Session 4: The Role of the Toxicologic Pathologist in Informing Regulatory Decisions and Guiding the Interpretation and Application of Data from New Technologies and Tools Toxicologic Pathology, 43: 90-97, 2015 Copyright # 2014 by The Author(s) ISSN: 0192-6233 print / 1533-1601 online DOI: 10.1177/0192623314555340
Evaluation of Clinical Pathology Data: Correlating Changes with Other Study Data NANCY E. EVERDS1 1
Amgen Inc, Seattle, Washington, USA ABSTRACT
During the conduct of in vivo toxicology studies, in-life, clinical pathology, and anatomic pathology parameters are collected and interpreted. These sets of parameters are evaluated in an integrative manner to determine the overall toxicity of a test article. For clinical pathology parameters, the inherent variability and physiologic factors affecting each analyte must be understood prior to interpretation. Changes in clinical pathology parameters that are considered to be test article–related are then assessed with respect to changes in the concurrent data sets such as clinical signs and anatomic pathology to determine the underlying pathophysiology. In this article, examples of hemolysis and hepatotoxicity are used to demonstrate the relationships among the various parameters and data sets. Whereas there was tight correlation of all data sets in the example of hemolysis in rats, the examples of altered enzymes and other biomarkers indicating liver injury and dysfunction were more often discordant with other data sets. Keywords:
clinical pathology; hematology; hemolysis; clinical chemistry; hepatotoxicity; biomarker; data interpretation; safety assessment.
concordant hematologic data and discordant clinical chemistry data will be presented.
INTRODUCTION During the conduct of in vivo toxicology studies, several correlative data sets are collected to evaluate the effect of a test article on animals. For most studies, these data sets include inlife, clinical pathology, and anatomic pathology parameters. The overall toxicity of the test article can be understood best if these parameter sets are evaluated in an integrative manner. From a clinical pathology point of view, the factors influencing the strength of data correlations include how tightly a parameter is regulated and the underlying pathophysiology causing a change in the parameter. To illustrate the relationships among sets of data collected during toxicity studies, examples of
TYPES
OF
CLINICAL PATHOLOGY PARAMETERS
Clinical pathology parameters are typically categorized by sample type (whole blood, plasma, serum) and/or instrumentation used for analysis (e.g., hematology, clinical chemistry, coagulation, and urinalysis), or by systems (e.g., inflammation parameters including leukocytes, fibrinogen, and serum proteins). Another way to classify clinical pathology parameters is by those that are regulated and those that have a wide range in health. Clinical pathology parameters that must be in a given range for health are generally those that are regulated, either tightly or loosely. When these parameters deviate from their normal homeostatic state, the increase or decrease in their concentration or activity can result in pathology. Parameters fitting these characteristics include blood cell counts and concentrations or activities of albumin, globulins, sodium, chloride, bicarbonate, potassium, calcium, and coagulation factors. For these parameters, clinical signs or other in-life measurement (e.g., electrocardiography) may provide the clearest correlation. These parameters thus serve as markers of altered homeostasis. For example, large decreases in coagulation factors indicated by prolonged prothrombin time and/ or activated partial thromboplastin time may result in petechial and/or ecchymotic hemorrhages in the skin and mucous
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The author(s) received no financial support for the research, authorship, and/or publication of this article. Address correspondence to: Nancy E. Everds, Amgen Inc., 1201 Amgen Ct. W, Seattle, WA 98119, USA; e-mail: [email protected]. Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; ARET, absolute reticulocyte counts; AST, aspartate aminotransferase; CK, creatine kinase; cTnT, cardiac troponin T; cTnI, cardiac troponin I; FABP3, fatty acid binding protein 3; GGT, gammaglutamyl transferase; GLDH, glutamate dehydrogenase; MCHC, mean cell hemoglobin concentration; MCV, mean cell volume; MIP, muscle injury panel; Myl3, myosin light chain 3; Myl3, parvalbumin; RBC, red blood cell; RDW, red blood cell distribution width; SDH, sorbitol dehydrogenase; sTnI, skeletal troponin I; TIMP, tissue inhibitor of metalloproteinase. 90
Vol. 43, No. 1, 2015
EVALUATION OF CLINICAL PATHOLOGY DATA
membrane of affected animals. Similarly, increased potassium may correlate with alterations in electrocardiograms. Some parameters are not regulated in order to maintain homeostasis. Alterations in these parameters generally do not affect health directly; rather these alterations are generally indicators of an underlying process that has resulted in the change. Most of the parameters that fit into this category are injury or adaptive markers rather than molecules required for good health. These parameters include serum liver enzymes such as alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), and gammaglutamyl transferase (GGT); parameters indicating liver function such as bilirubin or total bile acids; and structural proteins such as skeletal and cardiac troponins. Because these parameters generally indicate a response to a change in a particular tissue or metabolic pathway, they often correlate with structural changes that are detected by light microscopy and are less likely to correlate with in-life changes. These correlations between clinical pathology injury markers and histologic evidence of injury may be lacking when changes are minimal to mild (e.g., minimal or mild cardiac myocyte degeneration without changes in AST; 2–4 increases over controls for ALT activity without light microscopic evidence of liver injury; Boone et al. 2005). CONCORDANCE
DISCORDANCE AMONG DATA SETS FROM TOXICOLOGY STUDIES
91
challenges for clinicians if alternate markers of injury are not available. For example, muscle injury detected by light microscopy may not be reflected by clinical pathology changes during nonclinical studies. Progression of this a molecule with this characteristics may require special monitoring in the cliniic in order to minimize risk to humans. For pharmaceutical test articles, the presence of clinical pathology changes without anatomic pathology changes in nonclinical studies may require an understanding of the underlying mechanism for the discordance or the use of lower starting doses to allow progression of the molecule and may complicate clinical trials. For example, patients are routinely evaluated for liver injury during clinical trials. If a test article causes increased ALT without any histologic evidence of injury nonclinically, the clinician may be hampered in his or her ability to discern drug-related or incidental liver injury from potentially nonadverse increases in ALT. For industrial chemicals and agrichemicals, discordant findings in toxicology studies and associated uncertainties as to their relevance to humans may result in the precautionary need for more stringent limits (larger margins of safety) for human or environmental exposure. These limitations may in turn result in limits on production or the use of products with important societal benefits.
AND
The ideal result for a toxicity study occurs when most or all data correlates on a group and/or individual animal basis, and the correlations point to a clear mechanism of underlying toxicity. Concordance happens fairly frequently for regulated parameters, as described above, or when test articles have a clear and significant impact on organs or physiologic processes in classical or well-documented patterns. A typical example of the former is when tremors are observed in-life, resulting from test article–related decreases in serum calcium. Examples of the latter include multifocal moderate liver necrosis with increased ALT and AST. Because regulated parameters are those required to be maintained for health, they are unlikely to be discordant with other parameters. In contrast, injury markers are more frequently discordant for several reasons: (1) sampling—focal or multifocal lesion that was not included in a histologic section; (2) timing—the injury marker increased before histologic evidence or injury or returned to normal prior to histologic resolution; and (3) flux—a test article alters the production, circulation, excretion of an analyte or biomarker; and (4) interference—a test article interferes with the measurement of an analyte, cell, or related determination. Lack of concordance can cause challenges in interpretation and test article development. For pharmaceuticals, the lack of clinical pathology changes in the presence of anatomic pathology changes during nonclinical studies means that the typical parameters used to monitor toxicity may not be effective in clinical trials. These conditions may create monitoring
CONCORDANCE
OF
HEMATOLOGIC DATA
Red blood cells (RBCs) perform several functions essential for health (Hall and Guyton 2011; Latimer 2011). Their primary function is to carry oxygen to tissues and carbon dioxide away from tissues, and their production is tightly regulated based on their role in oxygen transport. A decrease in oxygen tension caused by decreased RBC mass (estimated by three parameters: RBC count, hemoglobin, and hematocrit) results in rapid increase in erythropoietin production and increased production and release of RBCs from the bone marrow. RBCs are often exposed to relatively high concentrations of test article and/or metabolite, and as anucleate cells are limited in their ability to respond to injury. Decreased RBC mass is a common finding in toxicity studies, and is a potential hazard for industrial compounds and a potential liability for pharmaceutical products. In general, there is good translatability of RBC toxicities in nonhuman species to humans (Olson et al. 2000). Decreases in RBC mass can result from decreased production, hemorrhage, or hemolysis. The underlying cause of decreased RBC mass determines what correlative findings will be observed for hematology parameters and other data sets. The strength of this correlation is mechanism-, time-, and speciesdependent. Decreased production of RBCs is due to primary or secondary effects on bone marrow and auxiliary erythropoietic organs (spleen primarily in rodents). Decreased RBC production during a toxicity study is characterized by decreased reticulocytes, correlative changes in RBC indices, and decreased splenic (rodents) and bone marrow erythropoiesis. The
92
EVERDS
TOXICOLOGIC PATHOLOGY
FIGURE 1.—The correlations among reticulocyte counts, spleen weights (% total body weight), and histology are shown for approximately 2,500 control and dosed rats on 4- to 14-day studies. Increased reticulocytes and splenic weights correlated well with each other and with the anatomic designation of increased hemopoiesis (arrow). Male rats are indicated by blue triangles, and females are indicated by red triangles.
