- Split View
-
Views
-
Cite
Cite
Dennis F. Lawler, Kevin Chase, Robert Teckenbrock, Karl G. Lark, Heritable Components of Feline Hematology, Clinical Chemistry, and Acid–Base Profiles, Journal of Heredity, Volume 97, Issue 6, November/December 2006, Pages 549–554, https://doi.org/10.1093/jhered/esl041
- Share Icon Share
Abstract
Four erythrocyte variables (erythrocyte count, hemoglobin, mean cell volume, packed cell volume), 14 serum variables (alanine transferase, albumin, alkaline phosphatase, calcium, chloride, cholesterol, creatinine, glucose, phosphorus, potassium, sodium, total protein, triglycerides, urea nitrogen), and 7 venous acid–base variables (base excess, bicarbonate, carbon dioxide partial pressure, oxygen partial pressure, oxygen saturation, pH, and total carbon dioxide) were evaluated for heritability in domestic cats (Felis catus). Values used for individual cats were expressed as the mean over all lifetime measurements, using 444–530 animals for clinical chemistry, 629 animals for acid–base, and 564 animals for erythrocyte metrics. Gender and age at death (where applicable) also were evaluated for correlation with variables. Heritabilities for clinical chemistry, acid–base, and erythrocyte variables ranged, respectively, from 0.13 to 0.78, from 0.23 to 0.59, and from 0.41 to 0.69 (P < 0.05). This result indicates that serum variability has a genetic basis and is segregating in this feline population. These findings may have important implications in both research and clinical medicine.
Hematology, clinical chemistry, and blood gas screening are used to support health maintenance programs and disease diagnosis in many species. Routine use of these variables in the clinical environment involves a synthesis with medical history, results of physical examination, and results of other diagnostic tests. Specific results often must be evaluated together because interpretation of one variable can depend on the outcomes of tests for other variables. These tests also are important components of studies that evaluate safety and efficacy of drugs and other compounds.
Important factors that lead to variability in expression of these measures among species include anatomy and function (ruminant vs. monogastric), dietary habits (herbivore, omnivore, carnivore), or species-associated behaviors. Among individuals of the same species, many physiological and environmental influences on these same variables also can be identified, including stress, obesity, pregnancy, growth, meal times and frequency, circadian fluctuations, and many diseases (Kaneko 1980). Thus, it can be appreciated that homeostasis as measured by these variables is largely circumstantial.
Several studies also have indicated a genetic basis for variability within a species (Havlik et al. 1977; Hamsten et al. 1986; Kalousdian et al. 1987; Randi and Ragni 1991; Boerwinkle et al. 1992; Heller et al. 1993; Snieder et al. 1999; Hunter et al. 2002; Bathum et al. 2004; Greenfield et al. 2004). Interesting comparative summaries of these works have appeared in some later reports (Kalousdian et al. 1987; Heller et al. 1993; Snieder et al. 1999), documenting quantitative genetic control over expression of some of these variables in plasma, serum, or urine. Among humans, these studies often have used twins to segregate environmental and genetic influences by comparing monozygotic and dizygotic pairs (Kalousdian et al. 1987). In addition, it is becoming important to understand the degree to which heritability comparisons within populations and among species can or should be generalized. Phenotypes with similar heritabilities can have very different implications in different populations or species, raising questions about potentially important gene–gene and/or gene–environment interactions.
The Purina cat population is a large colony of more than 11 000 loosely related individuals maintained under a consistent environmental and health care system over about 3 decades at a single site to conduct nutritional research. Using data from this colony, we undertook evaluation of a feline model to define the extent of genetic regulation of intraspecies variation in blood and serum expression of selected clinical measures. Our data provide quantitative heritability measurements for the erythrocyte, clinical biochemistry, and acid–base variables examined.
