The authors have declared that no competing interests exist.
Conceived and designed the experiments: VKB GCS EKH KAMP. Performed the experiments: VKB NPS EKH KAMP. Analyzed the data: VKB GCS CSC EJO MCZ JLM EKH KAMP. Contributed reagents/materials/analysis tools: MCZ JLM RJA KAMP. Wrote the paper: VKB EKH KAMP.
The increasing use of the common marmoset (
The common marmoset (
Gastrointestinal (GI) disease, including Marmoset Wasting Syndrome (MWS), and bone disease are endemic in many captive marmoset colonies and result in significant mortality. MWS, commonly referred to as “chronic wasting,” is one of the most important non-infectious marmoset diseases
Another significant cause of mortality in captive marmoset colonies is bone disease, particularly metabolic bone diseases (MBDs). MBDs can be characterized by fibrous osteodystrophy (classic MBD), osteopenia, or rickets, and are caused by high bone turnover secondary to calcium-phosphorus imbalance, malabsorption, vitamin D deficiency, and/or excess parathyroid hormone (PTH)
Bone disease and GI disease, both individually and in combination, have devastating effects on marmoset health and the potential to hinder long-term research projects when diagnosed late in the clinical course of disease. We hypothesize that bone disease and GI disease are associated in marmosets, and that antemortem biomarkers indicative of GI health and bone metabolism correlate with histopathologic findings. Early and accurate identification of marmosets predisposed to or in the early stages of disease would preserve research integrity by allowing for the early exclusion of affected animals from long-term projects. In this study, we establish that bone and GI disease are associated in marmosets, and multiple antemortem biomarkers, including serum albumin, body weight, radiographs, and parathyroid hormone, can be used to distinguish between affected and unaffected individuals.
All experimental procedures were approved and overseen by the Institutional Animal Care and Use Committee of Johns Hopkins University. Strict adherence to the Guide for the Care and Use of Laboratory Animals by the National Institutes of Health, the Animal Welfare Act by the United States Department of Agriculture, and the Weatherall Report by the Medical Research Council was observed.
All marmosets were housed in family units, in pairs, or singly as part of a large breeding and experimental colony at Johns Hopkins University, an Association for Assessment and Accreditation of Laboratory Animal Care International (AAALAC)-accredited institution. Animals were given ad libitum access to water and were fed a complete and balanced diet consisting of a custom homogenized blend of Teklad 8794N New World Primate Diet (Harlan Laboratories, Indianapolis, IN), Zupreem 9920.CS canned marmoset diet (Shawnee, KS), and Bio-Serv Newberne Hayes Vitamin Mix (San Diego, CA), supplemented with fruit and yogurt. Environmental enrichment in the form of visual and auditory contact with conspecifics, complex environments, and manipulanda were provided to all marmosets. Experienced animal care technicians observed animals at least once daily and more often if necessitated by experimental or medical need. The Johns Hopkins University Research Animal Resources veterinarians diagnosed, treated, and/or managed any medical illness or injury in the marmosets, and if euthanasia due to disease was warranted, marmosets were euthanized with an intravenous overdose of Euthasol (Virbac Corporation, Fort Worth, TX) under deep ketamine anesthesia. Of the 105 marmosets evaluated for this study, 62 marmosets served as colony breeders or were not on experimental study. The remaining 43 marmosets participated in neurobehavioral studies, of which 33 had surgically implanted headcaps
Bone and GI disease statuses for all animals in this study were determined by histopathologic analysis of postmortem tissues rather than the presence of clinical disease. Hematoxylin and eosin-stained slides from paraffin blocks containing bone and gastrointestinal tissue from 105 marmosets necropsied between October 1999 and February 2012 were examined by veterinary pathologists (Diplomates of the American College of Veterinary Pathologists). Animals were diagnosed as having abnormal bone if any of the following lesions were identified: abnormal growth plate (disorganized or abnormally oriented chondrocytes, thickened or fractured trabecular bones), abnormal cortices (thin – osteopenia, fractured without prior history of trauma, increased numbers of resorption cavities with increased osteoclastic +/– osteoblastic activity), and/or myelofibrosis. Animals were diagnosed as having GI disease if they had chronic mild inflammation in two or more segments or chronic moderate or severe inflammation in one or more segments of small or large intestine, including cecum. Animals with severe intestinal necrosis consistent with
Data on weight and bloodwork parameters (including albumin) were retrospectively obtained from clinical and laboratory records from deceased marmosets and through veterinary physical exam of live marmosets. Terminal weight records were obtained from catalogued necropsy reports, and historically obtained bloodwork data were collected from clinical records. All live marmosets in the colony underwent a yearly physical exam while sedated with 20 mg/kg ketamine, which included body weight, superficial dental examination, and fur coat quality check. Blood was collected from the femoral vein, and a complete blood count and serum chemistry panel were performed using a Hemavet 950 hematology analyzer (Drew Scientific Inc, Oxford, CT) and vetACE chemical analyzer (Alfa Wassermann, West Caldwell, NJ), respectively. All bloodwork data used in this study were collected within one year of death, and in cases of multiple available records, the data collected closest to the date of death were included in analyses. Progressive weight data were collected by the lab as part of a standard experimental protocol from five animals that died due to natural causes and were diagnosed with GI disease +/– bone disease at necropsy. Weight data from five age- and gender-matched animals diagnosed with neither bone nor GI disease at necropsy were used for comparison. Data were graphed as percentage peak body weight versus days before death, and the slopes of the best-fit lines for each of the animals’ body weights were calculated.