decreased reticulocytes are most prominent in rodents because reticulocytes normally make up a larger proportion of circulating RBCs than in nonrodent species. For the same reason, changes in RBC indices secondary to decreased production are generally only observed or are much more prominent in rodents compared to large animals and are the opposite of those observed with hemolysis or other regenerative effects on RBC mass. Changes in RBC indices associated with decreased RBC production in rodents include decreased mean cell volume (MCV) and red blood cell distribution width (RDW) and increased mean cell hemoglobin concentration (MCHC). Changes in anatomic pathology parameters include decreased erythropoiesis in spleen and decreased cellularity of bone marrow. These changes might not be apparent in short-term studies, especially for larger animals with longer RBC life span. Hemorrhage can be iatrogenic due to blood collection or as a result of the test article causing overt or occult hemorrhage. Hemorrhage is usually associated with robust regeneration of RBCs during shorter-term studies because typical laboratory diets are iron replete. For this reason, hematologic changes associated with hemorrhage may mimic those of hemolysis (see below). Chronic hemorrhage during longer-term studies may result in classical iron deficiency. Hemolysis is defined as destruction of RBCs, causing shortened life span of RBCs; in natural disease and toxicity, hemolysis is usually an extravascular process. Test articles can cause hemolysis by altering the physiology, metabolism, membrane,
shape, fluidity, immunogenicity, contents, or structure of the RBC. Altered RBCs are recognized by macrophages in the spleen and liver and are remodeled or removed from circulation (Ciesla 2012; Schmaier and Lazarus 2011). The hemoglobin of these cells is catabolized to protein and heme, and the heme is recycled into new RBCs. Hemolysis is associated with increased reticulocytes and correlative changes in RBC indices, increased splenic (rodents) and bone marrow erythropoiesis, increased spleen weights, and splenic congestion (rodents) and pigment. The above clinical and anatomic pathology effects are more often observed and/or are prominent in rodents due to the shorter life span of rodent RBCs and the consistency of changes among groups of rodents (especially rats). The underlying cause of hemolysis may not be identified in toxicity studies. The correlation among reticulocytes, spleen weights, and histology is shown graphically in Figure 1. In a retrospective evaluation of data from approximately 2,500 rats on 4 to 14 day studies, increased reticulocytes and splenic weights correlated well with each other and with the anatomic designation of increased hemopoiesis. The time of sample collection (beginning or resolving increased hemopoiesis) may explain why some animals with increased hemopoiesis histologically had normal reticulocytes and spleen weights or that some animals with increased reticulocytes may have normal histology. A commonly observed condition in toxicity studies is analogous to the syndrome of ‘‘anemia of chronic disease’’ in clinical medicine (Weiss and Goodnough 2005). This condition is a
Vol. 43, No. 1, 2015
EVALUATION OF CLINICAL PATHOLOGY DATA
FIGURE 2.—Red blood cell (RBC) mass parameters (mean + SD) were decreased in an exposure-related pattern in rats administered an agrichemical test article via nose-only inhalation chambers for 8 hr/day for 90 days at exposures of 0, 3, 30, and 300 mg/m3. *Indicates statistically significant difference from vehicle control (p < 0.05).
nonspecific reaction to toxicity in animal studies and has been observed with a variety of conditions including inflammatory, metabolic, and endocrine alterations. Anemia of chronic disease results from low-grade hemolysis, decreased RBC production, and iron sequestration by macrophages. The characteristic finding in these animals is minimal to mild decreased RBC mass with unchanged or slightly increased reticulocytes. Bone marrow may contain increased stainable iron in macrophages, which can be observed on hematoxylin/eosin stained sections as golden brown pigment or on Prussian blue-stained sections (Travlos 2006). Erythroid cell numbers are generally normal or slightly decreased. Unlike bone marrow, characteristic splenic changes have not been reported. Affected animals tend to be poor doers or animals with inflammatory, metabolic, or endocrine processes. The terms ‘‘chronic’’ and ‘‘anemia’’ are misnomers for this condition, in that the changes can occur in days rather than weeks, and the changes may be too minimal to be considered anemia. Because this condition is nonspecific, a weight of evidence approach can be used to support a diagnosis, and other potential causes for decreased RBC mass may need to be ruled out. The following data set illustrates some of these hematologic principles. In this study, rats were administered an agrichemical test article in nose-only inhalation chambers (8 hr/day) for 90 days at exposures of 0, 3, 30, and 300 mg/m3. Blood was collected on day 91, the morning after the last exposure. RBC mass parameters were decreased in an exposure-related pattern (Figure 2). In this study, clinical pathology correlates to decreased RBC mass relate primarily to the increased number of circulating reticulocytes (Figure 3). Reticulocytes are larger than mature RBCs and have excess cytoplasm and less concentrated hemoglobin, so groups with increased reticulocytes have increased
93
FIGURE 3.—Reticulocytes, mean cell volume (MCV), and red blood cell distribution width (RDW) were increased and mean cell hemoglobin concentration (MCHC) was decreased (mean + SD) in rats administered an agrichemical test article via nose-only inhalation chambers for 8 hr/day for 90 days at exposures of 0, 3, 30, and 300 mg/m3. *Indicates statistically significant difference from vehicle control (p < 0.05).
FIGURE 4.—Spleen weights (absolute and relative to body weight; mean + SD) were increased in rats administered an agrichemical test article in nose-only inhalation chambers for 8 hr/day for 90 days at exposures of 0, 3, 30, and 300 mg/m3. *Indicates statistically significant difference from vehicle control (p < 0.05).
MCV and decreased MCHC. The greater percentage of large reticulocytes increases the variability of the MCV, as reflected in the increased RDW, defined as the coefficient of variation in the mean cell volume. On previous studies with this test article, Heinz bodies were demonstrated, suggesting that hemolysis is due to oxidant injury. The anatomic pathology correlates to hemolysis are related mostly to increased erythropoiesis (Figure 4). Spleen weights are increased (both absolute and as a percentage of final body weight). These increases are considered to be due to the microscopic findings (see below). Histologic changes in the spleen are consistent with extravascular hemolysis and compensatory increased RBC production (Table 1). Increased splenic pigment is due to the storage of heme after increased phagocytosis of oxidant-damaged RBCs. In mice and rats, production of RBCs in times of increased demand
94
EVERDS
TOXICOLOGIC PATHOLOGY
TABLE 1.—Incidence of rats with histologic changes in spleen. Concentration (mg/m3) No. in group Extramedullary hematopoiesis, increased Minimal Mild Total Pigment, increased Minimal Mild Total Congestion Minimal Total
0 10
3 10
30 10
300 10
0 0 0
0 0 0
0 0 0
7 3 10
0 0 0
0 0 0
0 0 0
6 3 9
0 0
0 0
0 0
6 6
preferentially occurs in the spleen, as the spleen in these species is more efficient (more RBCs produced per precursor) than bone marrow in RBC production (Broudy et al. 1996; Ohno et al. 1993; Ou et al. 1980; Peschle et al. 1977). In addition, bone marrow of young rats is filled with hematopoietic, lymphoid, and stromal cells and has limited additional erythropoietic capacity compared to the spleen. Congestion (an indicator of increased RBC production) is commonly observed in rodent spleens with increased extramedullary erythropoiesis. The findings in this study indicate that the decreased RBC mass was due to hemolysis. Although these findings could have also been observed with hemorrhage, there was no overt or occult evidence for hemorrhage, and the RBC morphology from a previous study indicated oxidant damage. The findings were clearly not due to decreased production. CONCORDANCE
OF
HEPATIC DATA
The liver is a relatively frequent target for nonclinical and clinical toxicity. Clinical pathology markers relating to liver status are generally injury markers and include hepatocellular enzymes, hepatobiliary enzymes, and bile components (Boone et al. 2005; Latimer 2011; Solter 2005). In addition, there are proteins synthesized by the liver that fall into the ‘‘regulated’’ group (e.g., albumin, coagulation factors), but the synthesis of these proteins are generally only affected with profound decreases in hepatic synthetic function. The classic and most widely used hepatocellular injury markers are AST (occurring in cytosol and mitochondria) and ALT (found in cytosol; Boone et al. 2005). Both of these enzymes can be released by injury to muscle, so additional enzymes are sometimes used that are more liver specific. Sorbitol dehydrogenase (SDH) and glutamate dehydrogenase (GLDH) are liver-specific enzymes found in cytosol and mitochondria, respectively. SDH and GLDH can be used to further define the origin of increases in serum AST and ALT. ALP and GGT are enzymes located on the brush border of biliary cells and are more specific for hepatobiliary rather than hepatocellular changes. These enzymes can be induced by xenobiotics that undergo metabolism or by alterations in bile
FIGURE 5.—AST activity was increased for cynomolgus monkeys on day 2 of an exploratory toxicity study, compared to their respective second pretest value, likely as a result of numerous procedures being conducted on the first day of the study. Each line represents data from a single animal. Group 1 is a vehicle control group, and Groups 2 and 3 are low- and high-dose groups, respectively.