Materials and Methods
A feline breeding and adult colony numbering an average of about 900 adult domestic shorthair cats (Felis catus) was maintained over about 3 decades at a single site to conduct nutritional research (Nestle [Ralston] Purina Company, Gray Summit, MO). Birth dates within this population range over the years 1960–2000, with about 2500 cats born after 1990. In each generation, many cats were retained in the colony while some new cats were introduced in order to avoid the problems that accompany inbreeding. The deepest pedigree is 14 generations. This population represents a diverse genetic background with a wide range of consanguinities (f = 0.0–0.33), resulting in a wide range of phenotypes with enough relatives to provide power for genetic analysis.
Erythrocyte variables, clinical chemistry, and acid–base profiles are part of the nutritional database maintained on this population. A group of these variables were extracted for genetic analysis: 1) erythrocytes as erythrocyte count, hemoglobin, mean cell volume, and packed cell volume; 2) serum biochemistry as alanine transferase, albumin, alkaline phosphatase, calcium, chloride, cholesterol, creatinine, glucose, phosphorus, potassium, sodium, total protein, triglycerides, and urea nitrogen, and 3) venous acid–base metrics as base excess, bicarbonate, carbon dioxide partial pressure, oxygen partial pressure, oxygen saturation, pH, and total carbon dioxide. In order to increase the robustness of these estimates, likely nonphysiological values were removed, positively skewed nonnormal traits were log transformed to obtain a more normal distribution, and subsets of the data were used for which each subject was measured three or more times. Nonphysiological values were defined as those magnitudes that could not result from normal or abnormal physiological states. The “shapiro.test” of R (R Development Core Team 2005) was used to test for normal distributions of traits.
Results
Summary data for erythrocytes, clinical chemistry, and blood gas variables are listed. (Tables 1–3) For each variable, the heritability with its standard error is shown, together with the population mean and the range of values (minimum and maximum), the number of animals analyzed, and correlations with gender and age at death. Even with standardized sampling conditions (fasted baseline samples drawn in the morning using the same methods), a wide range of individual values were encountered, much of which could be explained as heritable variation.
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
Erythrocytes (106/μl) | 0.475 | 0.040 | 7.692 | 0.043 | 4.51 | 11.83 | 564 | −0.172 | −0.161 |
MCVc (fl) | 0.694 | 0.048 | 47.37 | 0.13 | 37.84 | 56.25 | 564 | 0.138 | 0.222 |
PCVd (%) | 0.438 | 0.041 | 36.33 | 0.19 | 21.6 | 56.7 | 564 | −0.122 | 0.093 |
Hemoglobin (g/dl) | 0.411 | 0.041 | 11.70 | 0.06 | 7.1 | 18.5 | 564 | −0.138 | −0.154 |
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
Erythrocytes (106/μl) | 0.475 | 0.040 | 7.692 | 0.043 | 4.51 | 11.83 | 564 | −0.172 | −0.161 |
MCVc (fl) | 0.694 | 0.048 | 47.37 | 0.13 | 37.84 | 56.25 | 564 | 0.138 | 0.222 |
PCVd (%) | 0.438 | 0.041 | 36.33 | 0.19 | 21.6 | 56.7 | 564 | −0.122 | 0.093 |
Hemoglobin (g/dl) | 0.411 | 0.041 | 11.70 | 0.06 | 7.1 | 18.5 | 564 | −0.138 | −0.154 |
Values in bold are statistically significant (P < 0.05).
Gender correlations to hematology: negative values represent males' higher means.
AOD—age at death correlations to hematology: negative values represent negative correlation.
Mean Cell Volume.
Packed Cell Volume.