Marmosets were sedated with 20 mg/kg ketamine and were positioned in a standardized manner using a restraint board prior to taking ventrodorsal radiographs with the FCR XC-2 digital radiography system (Fujifilm, Valhalla, NY). A mammographic aluminum stepwedge with 9 sequential steps of radiodensity (Gammex, Inc, Middleton, WI) was included in all radiographs. Images were imported into an image analysis software package (Elements, Nikon Imaging Software). The distal 25% of both femurs were gated as regions of interest, and the bone radiodensity fraction (BRF) within the region of interest was determined using the stepwedge as a standard. These values ranged from zero to one, with a BRF closer to one representative of more radiopaque (dense) bone.
Serum markers of bone disease were measured using commercially available enzyme-linked immunosorbent assays (ELISA) for bone alkaline phosphatase (Quidel Corporation, San Diego, CA), parathyroid hormone (ALPCO Diagnostics, Salem, NH), and serum cross laps (Immunodiagnostic Systems, Scottsdale, AZ), while serum and fecal markers of gastrointestinal disease were measured using ELISAs for C-reactive protein (Life Diagnostics, West Chester, PA), secretory IgA (ALPCO Diagnostics, Salem, NH), and calprotectin (Buhlmann Laboratories, Schoenenbuch, Switzerland). Additional biomarkers specific for bone or GI disease were examined and found not to cross-react with marmoset samples (
Due to insufficient numbers of unaffected marmosets in some categories to prove normal data distribution, nonparametric statistical analyses were used. Two-tailed Mann Whitney U tests were used to compare disease and non-diseased groups, and Spearman’s rank (two tailed) was used to measure correlation between data. Two-tailed Fisher’s exact (FE) tests were used to calculate the odds ratio, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV). A
Both bone disease and GI disease are common causes of morbidity and mortality in captive marmoset colonies, and we sought to elucidate whether diseases in these two organ systems were associated. We reviewed 105 marmoset necropsy records, with both bone and GI tissue available from 53 of these cases. Twenty-nine of these marmosets (55%) had both bone disease and GI disease, three (6%) had bone disease only, twelve (22%) had GI disease only, and nine (17%) had neither bone nor GI disease (
(A) Serum albumin and (B) body weight of marmosets diagnosed with BGS or no disease at necropsy. Solid horizontal lines represent median values for each group of data points, dotted horizontal lines denote cutoff values suitable to identify animals with BGS. (C) Correlation between body weight and albumin levels in individual animals. Solid black line represents best-fit line of the data points. Red data points denote animals with headcaps, gold data points denote animals on restricted diets.