composition, pressure, or flow. None of the hepatic enzyme tests measure liver function. The only routine clinical pathology parameters that measure liver function are bile constituents (bilirubin and total bile acids) and analytes that evaluate synthetic function of the liver (proteins, urea nitrogen, and lipids). Bilirubin concentrations are affected by altered bilirubin production (primarily from heme), uptake and processing by hepatocytes, and excretion through the biliary system. The concentration of total bile acids is affected by portal blood flow, uptake by the liver, and excretion through the biliary system. Hepatocellular injury markers can be released during reversible or nonreversible injury. During reversible injury, cells undergo cytoplasmic blebbing. Enzymes along with other cellular constituents are extruded from the cell inside of these blebs (Gores, Herman, and Lemasters 1990; Solter 2005). Frank leakage of these macromolecules does not occur during reversible injury, because the membrane defects required to leak enzymes would not be compatible with cell survival. Additionally, the timing of the appearance of liver enzymes in serum does not correspond to the magnitude of injury or the relative molecular size of enzymes, as one might expect if minor injury resulting in smaller defects allowed leak out of only smaller enzymes. In contrast, during cell death, large defects occur in the membrane with lysis of organelles and release of enzymes into the circulation. At the present time, there is conflicting information as to whether intracellular location of enzymes (cytosol versus mitochondria) correlates with reversibility of cell injury. ALT and AST are found not only in liver but also in muscle. During toxicity studies, increased ALT and AST (generally AST>ALT) can be observed as a result of intramuscular anesthetics or restraint and other handling procedures; these excursions are particularly problematic for nonhuman primate
Vol. 43, No. 1, 2015
EVALUATION OF CLINICAL PATHOLOGY DATA
95
TABLE 2.—Clinical and anatomic pathology findings in an exploratory rat study. Animal Number
Group
AST, U/L
ALT, U/L
Liver, microa
Muscle, microb
CK, U/L
1 2 3 4 5 6
Control
90 114 118 317 518 273
38 31 27 249 124 199
– – – Mild Minimal –
– – – – Minimal –
269 821 561 496 5,481 489
Treated
AST ¼ aspartate aminotransferase; ALT ¼ alanine aminotransferase; CK ¼ creatine kinase. a Relevant findings included degeneration of cells within hepatic sinusoids. b Relevant findings included degeneration of myocytes.
studies with low number of animals per group and numerous study procedures. For example, AST for cynomolgus monkeys on day 2 of an exploratory toxicity study is increased over the second pretest value (Figure 5). During the first day of this study, several procedures were conducted, including multiple restraints with blood collection. Procedure-related increases in ALT and AST also have been reported for mice; mean ALT of mice handled gently by the trunk is fourfold higher at 1-hr post-handling than mice handled by the tail (Swaim, Taylor, and Jersey 1985). Comparisons between procedures conducted during the study and the relative changes in ALT and AST are generally sufficient to attribute changes to procedure rather than the test article. However, more liverand muscle-specific enzymes or parameters (e.g., SDH and GLDH for liver; creatine kinase [CK] or troponin for muscle) may be useful in distinguishing effects due to muscle compared to those due to liver. Changes in transaminases due to handling and diagnostic procedures conducted on animals are particularly prominent during the early part of a study and may represent a response to emotional stress as much as physical perturbations. Psychological stress in rats and mice results in *fourfold increases in AST and smaller increases in ALT (Arakawa et al. 1997; Sanchez et al. 2002). These changes can be blocked by betaadrenergic blockers. Daily repetition of handling procedures over several days results in habituation and attenuation of the increase in serum enzymes. Newer markers of skeletal muscle injury show promise for discriminating injury to liver versus muscle. In a small exploratory study, 6 animals (3/group) were given either vehicle or a test article. Vacuolation was observed in hepatic sinusoidal cells of 2 animals (1 minimal and 1 mild), and skeletal muscle degeneration (minimal) was observed in 1 animal (Table 2). Novel markers of rat skeletal and cardiac muscle injury (muscle injury panel [MIP] 1: cardiac troponin I [cTnI], cardiac troponin T [cTnT], fatty acid binding protein 3 [FABP3], myosin light chain 3 [Myl3], and skeletal troponin I [sTnI]; MIP2: tissue inhibitor of metalloproteinase [TIMP], parvalbumin [Parv], and CK protein) and serum myoglobin were evaluated to determine if the AST and ALT increases in animals 4 and 6 could be due to muscle injury that is not detected by histopathology. In this small study, novel markers of skeletal muscle injury correlated with anatomic pathology changes in the single animal
FIGURE 6.—Exploratory serum markers of muscle injury were measured in 6 animals administered vehicle (control) or a test article (treated). For values above the limit of the Y-axis, actual results are reported in text boxes. Measurements for animal 5, the only animal with histologic evidence of muscle degeneration, were much higher for fatty acid binding protein 3 (FABP3), myosin light chain 3 (Myl3), skeletal troponin I (sTnI), creatine kinase ([CK], concentration), and myoglobin, compared to animals without degeneration. Of the exploratory parameters measured, only TIMP-1 did not show a clear difference between animals with histologic evidence of muscle degeneration and those without. All animals other than animal 5 had values below the limit of quantitation for sTnI and serum myoglobin.
with histologic evidence of skeletal muscle injury and were negative in the other treated animals (Figure 6). Markers of cardiac muscle injury were unchanged. Several pharmaceutical companies are conducting research to better characterize which of these investigational muscle markers may be useful in nonclinical studies. For nonclinical studies with significant histologic injury to the liver (mild to severe), there is generally good concordance of histology with clinical chemistry changes. Concordance is inconsistent when histologic changes are of low grade (minimal to mild) or focal, or when transaminase changes are of low magnitude (e.g., ALT and AST *24 control mean). This lack of concordance may be due in part to the effect of limited tissue and blood sampling, timing of collections, and circulating half-lives of transaminases. Novel hepatic-specific biomarkers are being evaluated by interindustry consortia for potential improved specificity and/or sensitivity compared to ALT
96
EVERDS