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
Erythrocytes (106/μl) | 0.475 | 0.040 | 7.692 | 0.043 | 4.51 | 11.83 | 564 | −0.172 | −0.161 |
MCVc (fl) | 0.694 | 0.048 | 47.37 | 0.13 | 37.84 | 56.25 | 564 | 0.138 | 0.222 |
PCVd (%) | 0.438 | 0.041 | 36.33 | 0.19 | 21.6 | 56.7 | 564 | −0.122 | 0.093 |
Hemoglobin (g/dl) | 0.411 | 0.041 | 11.70 | 0.06 | 7.1 | 18.5 | 564 | −0.138 | −0.154 |
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
Erythrocytes (106/μl) | 0.475 | 0.040 | 7.692 | 0.043 | 4.51 | 11.83 | 564 | −0.172 | −0.161 |
MCVc (fl) | 0.694 | 0.048 | 47.37 | 0.13 | 37.84 | 56.25 | 564 | 0.138 | 0.222 |
PCVd (%) | 0.438 | 0.041 | 36.33 | 0.19 | 21.6 | 56.7 | 564 | −0.122 | 0.093 |
Hemoglobin (g/dl) | 0.411 | 0.041 | 11.70 | 0.06 | 7.1 | 18.5 | 564 | −0.138 | −0.154 |
Values in bold are statistically significant (P < 0.05).
Gender correlations to hematology: negative values represent males' higher means.
AOD—age at death correlations to hematology: negative values represent negative correlation.
Mean Cell Volume.
Packed Cell Volume.
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
Albumin (g/dl) | 0.61 | 0.046 | 3.2 | 0.015 | 2.0 | 4.6 | 445 | −0.12 | −0.03 |
SAPc (IU/l) | 0.78 | 0.047 | 34.0 | 0.727 | 11.0 | 108.8 | 445 | −0.14 | −0.09 |
ALTd (IU/l) | 0.39 | 0.074 | 47.7 | 0.938 | 11.0 | 254.5 | 445 | 0.19 | 0.32 |
Calcium (mg/dl) | 0.39 | 0.058 | 9.2 | 0.025 | 8.0 | 11.4 | 445 | −0.04 | −0.07 |
Cholesterol (mg/dl) | 0.47 | 0.061 | 123.9 | 1.605 | 58.0 | 348.0 | 445 | −0.12 | −0.06 |
Chloride (mmol/l) | 0.28 | 0.053 | 120.9 | 0.115 | 114 | 144 | 445 | 0.20 | 0.11 |
Creatinine (mg/dl) | 0.25 | 0.050 | 1.3 | 0.013 | 0.8 | 3.1 | 445 | −0.27 | −0.03 |
Glucose (mg/dl) | 0.39 | 0.053 | 83.8 | 0.716 | 44.0 | 222.0 | 444 | 0.04 | 0.16 |
Potassium (meq/l) | 0.25 | 0.041 | 4.3 | 0.016 | 3.4 | 6.7 | 530 | −0.21 | −0.22 |
Sodium (meq/l) | 0.16 | 0.036 | 153.8 | 0.142 | 148.5 | 181.0 | 530 | −0.17 | −0.34 |
Phosphorus (mg/dl) | 0.20 | 0.033 | 5.0 | 0.032 | 3.1 | 7.8 | 530 | −0.06 | 0.06 |
Triglyceride (mg/dl) | 0.13 | 0.034 | 45.8 | 2.350 | 6 | 614 | 444 | −0.09 | 0.03 |
Total protein (g/dl) | 0.38 | 0.054 | 6.7 | 0.024 | 5.3 | 8.4 | 469 | 0.11 | 0.28 |
Urea nitrogen (mg/dl) | 0.23 | 0.034 | 25.1 | 0.243 | 14.2 | 74.3 | 466 | 0.01 | 0.08 |
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
Albumin (g/dl) | 0.61 | 0.046 | 3.2 | 0.015 | 2.0 | 4.6 | 445 | −0.12 | −0.03 |
