Serologic Evaluation of Clinical and Subclinical Secondary Hepatic Amyloidosis in Rhesus Macaques (Macaca mulatta)

Abstract

Secondary hepatic amyloidosis in nonhuman primates carries a grave prognosis once animals become clinically ill. The purpose of this study was to establish serologic parameters that potentially could be used to identify rhesus macaques undergoing subclinical development of secondary hepatic amyloidosis. A retrospective analysis was completed by using serum biochemical profiles from 26 histologically diagnosed amyloidotic macaques evaluated at 2 stages of disease, clinical and subclinical (3 to 32 mo prior to clinical signs of disease). Standard serum biochemistry values for cases were compared with institutional age- and gender-specific references ranges by construction of 95% confidence intervals for the difference between means. In addition, 19 histologically diagnosed amyloidotic macaques and 19 age-matched controls were assayed for changes in various parameters by using routinely banked, frozen (–80 °C) sera available from clinical and subclinical time points. Clinically amyloidotic animals displayed increased levels of alkaline phosphatase, aspartate aminotransferase, lactate dehydrogenase, gamma glutamyltranspeptidase, and macrophage colony-stimulating factor and significantly decreased quantities of albumin and total cholesterol. Subclinical amyloidotic animals displayed increased levels of alkaline phosphatase, aspartate aminotransferase, lactate dehydrogenase, and serum amyloid A and decreased concentrations of albumin and total cholesterol. The serologic parameters studied indicate a temporal relationship of these factors not previously described, show a clear pattern of disease progression, and could be useful in subclinical disease detection.

Amyloid is an eosinophilic substance made of insoluble fibrillar protein.³² When deposited extracellularly, amyloid causes displacement of tissue form and disruption of organ function.³² Persistent accretion of amyloid can result in organ failure and ultimately animal death.²² Clinical signs of disease depend on the tissues affected and the degree of involvement.³² Amyloidosis has been well documented in humans, other mammals, birds, and reptiles.³⁸ In humans, amyloidosis plays a key role in many diseases, including Alzheimer disease, type II diabetes, rheumatoid arthritis, and Down syndrome.^15,20,35,38

Amyloidosis generally is classified into 3 categories: primary, secondary, and hereditary. Primary amyloidosis consists of the immunoglobulin- and myeloma-associated types. Secondary (reactive) amyloidosis is associated with chronic inflammation.²⁴ Common causes of secondary amyloidosis in humans include rheumatoid arthritis, idiopathic colitis, infectious diseases, such as tuberculosis and leprosy, and malignant tumors, such as mesothelioma and Hodgkins disease.²⁸ Hereditary amyloid syndromes are rare and include Mediterranean fever, Muckle–Wells syndrome, and familial amyloid cardiomyopathy.^32,38

Secondary amyloidosis is the most common form of amyloidosis in animals.³⁸ Amyloidosis occurs in many species of nonhuman primates including the common marmoset (Callithrix jacchus),²³ squirrel monkey (Saimiri sciureus),³⁴ rhesus macaque (Macaca mulatta),^9,10 pigtailed macaque (Macaca nemestrina),^18,27 crab-eating macaque (Macaca fascicularis),²⁷ barbary ape (Macaca sylvanus),⁶ baboon (Papio spp.),¹⁷ mandrill (Papio sphinx), and chimpanzee (Pan troglodytes).^16,39 Although a definitive cause of secondary amyloidosis has not been identified in nonhuman primates, this condition has been associated with chronic inflammation due to rheumatoid arthritis,⁶ viral infection,¹⁸ parasitism,¹ respiratory disease,^27,30 trauma,³⁰ and bacterial enterocolitis.^27,30,31 Shigella spp. have received particular attention as a common etiology linking enterocolitis with amyloidosis.^4,7,38

Previous research on amyloidosis in nonhuman primates has yielded clinical and serologic profiles in end-stage amyloidotic animals, but little is known about the serologic status in the subclinical stages of disease. Amyloid can accumulate for as long as 3 y before severe organ disruption occurs¹⁴ and clinical signs of amyloidosis become evident.¹⁶ With appropriate analysis, detection of amyloidosis could occur much earlier than typically now achieved, thus allowing for targeted preventative therapy to potentially halt the progression of this insidious disease.

Materials and Methods

Animals.