(Lawrence 2011). Further data will be required to establish the utility of these newer markers in nonclinical studies. Several causes might underlie discordant finding of increased transaminase values without histologic correlates: preanalytical effects, positive assay interference, adaptation, or differences in production and/or clearance. As mentioned above, preanalytical effects (handling) and stress or other beta-adrenergic effects may cause transient increases in ALT and AST. Decreased clearance of circulating ALT and AST by pharmacologic inhibition of Kupffer cell function has been observed (Radi et al. 2011). This is analogous to the naturally occurring viral disease in mice caused by lactate dehydrogenase elevating virus, in which macrophages are destroyed and serum enzyme activities become elevated due to decreased clearance (Percy and Barthold 2007). Increased ALT and AST have been reported in mice and cynomolgus monkeys with decreased macrophage numbers, with means up to 10 baseline values. Induction of mixed function enzymes has been shown to be a factor in low-fold increases in ALT (Jackson et al. 2008). Adaptation to injury was postulated to be the cause of transient increases in ALT and AST that resolved despite continued dosing (Harrill et al. 2014). There is no clear mechanistic understanding about why this might happen, and thus the toxicologic significance is uncertain. Although it may be possible to discern the underlying cause of discordant increases in transaminases, the process often requires significant investment in investigative studies. CONCLUSION The concordance of data within a toxicity study varies by process, parameter, and species. In general, clinical pathology parameters that are regulated are concordant with other in-life and pathology parameters. Clinical pathology parameters that indicate injury may be more problematic in terms of specificity and concordance and may be more challenging to interpret. ACKNOWLEDGMENTS Steven R. Frame, Mehrdad Ameri, Esther Trueblood, Jeffrey Lawrence, Angela Wilcox for providing examples and review of the manuscript. AUTHORS’ CONTRIBUTION Everds, contributed to conception and design; to acquisition, analysis, and interpretation of drafted manuscript; and also critically revised manuscript. REFERENCES Arakawa, H., Kodama, H., Matsuoka, N., and Yamaguchi, I. (1997). Stress increases plasma enzyme activity in rats: Differential effects of adrenergic and cholinergic blockades. J Pharmacol Exp Ther 280, 1296–303. Boone, L., Meyer, D., Cusick, P., Ennulat, D., Bolliger, A. P., Everds, N., Meador, V., Elliott, G., Honor, D., Bounous, D., and Jordan, H. (2005). Selection and interpretation of clinical pathology indicators of hepatic injury in preclinical studies. Vet Clin Pathol 34, 182–88. Broudy, V. C., Lin, N. L., Priestley, G. V., Nocka, K., and Wolf, N. S. (1996). Interaction of stem cell factor and its receptor c-kit mediates lodgment and
TOXICOLOGIC PATHOLOGY
acute expansion of hematopoietic cells in the murine spleen. Blood 88, 75– 81. Ciesla, B. (2012). In Hematology in Practice. F. A. Davis, Philadelphia, PA. Gores, G. J., Herman, B., and Lemasters, J. J. (1990). Plasma membrane bleb formation and rupture: A common feature of hepatocellular injury. Hepatology 11, 690–98. Hall, J. E., and Guyton, A. C. (2011). Guyton and Hall Textbook of Medical Physiology. Saunders/Elsevier, Philadelphia, PA. Harrill, A. H., Eaddy, J. S., Rose, K., Cullen, J. M., Ramanathan, L., Wanaski, S., Collins, S., Ho, Y., Watkins, P. B., and Lecluyse, E. L. (2014). Liver biomarker and in vitro assessment confirm the hepatic origin of aminotransferase elevations lacking histopathological correlate in beagle dogs treated with GABAA receptor antagonist NP260. Toxicology Appl Pharmacol 277, 131–37. Jackson, E. R., Kilroy, C., Joslin, D. L., Schomaker, S. J., Pruimboom-Brees, I., and Amacher, D. E. (2008). The early effects of short-term dexamethasone administration on hepatic and serum alanine aminotransferase in the rat. Drug Chem Toxicol 31, 427–45. Latimer, K. S., ed. (2011). Duncan and Prasse’s Veterinary Laboratory Medicine: Clinical Pathology. Wiley-Blackwell, West Sussex, UK. Lawrence, J. (2011). New liver injury biomarkers. Accessed September 4, 2014. http://www.aasld.org/additionalmeetings/Documents/hepatotoxicity %20stc/3A-3_Lawrence.pdf. Ohno, H., Ogawa, M., Nishikawa, S., Hayashi, S., Kunisada, T., and Nishikawa, S. (1993). Conditions required for myelopoiesis in murine spleen. Immunol Lett 35, 197–204. Olson, H., Betton, G., Robinson, D., Thomas, K., Monro, A., Kolaja, G., Lilly, P., Sanders, J., Sipes, G., Bracken, W., Dorato, M., Van Deun, K., Smith, P., Berger, B., and Heller, A. (2000). Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul Toxicol Pharmacol 32, 56–67. Ou, L. C., Kim, D., Layton, W. M., Jr., and Smith, R. P. (1980). Splenic erythropoiesis in polycythemic response of the rat to high-altitude exposure. J Appl Physiol 48, 857–61. Percy, D. H., and Barthold, S. W. (2007). ‘‘Chapter 1: Mouse.’’ In Pathology of Laboratory Rodents and Rabbits, pp. 3–125. Blackwell Publishing, Ames Iowa. Peschle, C., Magli, M. C., Cillo, C., Lettieri, F., Genovese, A., Pizzella, F., and Soricelli, A. (1977). Kinetics of erythroid and myeloid stem cells in posthypoxia polycythaemia. Br J Haematol 37, 345–52. Radi, Z. A., Koza-Taylor, P. H., Bell, R. R., Obert, L. A., Runnels, H. A., Beebe, J. S., Lawton, M. P., and Sadis, S. (2011). Increased serum enzyme levels associated with kupffer cell reduction with no signs of hepatic or skeletal muscle injury. Am J Pathol 179, 240–47. Sanchez, O., Arnau, A., Pareja, M., Poch, E., Ramirez, I., and Soley, M. (2002). Acute stress-induced tissue injury in mice: Differences between emotional and social stress. Cell Stress Chaperones 7 36–46. Schmaier, A. H., and Lazarus, H. M. (2011). Concise Guide to Hematology. Wiley-Blackwell, Chichester, West Sussex, UK. Solter, P.F. (2005). Clinical pathology approaches to hepatic injury. Toxicol Pathol 33, 9–16. Swaim, L. D., Taylor, H. W., and Jersey, G. C. (1985). The effect of handling techniques on serum ALT activity in mice. J Appl Toxicol 5 160–62. Travlos, G. S. (2006). Normal structure, function, and histology of the bone marrow. Toxicol Pathol 34 548–65. Weiss, G., and Goodnough, L. T. (2005). Anemia of chronic disease. N Engl J Med 352 1011–23.
FURTHER READING GENERAL Hall, R. L., and Everds, N. E. (2008). Principles of clinical pathology for toxicology studies. In Principles and Methods of Toxicology (A. W. Hayes, ed.), pp. 1317–58. Crc Press, Boca Raton, FL. Latimer, K. S., ed. (2011). Duncan and Prasse’s Veterinary Laboratory Medicine: Clinical Pathology. Wiley-Blackwell, West Sussex, UK.
Vol. 43, No. 1, 2015
EVALUATION OF CLINICAL PATHOLOGY DATA
97
Thrall, M. A., Baker, D. C., Campbell, T. W., DeNicola, D. B., Fettman, M. J., Lassen, E. D., Rebar, A., and Weiser, G., eds. (2012). Veterinary Hematology and Clinical Chemistry. Wiley-Blackwell, West Sussex, UK.
Weiss, D. J., and Wardrop, K. J. (2010). Schalm’s Veterinary Hematology. Wiley-Blackwell, Ames, IA.
HEMATOLOGY
CLINICAL CHEMISTRY
Caro, J., and Erslev, A. J. (1995). Anemia of chronic renal failure. In Williams Hematology (E. Beutler, M. A. Lichtman, B. S. Coller, and T. J. Kipps. eds.), pp. 456. McGraw-Hill, New York. Erslev, A. J. (1995a). Anemia of chronic disease. In Williams Hematology (E. Beutler, M. A. Lichtman, B. S. Coller, and T. J. Kipps., eds.), pp. 518. McGraw-Hill, New York. Erslev, A. J. (1995b). Anemia of endocrine disorders. In Williams Hematology (E. Beutler, M. A. Lichtman, B. S. Coller, and T. J. Kipps. eds.), pp. 462. McGraw-Hill, New York. Evans, G. O. (2009). Animal Hematotoxicity. CRC Press, Boca Raton, FL. Everds, N. E. (2007). Hematology of the laboratory mouse. In The Mouse in Biomedical Research (J. G. Fox, et al., eds.), Vol. 3, pp. 133–70. Elsevier, Burlington, MA. Feldman, B. F., Zinkl, J. G., and Jain, N. C., eds. (2000). Schalm’s Veterinary Hematology. Wiley-Blackwell, West Sussex, UK. Kaushansky, K., Lichtman, M. A., Beutler, E., Kipps, T. J., Prchal, J., and Seligsohn, U. (2010). Williams Hematology. McGraw-Hill, New York.
Burtis, C. A., Ashwood, E. R., and Bruns, D. E., eds. (2005). Teitz Textbook of Clinical Chemistry and Molecular Diagnosis. W. B. Saunders, Philadelphia, PA. Kaneko, J. J., Harvey, J. W., and Bruss, M. L., eds. (2008). Clinical Biochemistry of Domestic Animals. Academic Press, San Diego, CA. Kaplan, L. A., Pesce, A., and Kazmierczak, S., eds. (2010). Clinical Chemistry: Theory, Analysis, Correlation. Mosby, St. Louis, MO. Kramer, J. W., and Hoffman, W. E. (1997). Clinical enzymology In Clinical Biochemistry of Domestic Animals (J. J. Kaneko, et al., eds.), pp. 303. Academic Press, San Diego, CA. Loeb, W. F., and Quimby, F. W. eds. (1999). The Clinical Chemistry of Laboratory Animals. Taylor Francis, Philadelphia, PA. Sherwin, J. E., and Sobenes, J. R. (1996). Liver function. In Clinical Chemistry: Theory, Analysis, Correlation (L. A. Kaplan and A. J. Pesce, eds.), pp. 505. Mosby, St. Louis, MO. Tenant, B. C. (1997). Hepatic function. In Clinical Biochemistry of Domestic Animals (J. J. Kaneko, et al., eds.), pp. 327. Academic Press, San Diego, CA.