Despite numerous case reports describing the clinical course of disease and postmortem findings of either bone or GI disease in marmosets
Of the clinical chemistry parameters examined in a subset of 31 marmosets, serum albumin differed significantly between animals with BGS compared to unaffected animals (Mann-Whitney [MW]
Test | Sensitivity | Specificity | PPV | NPV | ||
Serum Albumin < 3.5 g/dL | 90% | 100% | 100% | 67% | 0.046 | |
Body Weight < 325 g | 87% | 100% | 100% | 75% | 0.00050 | |
Serum Albumin < 3.5 g/dL or body weight < 325 g | 100% | 100% | 100% | 100% | 0.036 | |
Weight loss > 0.05% peak weight per day | 100% | 100% | 100% | 100% | 0.018 | |
Serum Albumin < 3.5 g/dL | 83% | 100% | 100% | 75% | 0.0015 | |
Body Weight < 325 g | 81% | 60% | 68% | 75% | 0.029 | |
Serum Albumin < 3.5 g/dL or body weight < 325 g | 86% | 67% | 86% | 67% | ||
Weight loss > 0.05% peak weight per day | 100% | 83% | 75% | 100% | 0.048 | |
BRF < 0.5 | 83% | 100% | 100% | 80% | 0.048 | |
PTH > 600 pg/mL | 75% | 100% | 100% | 73% | 0.0014 | |
Serum Albumin < 3.5 g/dL | 70% | 75% | 88% | 50% | 0.044 | |
Body Weight < 325 g | 75% | 93% | 96% | 59% | < 0.00010 | |
Serum Albumin < 3.5 g/dL or body weight < 325 g | 92% | 100% | 100% | 80% | 0.0027 | |
Weight loss > 0.05% peak weight per day | 100% | 100% | 100% | 100% | 0.0079 |
= positive predictive value, NPV = negative predictive value, BRF = bone radiodensity frequency, PTH = parathyroid hormone. PPV
Italicized
Serum calcium is bound by albumin in the blood, and values reported in serum chemistry panels can therefore be confounded by alterations in serum albumin level
In addition to hematology and serum chemistry parameters, we examined whether physical exam data, such as body weight, could be used to identify animals with BGS. Adult body weights in captive marmoset colonies range from 250 to 600 g, with most animals weighing 350 to 400 g
Negative correlations between marmoset age and serum albumin and body weight have been previously reported, with older adult marmosets generally having lower serum albumin levels and lower body weights compared to younger adult marmosets
Thirty-one percent of the marmosets included in this investigation were involved in neurobehavioral research that required the use of headcaps (identified by red data points in
While body weight and serum albumin provide a way to identify BGS prior to death, we wanted to identify a marker that could predict development of disease well before the terminal stage. Multiple sequential weight data were collected from five animals diagnosed with GI +/– bone disease at necropsy and five unaffected animals. Peak weight of each animal was determined, percentage of peak weight was calculated for all data available one year prior to death, and data were graphed as percentage peak body weight versus days before death. (
Percent of peak body weight of five unaffected (A) and five marmosets with GI +/– bone disease (B) for the year prior to death. Colors represent data points from individual animals, black lines represent best-fit line of the data points.
Marmosets with GI +/– bone disease trended towards lower peak weights than unaffected animals (MW
Affected | Unaffected | P value | |
5 | 5 | ||
3:2 | 2:3 | ||
5.0 (2.2 8.6) | 4.0 (2.8, 4.8) | ||
5:0 | 0:5 | ||
3:1 |
0:5 | ||
386 (304, 462) | 484 (386, 569) | ||
39.9 (23.5, 54.7) | 0.5 (–1.3, 5.1) | 0.0079 | |
–0.17 (–0.19, –0.14) | +0.036 (0.015, 0.058) | 0.0079 |
Data presented as median values with parentheses following denoting 95% confidence intervals of the mean.
Italicized
Bone tissue slides from one affected marmoset were nondiagnostic.
Percent weight change per day represented as the median slope and 95% confidence interval of that slope.
Two marmosets included in the above weight trend analyses, M84 and M104, were initially kept on diets targeted to maintain them at 90% of their free feeding weight as described above until four months prior to death; when persistent weight loss was noted, the diets were discontinued. The average percent peak weight loss per day actually increased following termination of the restricted diet for each animal (0.068% vs 0.24% for M84 and 0.094% vs 0.18% for M104, during and after cessation of the defined diet, respectively). Therefore, the restricted diet did not accelerate weight loss, and return to a free choice-feeding regimen did not halt, slow, or reverse the progression of disease.
We examined whether digital radiographs could quantify bone density and identify marmosets with bone disease. As part of our annual physical examination of the colony, ventrodorsal radiographs were taken with a digital radiography system, and the bone radiodensity fraction (BRF) from the images was determined using a stepwedge standard (
(A) Example of the initial radiographic image (left panel) and analyzed image (right panel) of an animal with bone disease, with BRF calculated from the distal quarter of each femur (white arrows). (B) BRF values in marmosets diagnosed with bone disease or with no bone disease at necropsy. (C) PTH levels in marmosets diagnosed with bone disease or with no bone disease at necropsy. BRF = bone radiodensity fraction, PTH = parathyroid hormone. Solid horizontal lines represent median values for each group of data points, dotted horizontal lines denote cutoff values to distinguish animals with bone disease.