Rhesus macaques (deceased or living) with histologically diagnosed hepatic amyloidosis (with or without other organ amyloid involvement) were considered for inclusion as study cases. Animals with no histologic evidence of amyloidosis at necropsy were considered for controls. Rhesus macaques with histologic evidence of hepatitis or cholecystitis were excluded from this study. All animals were negative for retroviral pathogens including SIV, simian retrovirus type D, and simian T-lymphotropic virus type. All animals in this study were housed at the California National Primate Research Center at the University of California, Davis, which is an AAALAC-accredited facility. Animals were housed over a 16-y period from 1990 to 2006 either intermittently or permanently paired (indoors in cages) or in large family groups (outdoors in cages). Animals received a diet of commercial monkey chow (Monkey Diet Jumbo 5037, Lab Diet, St Louis, MO) and ad libitum water and were supplemented with a rotation of fresh fruits, vegetables, and behavioral enrichment. All animals were maintained according to recommendations of the Guide for the Care and Use of Laboratory Animals.¹⁹ As part of routine colony management, serum samples are collected from animals approximately every 2 y, stored frozen (–80 °C), and are available for retrospective analyses.

Two separate populations of animals were used for serum biochemical analyses and immunoassays. There was no overlap between populations. For inclusion in the biochemical analyses, cases had a serum biochemical panel obtained at the time of histologic amyloid diagnosis (clinical disease) and a second panel obtained (3 to 32 mo) before diagnosis and clinical signs of disease (subclinical status). Institutional age- and gender-specific references ranges were used as control biochemistry values. For inclusion in the immunoassays, cases had a serum sample obtained and banked in conjunction with the date of histologic amyloid diagnosis (clinical disease) as well as a second sample (9 to 26 mo) before diagnosis and clinical signs of disease (subclinical status). Control animals for the immunoassays were age- and gender- matched to cases and had serum stored at the same intervals as their corresponding cases. The time intervals for serum biochemical analyses and immunoassays were chosen to include the largest sample size of animals (n = 26 and n = 19, respectively) within the shortest interval of time.

Serum biochemistry.

Serum biochemistry panels were performed at the University of California, Davis, Veterinary Medical Teaching Hospital's clinical laboratory. Tests were performed on an automated analyzer (Hitachi 917, Roche Biomedical, Indianapolis, IN). Standard serum biochemistry panels included tests for albumin, alkaline phosphatase, alanine transaminase, aspartate aminotransferase, blood urea nitrogen, calcium, chloride, carbon dioxide, creatine phosphokinase, creatinine, total bilirubin, gamma glutamyltranspeptidase, glucose, lactate dehydrogenase, phosphorus, potassium, sodium, total cholesterol, total protein, and triglycerides. Data for lactate dehydrogenase, aspartate aminotransferase, total cholesterol, and triglycerides were unavailable for 8 (4 female, 4 male) animals.

ELISA for serum amyloid A.

An ELISA for measuring serum amyloid A (SAA) in humans (Tridelta, Kildare, Ireland) was used with banked (–80 °C) serum from rhesus macaques according to the manufacturer's directions with 1 exception: samples were diluted 1:50 in 1× sample–conjugate diluent rather than the suggested 1:500 dilution. This change was instituted to maximize signal strength and prevent the possibility of wasting samples and reagents, while the option for further dilution remained available. Briefly, samples (including standards of known human SAA content) were added to microwells coated with a monoclonal antibody specific for SAA. Biotinylated antiSAA monoclonal antibody was then added to each microwell. Streptavidin–horseradish peroxidase conjugate was added, followed by tetramethylbenzidine substrate. Absorbance at 450 µm was read by using a microtiter plate absorbance reader (Tecan, San Jose, CA). Results were expressed as OD₄₅₀, and concentrations were not extrapolated from standard curves because there are no published standards for macaque SAA.

ELISA for macrophage colony-stimulating factor.

An ELISA for measuring macrophage colony-stimulating factor (mCSF) in humans (Invitrogen, Carlsbad, CA) was performed on banked serum (–80 °C) according to the manufacturer's directions. Briefly, samples (including standards of known human mCSF content) were added to microwells coated with a monoclonal antibody specific for mCSF. Biotinylated antimCSF monoclonal antibody then was added to each microwell. Streptavidin–horseradish peroxidase conjugate was added, followed by tetramethylbenzidine substrate. Absorbance at 450 µm was read using a microtiter plate absorbance reader (Tecan). Results were expressed as OD₄₅₀, and concentrations were not extrapolated from standard curves because there are no published standards for macaque mCSF.

Statistical analysis.