For reprints and permissions queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.
Evaluation of Clinical Pathology Data: Correlating Changes with Other Study Data NANCY E. EVERDS1 1
Amgen Inc, Seattle, Washington, USA ABSTRACT
During the conduct of in vivo toxicology studies, in-life, clinical pathology, and anatomic pathology parameters are collected and interpreted. These sets of parameters are evaluated in an integrative manner to determine the overall toxicity of a test article. For clinical pathology parameters, the inherent variability and physiologic factors affecting each analyte must be understood prior to interpretation. Changes in clinical pathology parameters that are considered to be test article–related are then assessed with respect to changes in the concurrent data sets such as clinical signs and anatomic pathology to determine the underlying pathophysiology. In this article, examples of hemolysis and hepatotoxicity are used to demonstrate the relationships among the various parameters and data sets. Whereas there was tight correlation of all data sets in the example of hemolysis in rats, the examples of altered enzymes and other biomarkers indicating liver injury and dysfunction were more often discordant with other data sets. Keywords:
clinical pathology; hematology; hemolysis; clinical chemistry; hepatotoxicity; biomarker; data interpretation; safety assessment.
concordant hematologic data and discordant clinical chemistry data will be presented.
INTRODUCTION During the conduct of in vivo toxicology studies, several correlative data sets are collected to evaluate the effect of a test article on animals. For most studies, these data sets include inlife, clinical pathology, and anatomic pathology parameters. The overall toxicity of the test article can be understood best if these parameter sets are evaluated in an integrative manner. From a clinical pathology point of view, the factors influencing the strength of data correlations include how tightly a parameter is regulated and the underlying pathophysiology causing a change in the parameter. To illustrate the relationships among sets of data collected during toxicity studies, examples of
TYPES
OF
CLINICAL PATHOLOGY PARAMETERS
Clinical pathology parameters are typically categorized by sample type (whole blood, plasma, serum) and/or instrumentation used for analysis (e.g., hematology, clinical chemistry, coagulation, and urinalysis), or by systems (e.g., inflammation parameters including leukocytes, fibrinogen, and serum proteins). Another way to classify clinical pathology parameters is by those that are regulated and those that have a wide range in health. Clinical pathology parameters that must be in a given range for health are generally those that are regulated, either tightly or loosely. When these parameters deviate from their normal homeostatic state, the increase or decrease in their concentration or activity can result in pathology. Parameters fitting these characteristics include blood cell counts and concentrations or activities of albumin, globulins, sodium, chloride, bicarbonate, potassium, calcium, and coagulation factors. For these parameters, clinical signs or other in-life measurement (e.g., electrocardiography) may provide the clearest correlation. These parameters thus serve as markers of altered homeostasis. For example, large decreases in coagulation factors indicated by prolonged prothrombin time and/ or activated partial thromboplastin time may result in petechial and/or ecchymotic hemorrhages in the skin and mucous
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The author(s) received no financial support for the research, authorship, and/or publication of this article. Address correspondence to: Nancy E. Everds, Amgen Inc., 1201 Amgen Ct. W, Seattle, WA 98119, USA; e-mail: [email protected]. Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; ARET, absolute reticulocyte counts; AST, aspartate aminotransferase; CK, creatine kinase; cTnT, cardiac troponin T; cTnI, cardiac troponin I; FABP3, fatty acid binding protein 3; GGT, gammaglutamyl transferase; GLDH, glutamate dehydrogenase; MCHC, mean cell hemoglobin concentration; MCV, mean cell volume; MIP, muscle injury panel; Myl3, myosin light chain 3; Myl3, parvalbumin; RBC, red blood cell; RDW, red blood cell distribution width; SDH, sorbitol dehydrogenase; sTnI, skeletal troponin I; TIMP, tissue inhibitor of metalloproteinase. 90
Vol. 43, No. 1, 2015
EVALUATION OF CLINICAL PATHOLOGY DATA
membrane of affected animals. Similarly, increased potassium may correlate with alterations in electrocardiograms. Some parameters are not regulated in order to maintain homeostasis. Alterations in these parameters generally do not affect health directly; rather these alterations are generally indicators of an underlying process that has resulted in the change. Most of the parameters that fit into this category are injury or adaptive markers rather than molecules required for good health. These parameters include serum liver enzymes such as alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), and gammaglutamyl transferase (GGT); parameters indicating liver function such as bilirubin or total bile acids; and structural proteins such as skeletal and cardiac troponins. Because these parameters generally indicate a response to a change in a particular tissue or metabolic pathway, they often correlate with structural changes that are detected by light microscopy and are less likely to correlate with in-life changes. These correlations between clinical pathology injury markers and histologic evidence of injury may be lacking when changes are minimal to mild (e.g., minimal or mild cardiac myocyte degeneration without changes in AST; 2–4 increases over controls for ALT activity without light microscopic evidence of liver injury; Boone et al. 2005). CONCORDANCE
DISCORDANCE AMONG DATA SETS FROM TOXICOLOGY STUDIES
91
challenges for clinicians if alternate markers of injury are not available. For example, muscle injury detected by light microscopy may not be reflected by clinical pathology changes during nonclinical studies. Progression of this a molecule with this characteristics may require special monitoring in the cliniic in order to minimize risk to humans. For pharmaceutical test articles, the presence of clinical pathology changes without anatomic pathology changes in nonclinical studies may require an understanding of the underlying mechanism for the discordance or the use of lower starting doses to allow progression of the molecule and may complicate clinical trials. For example, patients are routinely evaluated for liver injury during clinical trials. If a test article causes increased ALT without any histologic evidence of injury nonclinically, the clinician may be hampered in his or her ability to discern drug-related or incidental liver injury from potentially nonadverse increases in ALT. For industrial chemicals and agrichemicals, discordant findings in toxicology studies and associated uncertainties as to their relevance to humans may result in the precautionary need for more stringent limits (larger margins of safety) for human or environmental exposure. These limitations may in turn result in limits on production or the use of products with important societal benefits.
AND
The ideal result for a toxicity study occurs when most or all data correlates on a group and/or individual animal basis, and the correlations point to a clear mechanism of underlying toxicity. Concordance happens fairly frequently for regulated parameters, as described above, or when test articles have a clear and significant impact on organs or physiologic processes in classical or well-documented patterns. A typical example of the former is when tremors are observed in-life, resulting from test article–related decreases in serum calcium. Examples of the latter include multifocal moderate liver necrosis with increased ALT and AST. Because regulated parameters are those required to be maintained for health, they are unlikely to be discordant with other parameters. In contrast, injury markers are more frequently discordant for several reasons: (1) sampling—focal or multifocal lesion that was not included in a histologic section; (2) timing—the injury marker increased before histologic evidence or injury or returned to normal prior to histologic resolution; and (3) flux—a test article alters the production, circulation, excretion of an analyte or biomarker; and (4) interference—a test article interferes with the measurement of an analyte, cell, or related determination. Lack of concordance can cause challenges in interpretation and test article development. For pharmaceuticals, the lack of clinical pathology changes in the presence of anatomic pathology changes during nonclinical studies means that the typical parameters used to monitor toxicity may not be effective in clinical trials. These conditions may create monitoring
CONCORDANCE
OF
HEMATOLOGIC DATA
Red blood cells (RBCs) perform several functions essential for health (Hall and Guyton 2011; Latimer 2011). Their primary function is to carry oxygen to tissues and carbon dioxide away from tissues, and their production is tightly regulated based on their role in oxygen transport. A decrease in oxygen tension caused by decreased RBC mass (estimated by three parameters: RBC count, hemoglobin, and hematocrit) results in rapid increase in erythropoietin production and increased production and release of RBCs from the bone marrow. RBCs are often exposed to relatively high concentrations of test article and/or metabolite, and as anucleate cells are limited in their ability to respond to injury. Decreased RBC mass is a common finding in toxicity studies, and is a potential hazard for industrial compounds and a potential liability for pharmaceutical products. In general, there is good translatability of RBC toxicities in nonhuman species to humans (Olson et al. 2000). Decreases in RBC mass can result from decreased production, hemorrhage, or hemolysis. The underlying cause of decreased RBC mass determines what correlative findings will be observed for hematology parameters and other data sets. The strength of this correlation is mechanism-, time-, and speciesdependent. Decreased production of RBCs is due to primary or secondary effects on bone marrow and auxiliary erythropoietic organs (spleen primarily in rodents). Decreased RBC production during a toxicity study is characterized by decreased reticulocytes, correlative changes in RBC indices, and decreased splenic (rodents) and bone marrow erythropoiesis. The
92
EVERDS
TOXICOLOGIC PATHOLOGY
FIGURE 1.—The correlations among reticulocyte counts, spleen weights (% total body weight), and histology are shown for approximately 2,500 control and dosed rats on 4- to 14-day studies. Increased reticulocytes and splenic weights correlated well with each other and with the anatomic designation of increased hemopoiesis (arrow). Male rats are indicated by blue triangles, and females are indicated by red triangles.