Biochemical parameters specific to either the skeletal or the GI system were also examined as potential antemortem tests for BGS. Three serum markers specific for disease of the skeletal system were examined: bone alkaline phosphatase (BAP), parathyroid hormone (PTH), and carboxy-terminal collagen crosslinks (serum cross laps or SCL). BAP is an enzyme produced by osteoblasts and osteoclasts that is used as an indicator of general bone turnover
One serum and two fecal biochemical parameters specific for disease of the GI system were examined. C-reactive peptide (CRP) is a serum acute phase inflammatory protein that is used to detect systemic inflammatory disease, including GI disease, in both humans and nonhuman primates
Marmosets are commonly employed for long-term research studies, and development of confounding disease after a study is underway results in a considerable waste of effort and funds. Early identification of diseased animals prior to the terminal stage could prevent some of these losses and allow labs to remove animals from study or relegate them to acute procedures. Bone disease and GI disease are two conditions with unknown causes commonly found in captive marmosets, and in the present study we demonstrate that bone and GI disease are associated. We also establish retrospectively that low body weight (< 325 g) and low serum albumin (< 3.5 g/dL) are sufficient to identify animals affected with BGS, and progressive weight loss of more than 0.05% body weight per day is predictive for the later development of disease. We furthermore show that quantitative analysis of digital radiographs and serum PTH levels can be used to identify marmosets with bone disease.
The power of this panel of biomarkers lies in its accessibility to the research, zoo, and veterinary communities. The equipment necessary to measure body weight, collect blood, and take radiographs is commonly found in animal facilities, and the tests are economical to perform. When used together, body weight and serum albumin offer a powerful method for identifying marmosets with BGS or GI disease, with 100% PPV and NPV for BGS.
Body weight has proved to be a strong indicator of affected animals and an effective predictor of disease development. Weight loss has been identified as characteristic of MWS, but other than a comment in one study of a marmoset with osteomalacia
Marmosets with BGS or bone disease possessed considerably lower serum albumin levels compared to unaffected animals, and a 3.5 g/dL cutoff could significantly distinguish between affected and unaffected animals. Hypoalbuminemia has been reported in humans with osteopenia
We additionally aimed to evaluate conventional radiography as a quantitative method to evaluate bone density for the identification of marmosets with bone disease. Radiological imaging provides a way to evaluate the skeletal system in a live animal. Skeletal integrity in marmosets has been evaluated by several modalities both ante- and postmortem, including conventional X-ray, computed tomography (CT), and dual energy X-ray absorptiometry (DEXA)
We originally evaluated a number of biochemical parameters specific for either bone or GI disease in humans, such as serum BAP, PTH, SCL, CRP, fecal sIgA, and fecal calprotectin, to determine if they could identify marmosets with BGS better than the more general measures of body weight, albumin or radiographs. Of the tests examined, none of these markers proved sufficient to identify BGS, and only PTH levels were found to significantly differ between marmosets with bone disease and without bone disease. PTH increases calcium release from bone into the blood through indirect stimulation of osteoclasts, making it a good indicator of bone turnover
While investigating the mechanism of BGS is beyond the scope of this project, our findings support an association between bone and GI disease, with marmosets in our colony 7.25 times more likely to have lesions in both skeletal tissue and GI tissue than in only one tissue or the other. Such a link between bone and GI disease supports the pathogenesis recently proposed by Jarcho
The widespread prevalence of this spontaneous disease in captive marmosets presents the opportunity to evaluate BGS as a potential model for human disease, similar to how another New World monkey, the cotton top tamarin (
At this time, bone and/or GI disease in marmosets remain untreatable. If GI disease truly precedes bone disease in marmosets as we suspect, early identification of affected animals may facilitate treatment and prevention of BGS. Several accounts report on therapeutic interventions that have reversed progression of MWS
In conclusion, we have found that bone disease and GI disease are indeed associated in marmosets, and the noninvasive antemortem tests of serum albumin and body weight can be used to identify affected animals prior to the terminal stage, especially when used in tandem. Furthermore, progressive weight trends can predict which animals will develop BGS prior to the terminal stage of disease, allowing for removal of these animals from experimental studies before significant long-term investments have been made. The inciting factor(s) that cause BGS remain unknown, and at this time no consistent, effective therapies are available to slow or reverse the disease process. Now that reliable antemortem biomarkers for BGS have been identified, further examination of the pathogenesis of BGS and the marmoset’s potential as a model for human bone and/or GI disease is warranted.
(TIFF)
(TIF)
(DOC)
(DOC)
(DOC)
We would like to thank Xiaoqin Wang and his students for allowing us to collect the clinical data featured in this study from their research marmosets, and Xuhang Li for consultation on biomarkers of GI disease. We also thank Melanie Albano, Jenny Estes, Zachary Freeman, Caroline Garrett, Tracey Graham, Kristy Koenig, Theresa Meade, and Kelly Rice for their assistance in collecting samples analyzed as part of these experiments. Additionally, we are indebted to all of the veterinary pathologists and pathology trainees who performed necropsies and maintained the pathology database.