Standard serum biochemistry values for cases were compared with institutional age- and gender-specific references ranges by construction of 95% confidence intervals for the difference between means by using a t test for independent samples. SAA and mCSF results were analyzed for changes over time within cases by using paired t tests and for differences between cases and controls by using t tests for independent samples. The levels of the immunoassay markers were analyzed further by using receiver operating characteristic curves to determine the cutoff OD levels that resulted in the optimal diagnostic accuracy for each marker (amyloid sensitivity, specificity, and positive likelihood ratio). A logistic model of the probability of developing amylodosis using SAA and mCSF as independent variables was constructed by using Stata 9 (StataCorp, College Station, TX) statistical software. Statistical significance for each analysis was defined as a P value of less than 0.05.

Results

Serum biochemistry.

Clinically amyloidotic animals displayed increased levels of alkaline phosphatase, aspartate aminotransferase, lactate dehydrogenase, and gamma glutamyltranspeptidase and decreased concentrations of albumin and total cholesterol (Table 1). Subclinical amyloidotic animals displayed increased quantities of alkaline phosphatase, aspartate aminotransferase, and lactate dehydrogenase but decreased amounts of albumin and total cholesterol (Table 2). All other serum biochemical values were not significantly different from colony references.

Table 1.

Statistically significant biochemical results comparing clinically amyloidotic adult cases to age- and gender-matched controls

Factor	Sex	Amyloid status	N	Mean	95% Confidence interval	Pa
Albumin (g/dL)	F	+	12	2.6	2.3—2.8	<0.01
		—	24	3.4	3.3—3.5
	M	+	14	2.4	2.0—2.7	<0.01
		—	24	3.6	3.5—3.7
Alkaline phosphatase (U/L)	F	+	12	352.8	233.4—472.3	<0.01
		—	24	143.1	112.6—173.7
	M	+	14	323.8	229.1—418.4	<0.01
		—	24	102.2	88.6—115.8
Aspartate aminotransferase (U/L)	M	+	10b	47.3	29.3—65.3	<0.01
		—	24	24.5	22.4—26.5
Cholesterol (mg/dL)	F	+	8b	124.6	92.4—156.8	<0.01
		—	24	178.0	166.3—189.6
	M	+	10b	112.0	84.1—139.9	<0.01
		—	24	148.0	135.3—160.6
Gamma glutamyltransferase (U/L)	F	+	12	83.5	54.4—112.6	<0.01
		—	24	46.4	52.0—50.8
	M	+	14	85.2	72.5—97.9	<0.01
		—	24	51.3	44.4—58.2
Lactate dehydrogenase (U/L)	F	+	8b	423.3	233.8—612.7	<0.05
		—	24	277.0	211.9—342.2
	M	+	10b	691.4	388.5—994.3	<0.01
		—	24	333.9	270.6—397.2
Total protein (g/dL)	M	+	14	6.4	5.7—7.2	<0.05
		—	24	7.1	7.0—7.2

^a2-tailed t test for independent samples

^bIncomplete biochemistry panels from 4 cases

Table 2.

Statistically significant biochemical results comparing subclinical amyloidotic adult cases to age- and gender-matched controls

Factor	Gender	Amyloid status	N	Mean	95% Confidence interval	Pa
Albumin (g/dL)	F	+	12	2.7	2.3—3.1	<0.01
		—	24	3.4	3.3—3.5
	M	+	14	2.7	2.2—3.1	<0.01
		—	24	3.6	3.5—3.7
Alkaline phosphatase (U/L)	F	+	12	228.3	144.4—312.3	<0.01
		—	24	143.1	112.6—173.7
	M	+	14	257.3	107.9—406.6	<0.05
		—	24	102.2	88.6—115.8
Aspartate aminotransferase (U/L)	F	+	8b	58.8	25.3—92.2	<0.01
		—	24	27.8	24.8—30.9
	M	+	10b	45.9	34.7—57.1	<0.01
		—	24	24.5	22.4—26.5
Cholesterol (mg/dL)	F	+	8b	141.6	112.5—170.7	<0.01
		—	24	178.0	166.3—189.6
	M	+	10b	111.7	97.1—126.3	<0.01
		—	24	148.0	135.3—160.6
Lactate dehydrogenase (U/L)	F	+	8b	666.8	279.1—1054.4	<0.01
		—	24	277.0	211.9—342.2
	M	+	10b	651.5	340.5—962.5	<0.01
		—	24	333.9	270.6—397.2
Total protein (g/dL)	M	+	14	6.6	6.1—7.1	<0.05
		—	24	7.1	7.0—7.2

^a2-tailed t test for independent samples

^bIncomplete biochemistry panels from 4 cases

Immunoassays.