decreased reticulocytes are most prominent in rodents because reticulocytes normally make up a larger proportion of circulating RBCs than in nonrodent species. For the same reason, changes in RBC indices secondary to decreased production are generally only observed or are much more prominent in rodents compared to large animals and are the opposite of those observed with hemolysis or other regenerative effects on RBC mass. Changes in RBC indices associated with decreased RBC production in rodents include decreased mean cell volume (MCV) and red blood cell distribution width (RDW) and increased mean cell hemoglobin concentration (MCHC). Changes in anatomic pathology parameters include decreased erythropoiesis in spleen and decreased cellularity of bone marrow. These changes might not be apparent in short-term studies, especially for larger animals with longer RBC life span. Hemorrhage can be iatrogenic due to blood collection or as a result of the test article causing overt or occult hemorrhage. Hemorrhage is usually associated with robust regeneration of RBCs during shorter-term studies because typical laboratory diets are iron replete. For this reason, hematologic changes associated with hemorrhage may mimic those of hemolysis (see below). Chronic hemorrhage during longer-term studies may result in classical iron deficiency. Hemolysis is defined as destruction of RBCs, causing shortened life span of RBCs; in natural disease and toxicity, hemolysis is usually an extravascular process. Test articles can cause hemolysis by altering the physiology, metabolism, membrane,
shape, fluidity, immunogenicity, contents, or structure of the RBC. Altered RBCs are recognized by macrophages in the spleen and liver and are remodeled or removed from circulation (Ciesla 2012; Schmaier and Lazarus 2011). The hemoglobin of these cells is catabolized to protein and heme, and the heme is recycled into new RBCs. Hemolysis is associated with increased reticulocytes and correlative changes in RBC indices, increased splenic (rodents) and bone marrow erythropoiesis, increased spleen weights, and splenic congestion (rodents) and pigment. The above clinical and anatomic pathology effects are more often observed and/or are prominent in rodents due to the shorter life span of rodent RBCs and the consistency of changes among groups of rodents (especially rats). The underlying cause of hemolysis may not be identified in toxicity studies. The correlation among reticulocytes, spleen weights, and histology is shown graphically in Figure 1. In a retrospective evaluation of data from approximately 2,500 rats on 4 to 14 day studies, increased reticulocytes and splenic weights correlated well with each other and with the anatomic designation of increased hemopoiesis. The time of sample collection (beginning or resolving increased hemopoiesis) may explain why some animals with increased hemopoiesis histologically had normal reticulocytes and spleen weights or that some animals with increased reticulocytes may have normal histology. A commonly observed condition in toxicity studies is analogous to the syndrome of ‘‘anemia of chronic disease’’ in clinical medicine (Weiss and Goodnough 2005). This condition is a
Vol. 43, No. 1, 2015
EVALUATION OF CLINICAL PATHOLOGY DATA
FIGURE 2.—Red blood cell (RBC) mass parameters (mean + SD) were decreased in an exposure-related pattern in rats administered an agrichemical test article via nose-only inhalation chambers for 8 hr/day for 90 days at exposures of 0, 3, 30, and 300 mg/m3. *Indicates statistically significant difference from vehicle control (p < 0.05).
nonspecific reaction to toxicity in animal studies and has been observed with a variety of conditions including inflammatory, metabolic, and endocrine alterations. Anemia of chronic disease results from low-grade hemolysis, decreased RBC production, and iron sequestration by macrophages. The characteristic finding in these animals is minimal to mild decreased RBC mass with unchanged or slightly increased reticulocytes. Bone marrow may contain increased stainable iron in macrophages, which can be observed on hematoxylin/eosin stained sections as golden brown pigment or on Prussian blue-stained sections (Travlos 2006). Erythroid cell numbers are generally normal or slightly decreased. Unlike bone marrow, characteristic splenic changes have not been reported. Affected animals tend to be poor doers or animals with inflammatory, metabolic, or endocrine processes. The terms ‘‘chronic’’ and ‘‘anemia’’ are misnomers for this condition, in that the changes can occur in days rather than weeks, and the changes may be too minimal to be considered anemia. Because this condition is nonspecific, a weight of evidence approach can be used to support a diagnosis, and other potential causes for decreased RBC mass may need to be ruled out. The following data set illustrates some of these hematologic principles. In this study, rats were administered an agrichemical test article in nose-only inhalation chambers (8 hr/day) for 90 days at exposures of 0, 3, 30, and 300 mg/m3. Blood was collected on day 91, the morning after the last exposure. RBC mass parameters were decreased in an exposure-related pattern (Figure 2). In this study, clinical pathology correlates to decreased RBC mass relate primarily to the increased number of circulating reticulocytes (Figure 3). Reticulocytes are larger than mature RBCs and have excess cytoplasm and less concentrated hemoglobin, so groups with increased reticulocytes have increased
93
FIGURE 3.—Reticulocytes, mean cell volume (MCV), and red blood cell distribution width (RDW) were increased and mean cell hemoglobin concentration (MCHC) was decreased (mean + SD) in rats administered an agrichemical test article via nose-only inhalation chambers for 8 hr/day for 90 days at exposures of 0, 3, 30, and 300 mg/m3. *Indicates statistically significant difference from vehicle control (p < 0.05).
FIGURE 4.—Spleen weights (absolute and relative to body weight; mean + SD) were increased in rats administered an agrichemical test article in nose-only inhalation chambers for 8 hr/day for 90 days at exposures of 0, 3, 30, and 300 mg/m3. *Indicates statistically significant difference from vehicle control (p < 0.05).
MCV and decreased MCHC. The greater percentage of large reticulocytes increases the variability of the MCV, as reflected in the increased RDW, defined as the coefficient of variation in the mean cell volume. On previous studies with this test article, Heinz bodies were demonstrated, suggesting that hemolysis is due to oxidant injury. The anatomic pathology correlates to hemolysis are related mostly to increased erythropoiesis (Figure 4). Spleen weights are increased (both absolute and as a percentage of final body weight). These increases are considered to be due to the microscopic findings (see below). Histologic changes in the spleen are consistent with extravascular hemolysis and compensatory increased RBC production (Table 1). Increased splenic pigment is due to the storage of heme after increased phagocytosis of oxidant-damaged RBCs. In mice and rats, production of RBCs in times of increased demand
94
EVERDS
TOXICOLOGIC PATHOLOGY
TABLE 1.—Incidence of rats with histologic changes in spleen. Concentration (mg/m3) No. in group Extramedullary hematopoiesis, increased Minimal Mild Total Pigment, increased Minimal Mild Total Congestion Minimal Total
0 10
3 10
30 10
300 10
0 0 0
0 0 0
0 0 0
7 3 10
0 0 0
0 0 0
0 0 0
6 3 9
0 0
0 0
0 0
6 6
preferentially occurs in the spleen, as the spleen in these species is more efficient (more RBCs produced per precursor) than bone marrow in RBC production (Broudy et al. 1996; Ohno et al. 1993; Ou et al. 1980; Peschle et al. 1977). In addition, bone marrow of young rats is filled with hematopoietic, lymphoid, and stromal cells and has limited additional erythropoietic capacity compared to the spleen. Congestion (an indicator of increased RBC production) is commonly observed in rodent spleens with increased extramedullary erythropoiesis. The findings in this study indicate that the decreased RBC mass was due to hemolysis. Although these findings could have also been observed with hemorrhage, there was no overt or occult evidence for hemorrhage, and the RBC morphology from a previous study indicated oxidant damage. The findings were clearly not due to decreased production. CONCORDANCE
OF
HEPATIC DATA
The liver is a relatively frequent target for nonclinical and clinical toxicity. Clinical pathology markers relating to liver status are generally injury markers and include hepatocellular enzymes, hepatobiliary enzymes, and bile components (Boone et al. 2005; Latimer 2011; Solter 2005). In addition, there are proteins synthesized by the liver that fall into the ‘‘regulated’’ group (e.g., albumin, coagulation factors), but the synthesis of these proteins are generally only affected with profound decreases in hepatic synthetic function. The classic and most widely used hepatocellular injury markers are AST (occurring in cytosol and mitochondria) and ALT (found in cytosol; Boone et al. 2005). Both of these enzymes can be released by injury to muscle, so additional enzymes are sometimes used that are more liver specific. Sorbitol dehydrogenase (SDH) and glutamate dehydrogenase (GLDH) are liver-specific enzymes found in cytosol and mitochondria, respectively. SDH and GLDH can be used to further define the origin of increases in serum AST and ALT. ALP and GGT are enzymes located on the brush border of biliary cells and are more specific for hepatobiliary rather than hepatocellular changes. These enzymes can be induced by xenobiotics that undergo metabolism or by alterations in bile
FIGURE 5.—AST activity was increased for cynomolgus monkeys on day 2 of an exploratory toxicity study, compared to their respective second pretest value, likely as a result of numerous procedures being conducted on the first day of the study. Each line represents data from a single animal. Group 1 is a vehicle control group, and Groups 2 and 3 are low- and high-dose groups, respectively.
