not regulate bone mass. Because, ultimately, results in humans are the most clinically relevant, we sought to revisit this question in kindreds with HBM-causing mutations in LRP5.
Given the above noted difficulties in measuring serotonin, we used two different methodologies. One was a high-sensitivity ELISA that is was used to measure concentrations in PPP. This was necessary because the PPP serotonin values fell below the lowest standard in the standard curve of the assay used by prior investigators (Serotonin ELISA; Immuno-Biological Laboratories; Minneapolis, MN, USA). The other method we used, HPLC, has been widely used for measuring PPP serotonin. This is a complex technique but several of the authors (SM, NA, and MB) have had considerable experience with this methodology.
By both of these methods, there were no statistically significant differences in PPP and PP serotonin between affected individuals with HBM-causing mutations and controls. No prior studies have used the ultrasensitive ELISA to measure circulating serotonin in humans with HBM LRP5 mutations so our results using this assay cannot be compared to those of earlier investigators. Further, we added ascorbic acid, which stabilizes serotonin, to the blood used for measuring PPP serotonin. To our knowledge, this was not done by other investigators. HPLC was not used in earlier studies in patients with HBM LRP5 mutations so again it is difficult to compare our results to prior studies. We did use the less sensitive ELISA (Serotonin ELISA) to measure serum serotonin levels. The results in our control subjects are comparable to those reported by Frost and colleagues and Mödder and colleagues, although somewhat higher than the values in control subjects reported by Yadav and colleagues. Our control group was well-matched for age to our study population although the male:female ratio was not equivalent in the two groups (Supplemental Table 1). We do not believe this materially affected our results because it has been previously published that sex does not influence circulating levels of serotonin. Thus, results from two very different assays do not support the notion that levels of serotonin in PPP are consistently lower in patients with HBM-causing mutations in LRP5 as compared to levels in age-matched controls. In addition, the results in PP serotonin and serum serotonin do not indicate a consistent difference between the two groups. (Although serum serotonin was significantly higher in the individuals with HBM-causing mutations, this was because the serum serotonin was considerably higher in the two individuals bearing the N198S mutation.)
Because we were only able to recruit 2 individuals from the N198S kindred, we undertook a subanalysis restricted to subjects with the G171V mutation. In this subanalysis by HPLC, there were no statistically significant differences in PPP or PP serotonin. The reanalysis of the ELISA data demonstrated that PP serotonin levels were significantly lower in subjects with the G171V mutation than in age-matched controls (p = 0.01) (Fig. 2B). This is because the 2 N198S subjects had high PP serotonin levels (average value = 2540.0 ng/mL in N198S versus 646.1 ± 156.8 ng/mL in G171V). In the reanalysis restricted to the subjects with the G171V mutation, serum levels of serotonin were not different in the affected individuals and controls, unlike the analysis of the combined group in which serum serotonin was higher in affected individuals. Again, this is because the 2 subjects with the N198S mutation had higher levels of serum serotonin (average value = 503.5 ng/mL in N198S versus 193.2 ± 26.5 ng/mL in G171V). Frost and colleagues[10, 11] studied a kindred with a different HBM-causing LRP5 mutation (T253I) than either of those reported here. Because we noted possible differences between the G171V and N198S mutations, at least by ELISA, this raises the interesting possibility that different mutations may differentially affect serotonin metabolism.
The differences in our findings compared to those in the studies by Yadav and colleagues and Frost and colleagues[10, 11] may be due to several factors. These investigators did not use a high-sensitivity ELISA to measure PPP serotonin and, thus all of their results fell below the lowest value in the standard curve requiring extrapolation of the results. As noted, we added ascorbic acid to stabilize serotonin in the preparation of the PPP samples. Frost and colleagues studied HBM volunteers with a different genotype and perhaps some of the divergent findings could be explained by that fact. The findings using serum serotonin measurements reported by previous investigators are inherently limited.[6, 11, 12] During the preparation of the serum sample, clot formation occurs resulting in variable degrees of platelet aggregation and activation and in turn greater variability in serum serotonin measurements. Thus, the relevance of serum measurements to the hypothesis that LRP5 regulates serotonin production by enterochromaffin cells is not clear.
Our study also has limitations. The number of affected study subjects was relatively small. However, Frost and colleagues saw significant differences in levels of PPP with fewer numbers of subjects. Our control group was also relatively small, although it was carefully age-matched.
In summary, our data do not support the notion that there are consistent changes in circulating levels of serotonin in patients bearing HBM-causing LRP5 mutations. The validity of the model suggesting that LRP5 mediates its actions on bone via serotonin should be testable in the near future, because orally available selective serotonin inhibitors are already being used in phase 2 clinical trials.
All authors state that they have no conflicts of interest.
This work was supported by a grant from the Women's Health Research at Yale Program and by the Yale Bone Center. This work was also made possible by CTSA Grant Number UL1 RR024139 from the National Center for Research Resources (NCRR) and the National Center for Advancing Translational Science (NCATS), components of the National Institutes of Health (NIH), and NIH roadmap for Medical Research. Its contents are solely the responsibility of the authors and do not necessarily represent the official view of NIH. This work is dedicated to John M. Packard, MD. Without his commitment to this project, it would have been impossible to complete. We are also deeply indebted to the families and normal volunteers who so generously agreed to participate in this study.
Authors' roles: Study design: KI and GL. Data collection: KI, GL, DF, CS, BHS, YC, JB, BS, NA, SM, and MB. Data analysis: KI, GL, CS, YC, NA, SM, and MB. Data interpretation: KI and GL. Drafting manuscript: KI and GL. Approving final version of manuscript: KI, GL, NA, SM, and MB.