Clinically amyloidotic animals displayed increased mCSF levels whereas subclinical amyloidotic animals displayed increased SAA values (Table 3). Within cases, SAA decreased significantly over time as measured by Δ_OD (–0.42, P < 0.01). In addition, receiver operating characteristic curves were used to evaluate the individual diagnostic value of SAA and mCSF. Cutoff values were set at the level that resulted in the optimal positive likelihood ratio for each marker. The sensitivity, specificity, and likelihood ratio for each marker developed by using these values are shown in Table 4. According to logistic regression, SAA and mCSF used in combination are predictive of amyloidosis (likelihood ratio x² = 17.4; P>x² = 0.01; Table 5). SAA during the subclinical stage and mCSF during the clinical stage of amyloidosis alone are significantly responsible for the predictability of this model.

Table 3.

Immunoassay results comparing amyloidotic cases to age- and gender-matched controls

				OD450
Factor	Disease stage	Amyloid status	N	Mean	SEM	Pa
SAA	Preclinical	+	19	0.97	0.15	<0.05
		—	19	0.49	0.12
	Clinical	+	19	0.55	0.13	>0.1
		—	19	0.72	0.15
mCSF	Preclinical	+	19	1.19	0.16	>0.1
		—	19	0.90	0.07
	Clinical	+	19	1.45	0.19	<0.05
		—	19	0.98	0.06

^a2-tailed paired t test

Table 4.

Cut-off OD values for preclinically amyloidotic animals and resulting positive likelihood ratios

Factor	Cut-off OD	Sensitivity	Specificity	Positive likelihood ratio
SAA	1.59	32%	95%	6.0
mCSF	1.31	32%	95%	6.0

Table 5.

Coefficients, standard errors, and P values of the logistic regression model of combined SAA and mCSF factors

Factor	Disease stage	Coefficient	SEM	z	P > |z|
SAA	Preclinical	2.03	0.90	2.25	0.024
	Clinical	−1.85	0.97	−1.90	0.057
mCSF	Preclinical	0.47	1.34	0.35	0.728
	Clinical	3.34	1.67	2.00	0.046

LR x² = 17.4

P > x² = 0.002

Discussion

Amyloidosis in nonhuman primates is regarded as a chronic, progressive, intractable disease that is usually not clinically apparent until considerable organ damage has occurred.¹⁶ The most common tissue sites of amyloid deposition in nonhuman primates are the spleen, liver, kidneys, intestines, mesenteric lymph nodes, and adrenal glands.^17,31,38 Nonhuman primates with secondary amyloidosis may present clinically with cachexia and muscle wasting, recurrent diarrhea, hepatomegaly, anemia, hypoproteinemia, hypoalbuminemia, elevated liver enzymes, or arthralgia (or a combination of these symptoms).^4,7,16,18,23 The prognosis for nonhuman primates with clinical signs of amyloidosis is grave.¹⁶ Amyloid deposition may lead to impaired organ function and altered drug metabolism, which also renders these animals poor candidates for anesthesia and surgery and may be a confounding variable in biomedical research experiments.¹⁸

The clinically amyloidotic macaques studied here displayed increased alkaline phosphatase and gamma glutamyltranspeptidase levels, which together indicate cholestasis due to the canalicular deposition of amyloid in the liver.¹⁶ Interestingly, total bilirubin was not increased significantly, as would be expected in cholestasis. The findings obtained are consistent with those in the literature.¹⁸ Alkaline phosphatase, but not gamma glutamyltranspeptidase, was increased in the subclinical amyloidotic macaques. This result is somewhat unexpected due to the sensitive nature of gamma glutamyltranspeptidase compared with alkaline phosphatase as a marker of hepatobiliary disease. Juvenile macaques were excluded from this portion of the study to rule-out reactivity with the bone isoenzyme of alkaline phosphatase.

During both stages of disease, amyloidotic macaques displayed decreased levels of albumin and total cholesterol. These findings were expected, because both albumin and cholesterol are known negative acute-phase reactants; they are decreased due to acute and chronic inflammation,⁵ which is a hallmark of secondary amyloidosis. Hypoalbuminemia is accompanied by a loss in the amyloid-degrading ability,²⁵ which further perpetuates the cycle of amyloid formation. Hypocholesterolemia has not been observed previously in macaques with amyloidosis. Conversely, hypercholesterolemia was found in amyloidotic pigtailed macaques and attributed to cholestasis.¹⁸

The decreases in albumin and total protein (males only) may also be attributed in part to protein-losing enteropathy. Of the 26 amyloidotic macaques studied, 5 (19.2%) had intestinal amyloidosis in conjunction with the hepatic distribution. However, the albumin levels of these 5 animals were not significantly different from those of the other 21 amyloidotic macaques. For this reason, the etiology of hypoalbuminemia can be attributed to the nature of the acute-phase reactants, gastrointestinal losses, and decreased hepatic production.