composition, pressure, or flow. None of the hepatic enzyme tests measure liver function. The only routine clinical pathology parameters that measure liver function are bile constituents (bilirubin and total bile acids) and analytes that evaluate synthetic function of the liver (proteins, urea nitrogen, and lipids). Bilirubin concentrations are affected by altered bilirubin production (primarily from heme), uptake and processing by hepatocytes, and excretion through the biliary system. The concentration of total bile acids is affected by portal blood flow, uptake by the liver, and excretion through the biliary system. Hepatocellular injury markers can be released during reversible or nonreversible injury. During reversible injury, cells undergo cytoplasmic blebbing. Enzymes along with other cellular constituents are extruded from the cell inside of these blebs (Gores, Herman, and Lemasters 1990; Solter 2005). Frank leakage of these macromolecules does not occur during reversible injury, because the membrane defects required to leak enzymes would not be compatible with cell survival. Additionally, the timing of the appearance of liver enzymes in serum does not correspond to the magnitude of injury or the relative molecular size of enzymes, as one might expect if minor injury resulting in smaller defects allowed leak out of only smaller enzymes. In contrast, during cell death, large defects occur in the membrane with lysis of organelles and release of enzymes into the circulation. At the present time, there is conflicting information as to whether intracellular location of enzymes (cytosol versus mitochondria) correlates with reversibility of cell injury. ALT and AST are found not only in liver but also in muscle. During toxicity studies, increased ALT and AST (generally AST>ALT) can be observed as a result of intramuscular anesthetics or restraint and other handling procedures; these excursions are particularly problematic for nonhuman primate
Vol. 43, No. 1, 2015
EVALUATION OF CLINICAL PATHOLOGY DATA
95
TABLE 2.—Clinical and anatomic pathology findings in an exploratory rat study. Animal Number
Group
AST, U/L
ALT, U/L
Liver, microa
Muscle, microb
CK, U/L
1 2 3 4 5 6
Control
90 114 118 317 518 273
38 31 27 249 124 199
– – – Mild Minimal –
– – – – Minimal –
269 821 561 496 5,481 489
Treated
AST ¼ aspartate aminotransferase; ALT ¼ alanine aminotransferase; CK ¼ creatine kinase. a Relevant findings included degeneration of cells within hepatic sinusoids. b Relevant findings included degeneration of myocytes.
studies with low number of animals per group and numerous study procedures. For example, AST for cynomolgus monkeys on day 2 of an exploratory toxicity study is increased over the second pretest value (Figure 5). During the first day of this study, several procedures were conducted, including multiple restraints with blood collection. Procedure-related increases in ALT and AST also have been reported for mice; mean ALT of mice handled gently by the trunk is fourfold higher at 1-hr post-handling than mice handled by the tail (Swaim, Taylor, and Jersey 1985). Comparisons between procedures conducted during the study and the relative changes in ALT and AST are generally sufficient to attribute changes to procedure rather than the test article. However, more liverand muscle-specific enzymes or parameters (e.g., SDH and GLDH for liver; creatine kinase [CK] or troponin for muscle) may be useful in distinguishing effects due to muscle compared to those due to liver. Changes in transaminases due to handling and diagnostic procedures conducted on animals are particularly prominent during the early part of a study and may represent a response to emotional stress as much as physical perturbations. Psychological stress in rats and mice results in *fourfold increases in AST and smaller increases in ALT (Arakawa et al. 1997; Sanchez et al. 2002). These changes can be blocked by betaadrenergic blockers. Daily repetition of handling procedures over several days results in habituation and attenuation of the increase in serum enzymes. Newer markers of skeletal muscle injury show promise for discriminating injury to liver versus muscle. In a small exploratory study, 6 animals (3/group) were given either vehicle or a test article. Vacuolation was observed in hepatic sinusoidal cells of 2 animals (1 minimal and 1 mild), and skeletal muscle degeneration (minimal) was observed in 1 animal (Table 2). Novel markers of rat skeletal and cardiac muscle injury (muscle injury panel [MIP] 1: cardiac troponin I [cTnI], cardiac troponin T [cTnT], fatty acid binding protein 3 [FABP3], myosin light chain 3 [Myl3], and skeletal troponin I [sTnI]; MIP2: tissue inhibitor of metalloproteinase [TIMP], parvalbumin [Parv], and CK protein) and serum myoglobin were evaluated to determine if the AST and ALT increases in animals 4 and 6 could be due to muscle injury that is not detected by histopathology. In this small study, novel markers of skeletal muscle injury correlated with anatomic pathology changes in the single animal
FIGURE 6.—Exploratory serum markers of muscle injury were measured in 6 animals administered vehicle (control) or a test article (treated). For values above the limit of the Y-axis, actual results are reported in text boxes. Measurements for animal 5, the only animal with histologic evidence of muscle degeneration, were much higher for fatty acid binding protein 3 (FABP3), myosin light chain 3 (Myl3), skeletal troponin I (sTnI), creatine kinase ([CK], concentration), and myoglobin, compared to animals without degeneration. Of the exploratory parameters measured, only TIMP-1 did not show a clear difference between animals with histologic evidence of muscle degeneration and those without. All animals other than animal 5 had values below the limit of quantitation for sTnI and serum myoglobin.
with histologic evidence of skeletal muscle injury and were negative in the other treated animals (Figure 6). Markers of cardiac muscle injury were unchanged. Several pharmaceutical companies are conducting research to better characterize which of these investigational muscle markers may be useful in nonclinical studies. For nonclinical studies with significant histologic injury to the liver (mild to severe), there is generally good concordance of histology with clinical chemistry changes. Concordance is inconsistent when histologic changes are of low grade (minimal to mild) or focal, or when transaminase changes are of low magnitude (e.g., ALT and AST *24 control mean). This lack of concordance may be due in part to the effect of limited tissue and blood sampling, timing of collections, and circulating half-lives of transaminases. Novel hepatic-specific biomarkers are being evaluated by interindustry consortia for potential improved specificity and/or sensitivity compared to ALT
96
EVERDS
(Lawrence 2011). Further data will be required to establish the utility of these newer markers in nonclinical studies. Several causes might underlie discordant finding of increased transaminase values without histologic correlates: preanalytical effects, positive assay interference, adaptation, or differences in production and/or clearance. As mentioned above, preanalytical effects (handling) and stress or other beta-adrenergic effects may cause transient increases in ALT and AST. Decreased clearance of circulating ALT and AST by pharmacologic inhibition of Kupffer cell function has been observed (Radi et al. 2011). This is analogous to the naturally occurring viral disease in mice caused by lactate dehydrogenase elevating virus, in which macrophages are destroyed and serum enzyme activities become elevated due to decreased clearance (Percy and Barthold 2007). Increased ALT and AST have been reported in mice and cynomolgus monkeys with decreased macrophage numbers, with means up to 10 baseline values. Induction of mixed function enzymes has been shown to be a factor in low-fold increases in ALT (Jackson et al. 2008). Adaptation to injury was postulated to be the cause of transient increases in ALT and AST that resolved despite continued dosing (Harrill et al. 2014). There is no clear mechanistic understanding about why this might happen, and thus the toxicologic significance is uncertain. Although it may be possible to discern the underlying cause of discordant increases in transaminases, the process often requires significant investment in investigative studies. CONCLUSION The concordance of data within a toxicity study varies by process, parameter, and species. In general, clinical pathology parameters that are regulated are concordant with other in-life and pathology parameters. Clinical pathology parameters that indicate injury may be more problematic in terms of specificity and concordance and may be more challenging to interpret. ACKNOWLEDGMENTS Steven R. Frame, Mehrdad Ameri, Esther Trueblood, Jeffrey Lawrence, Angela Wilcox for providing examples and review of the manuscript. AUTHORS’ CONTRIBUTION Everds, contributed to conception and design; to acquisition, analysis, and interpretation of drafted manuscript; and also critically revised manuscript. REFERENCES Arakawa, H., Kodama, H., Matsuoka, N., and Yamaguchi, I. (1997). Stress increases plasma enzyme activity in rats: Differential effects of adrenergic and cholinergic blockades. J Pharmacol Exp Ther 280, 1296–303. Boone, L., Meyer, D., Cusick, P., Ennulat, D., Bolliger, A. P., Everds, N., Meador, V., Elliott, G., Honor, D., Bounous, D., and Jordan, H. (2005). Selection and interpretation of clinical pathology indicators of hepatic injury in preclinical studies. Vet Clin Pathol 34, 182–88. Broudy, V. C., Lin, N. L., Priestley, G. V., Nocka, K., and Wolf, N. S. (1996). Interaction of stem cell factor and its receptor c-kit mediates lodgment and
TOXICOLOGIC PATHOLOGY
acute expansion of hematopoietic cells in the murine spleen. Blood 88, 75– 81. Ciesla, B. (2012). In Hematology in Practice. F. A. Davis, Philadelphia, PA. Gores, G. J., Herman, B., and Lemasters, J. J. (1990). Plasma membrane bleb formation and rupture: A common feature of hepatocellular injury. Hepatology 11, 690–98. Hall, J. E., and Guyton, A. C. (2011). Guyton and Hall Textbook of Medical Physiology. Saunders/Elsevier, Philadelphia, PA. Harrill, A. H., Eaddy, J. S., Rose, K., Cullen, J. M., Ramanathan, L., Wanaski, S., Collins, S., Ho, Y., Watkins, P. B., and Lecluyse, E. L. (2014). Liver biomarker and in vitro assessment confirm the hepatic origin of aminotransferase elevations lacking histopathological correlate in beagle dogs treated with GABAA receptor antagonist NP260. Toxicology Appl Pharmacol 277, 131–37. Jackson, E. R., Kilroy, C., Joslin, D. L., Schomaker, S. J., Pruimboom-Brees, I., and Amacher, D. E. (2008). The early effects of short-term dexamethasone administration on hepatic and serum alanine aminotransferase in the rat. Drug Chem Toxicol 31, 427–45. Latimer, K. S., ed. (2011). Duncan and Prasse’s Veterinary Laboratory Medicine: Clinical Pathology. Wiley-Blackwell, West Sussex, UK. Lawrence, J. (2011). New liver injury biomarkers. Accessed September 4, 2014. http://www.aasld.org/additionalmeetings/Documents/hepatotoxicity %20stc/3A-3_Lawrence.pdf. Ohno, H., Ogawa, M., Nishikawa, S., Hayashi, S., Kunisada, T., and Nishikawa, S. (1993). Conditions required for myelopoiesis in murine spleen. Immunol Lett 35, 197–204. Olson, H., Betton, G., Robinson, D., Thomas, K., Monro, A., Kolaja, G., Lilly, P., Sanders, J., Sipes, G., Bracken, W., Dorato, M., Van Deun, K., Smith, P., Berger, B., and Heller, A. (2000). Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul Toxicol Pharmacol 32, 56–67. Ou, L. C., Kim, D., Layton, W. M., Jr., and Smith, R. P. (1980). Splenic erythropoiesis in polycythemic response of the rat to high-altitude exposure. J Appl Physiol 48, 857–61. Percy, D. H., and Barthold, S. W. (2007). ‘‘Chapter 1: Mouse.’’ In Pathology of Laboratory Rodents and Rabbits, pp. 3–125. Blackwell Publishing, Ames Iowa. Peschle, C., Magli, M. C., Cillo, C., Lettieri, F., Genovese, A., Pizzella, F., and Soricelli, A. (1977). Kinetics of erythroid and myeloid stem cells in posthypoxia polycythaemia. Br J Haematol 37, 345–52. Radi, Z. A., Koza-Taylor, P. H., Bell, R. R., Obert, L. A., Runnels, H. A., Beebe, J. S., Lawton, M. P., and Sadis, S. (2011). Increased serum enzyme levels associated with kupffer cell reduction with no signs of hepatic or skeletal muscle injury. Am J Pathol 179, 240–47. Sanchez, O., Arnau, A., Pareja, M., Poch, E., Ramirez, I., and Soley, M. (2002). Acute stress-induced tissue injury in mice: Differences between emotional and social stress. Cell Stress Chaperones 7 36–46. Schmaier, A. H., and Lazarus, H. M. (2011). Concise Guide to Hematology. Wiley-Blackwell, Chichester, West Sussex, UK. Solter, P.F. (2005). Clinical pathology approaches to hepatic injury. Toxicol Pathol 33, 9–16. Swaim, L. D., Taylor, H. W., and Jersey, G. C. (1985). The effect of handling techniques on serum ALT activity in mice. J Appl Toxicol 5 160–62. Travlos, G. S. (2006). Normal structure, function, and histology of the bone marrow. Toxicol Pathol 34 548–65. Weiss, G., and Goodnough, L. T. (2005). Anemia of chronic disease. N Engl J Med 352 1011–23.
FURTHER READING GENERAL Hall, R. L., and Everds, N. E. (2008). Principles of clinical pathology for toxicology studies. In Principles and Methods of Toxicology (A. W. Hayes, ed.), pp. 1317–58. Crc Press, Boca Raton, FL. Latimer, K. S., ed. (2011). Duncan and Prasse’s Veterinary Laboratory Medicine: Clinical Pathology. Wiley-Blackwell, West Sussex, UK.
Vol. 43, No. 1, 2015
EVALUATION OF CLINICAL PATHOLOGY DATA
97
Thrall, M. A., Baker, D. C., Campbell, T. W., DeNicola, D. B., Fettman, M. J., Lassen, E. D., Rebar, A., and Weiser, G., eds. (2012). Veterinary Hematology and Clinical Chemistry. Wiley-Blackwell, West Sussex, UK.
Weiss, D. J., and Wardrop, K. J. (2010). Schalm’s Veterinary Hematology. Wiley-Blackwell, Ames, IA.
HEMATOLOGY
CLINICAL CHEMISTRY
Caro, J., and Erslev, A. J. (1995). Anemia of chronic renal failure. In Williams Hematology (E. Beutler, M. A. Lichtman, B. S. Coller, and T. J. Kipps. eds.), pp. 456. McGraw-Hill, New York. Erslev, A. J. (1995a). Anemia of chronic disease. In Williams Hematology (E. Beutler, M. A. Lichtman, B. S. Coller, and T. J. Kipps., eds.), pp. 518. McGraw-Hill, New York. Erslev, A. J. (1995b). Anemia of endocrine disorders. In Williams Hematology (E. Beutler, M. A. Lichtman, B. S. Coller, and T. J. Kipps. eds.), pp. 462. McGraw-Hill, New York. Evans, G. O. (2009). Animal Hematotoxicity. CRC Press, Boca Raton, FL. Everds, N. E. (2007). Hematology of the laboratory mouse. In The Mouse in Biomedical Research (J. G. Fox, et al., eds.), Vol. 3, pp. 133–70. Elsevier, Burlington, MA. Feldman, B. F., Zinkl, J. G., and Jain, N. C., eds. (2000). Schalm’s Veterinary Hematology. Wiley-Blackwell, West Sussex, UK. Kaushansky, K., Lichtman, M. A., Beutler, E., Kipps, T. J., Prchal, J., and Seligsohn, U. (2010). Williams Hematology. McGraw-Hill, New York.
Burtis, C. A., Ashwood, E. R., and Bruns, D. E., eds. (2005). Teitz Textbook of Clinical Chemistry and Molecular Diagnosis. W. B. Saunders, Philadelphia, PA. Kaneko, J. J., Harvey, J. W., and Bruss, M. L., eds. (2008). Clinical Biochemistry of Domestic Animals. Academic Press, San Diego, CA. Kaplan, L. A., Pesce, A., and Kazmierczak, S., eds. (2010). Clinical Chemistry: Theory, Analysis, Correlation. Mosby, St. Louis, MO. Kramer, J. W., and Hoffman, W. E. (1997). Clinical enzymology In Clinical Biochemistry of Domestic Animals (J. J. Kaneko, et al., eds.), pp. 303. Academic Press, San Diego, CA. Loeb, W. F., and Quimby, F. W. eds. (1999). The Clinical Chemistry of Laboratory Animals. Taylor Francis, Philadelphia, PA. Sherwin, J. E., and Sobenes, J. R. (1996). Liver function. In Clinical Chemistry: Theory, Analysis, Correlation (L. A. Kaplan and A. J. Pesce, eds.), pp. 505. Mosby, St. Louis, MO. Tenant, B. C. (1997). Hepatic function. In Clinical Biochemistry of Domestic Animals (J. J. Kaneko, et al., eds.), pp. 327. Academic Press, San Diego, CA.
For reprints and permissions queries, please visit SAGE’s Web site at http://www.sagepub.com/journalsPermissions.nav.