Lactate dehydrogenase and aspartate aminotransferase were increased during both stages of disease. In the absence of significant increases in alanine aminotransferase, these findings can be attributed to the effect of muscle-wasting and cachexia rather than as enzymes of hepatocellular damage.

Serum amyloid A is an acute-phase protein that is produced in the liver and thought to cause chemotaxis and adhesion of phagocytic cells and lymphocytes during inflammation.^9,12 Levels of SAA levels can increase as much as 1000-fold in acute inflammation.^12,33 In addition, SAA is thought to be a precursor protein for amyloid production. Although necessary, SAA alone is not sufficient for amyloid formation.^10,24,36 Persistently increased SAA levels are due to a reduced or incomplete ability to degrade SAA that progresses to amyloid formation.^10,35 Quantitation of SAA recently was found to be better than organ function tests as a clinical monitor of amyloid burden in humans.¹³ In a previous study,¹⁸ end-stage amyloidotic pigtailed macaques had significantly higher SAA levels than did nonamyloidotic controls. That correlation was not demonstrated in the current study. However, the subclinical amyloidotic macaques in the current study showed significantly increased SAA. Perhaps this subclinical increase is a reflection of the underlying cause of the amyloidosis, the initial inflammatory insult.

Macrophages play an essential role in the formation of amyloid.²¹ Although SAA is chemotaxic for macrophages,² the incomplete degradation of SAA by tissue macrophages allows for amyloid formation.¹¹ In the presence of amyloid protein in vitro, activated macrophages taken from mice predisposed to amyloidosis lack the ability to catabolize SAA, whereas macrophages from control mice can completely degrade SAA.¹¹ Macrophage colony-stimulating factor is the key regulator of the proliferation, activation, and survival of mononuclear phagocytes.^3,8 In vitro, phagocytosis of β-amyloid protein by murine microglial cells is enhanced by the presence of mCSF.²⁶ Mice deficient in mCSF display a reduction in microglial cell number and exhibit β-amyloid plaques in their brains similar to those seen in Alzheimer disease.³⁷ The role of macrophages in amyloidogenesis appears to be a double-edged sword: macrophages are necessary for both the formation and phagocytosis of amyloid.

In humans, mCSF recently was implicated as a potential marker for renal amyloidosis.²⁹ In a previous study,²⁹ serum mCSF was more sensitive than organ function tests for diagnosing renal amyloidosis, and mCSF was more specific than SAA as a marker of renal amyloidosis. In the current study, mCSF was significantly increased only in the clinically amyloidotic macaques.

The roles of cholesterol, SAA, and macrophages in the pathogenesis of amyloidosis are particularly interesting due to the fact that whereas free SAA is chemotaxic for monocytes and leukocytes, SAA bound to high-density lipoprotein lacks this effect.² Perhaps there is a protective dilutional effect of high-density lipoprotein in modulating the chemotaxic effects of SAA. Decreasing high-density lipoprotein might cause this system to spiral out of control. The ratio of high-density to low-density lipoprotein in cholesterol has not been studied in amyloidosis but should be evaluated to better understand this relationship.

In the present study, immunoassay results were displayed in terms of optical densities rather than concentrations extrapolated from standard curves. We followed this route because macaque standards for these assays do not exist, therefore any calculated concentrations would likely be inaccurate. A potential consequence of our choice is the difficulty in translating the current results to future analyses. One way to resolve this dilemma is to provide calibration factors based on the results of these immunoassays that could be used to standardize future results. In future studies, equivalent cutoff ODs could be calibrated back to the cutoffs established in the present study (1.59 SAA or 1.31 mCSF) by multiplying these original cutoffs by a ratio of the measured OD for the 50 ng (SAA) or 62.5 pg (mCSF) standard in subsequent assays to our original ODs for the same standards (SA, 0.55; mCSF, 0.51).

In summary, serum biochemistry profiles, SAA and mCSF can be useful during early detection of secondary hepatic amyloidosis in rhesus macaques. These assays, when used in combination, can help to identify animals at risk for developing hepatic amyloidosis before overt clinical disease occurs. Furthermore, our findings shed light on the pathogenesis of secondary amyloidosis, during which SAA (early) and mCSF (late) increase as the disease progresses.