SAPc (IU/l) | 0.78 | 0.047 | 34.0 | 0.727 | 11.0 | 108.8 | 445 | −0.14 | −0.09 |
ALTd (IU/l) | 0.39 | 0.074 | 47.7 | 0.938 | 11.0 | 254.5 | 445 | 0.19 | 0.32 |
Calcium (mg/dl) | 0.39 | 0.058 | 9.2 | 0.025 | 8.0 | 11.4 | 445 | −0.04 | −0.07 |
Cholesterol (mg/dl) | 0.47 | 0.061 | 123.9 | 1.605 | 58.0 | 348.0 | 445 | −0.12 | −0.06 |
Chloride (mmol/l) | 0.28 | 0.053 | 120.9 | 0.115 | 114 | 144 | 445 | 0.20 | 0.11 |
Creatinine (mg/dl) | 0.25 | 0.050 | 1.3 | 0.013 | 0.8 | 3.1 | 445 | −0.27 | −0.03 |
Glucose (mg/dl) | 0.39 | 0.053 | 83.8 | 0.716 | 44.0 | 222.0 | 444 | 0.04 | 0.16 |
Potassium (meq/l) | 0.25 | 0.041 | 4.3 | 0.016 | 3.4 | 6.7 | 530 | −0.21 | −0.22 |
Sodium (meq/l) | 0.16 | 0.036 | 153.8 | 0.142 | 148.5 | 181.0 | 530 | −0.17 | −0.34 |
Phosphorus (mg/dl) | 0.20 | 0.033 | 5.0 | 0.032 | 3.1 | 7.8 | 530 | −0.06 | 0.06 |
Triglyceride (mg/dl) | 0.13 | 0.034 | 45.8 | 2.350 | 6 | 614 | 444 | −0.09 | 0.03 |
Total protein (g/dl) | 0.38 | 0.054 | 6.7 | 0.024 | 5.3 | 8.4 | 469 | 0.11 | 0.28 |
Urea nitrogen (mg/dl) | 0.23 | 0.034 | 25.1 | 0.243 | 14.2 | 74.3 | 466 | 0.01 | 0.08 |
Values in bold are statistically significant (P < 0.05).
Gender correlations to hematology: negative values represent males' higher means.
AOD—age at death correlations to hematology: negative values represent negative correlation.
SAP, serum alkaline phosphatase.
ALT, alanine transferase.
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
Albumin (g/dl) | 0.61 | 0.046 | 3.2 | 0.015 | 2.0 | 4.6 | 445 | −0.12 | −0.03 |
SAPc (IU/l) | 0.78 | 0.047 | 34.0 | 0.727 | 11.0 | 108.8 | 445 | −0.14 | −0.09 |
ALTd (IU/l) | 0.39 | 0.074 | 47.7 | 0.938 | 11.0 | 254.5 | 445 | 0.19 | 0.32 |
Calcium (mg/dl) | 0.39 | 0.058 | 9.2 | 0.025 | 8.0 | 11.4 | 445 | −0.04 | −0.07 |
Cholesterol (mg/dl) | 0.47 | 0.061 | 123.9 | 1.605 | 58.0 | 348.0 | 445 | −0.12 | −0.06 |
Chloride (mmol/l) | 0.28 | 0.053 | 120.9 | 0.115 | 114 | 144 | 445 | 0.20 | 0.11 |
Creatinine (mg/dl) | 0.25 | 0.050 | 1.3 | 0.013 | 0.8 | 3.1 | 445 | −0.27 | −0.03 |
Glucose (mg/dl) | 0.39 | 0.053 | 83.8 | 0.716 | 44.0 | 222.0 | 444 | 0.04 | 0.16 |
Potassium (meq/l) | 0.25 | 0.041 | 4.3 | 0.016 | 3.4 | 6.7 | 530 | −0.21 | −0.22 |
Sodium (meq/l) | 0.16 | 0.036 | 153.8 | 0.142 | 148.5 | 181.0 | 530 | −0.17 | −0.34 |
Phosphorus (mg/dl) | 0.20 | 0.033 | 5.0 | 0.032 | 3.1 | 7.8 | 530 | −0.06 | 0.06 |
Triglyceride (mg/dl) | 0.13 | 0.034 | 45.8 | 2.350 | 6 | 614 | 444 | −0.09 | 0.03 |
Total protein (g/dl) | 0.38 | 0.054 | 6.7 | 0.024 | 5.3 | 8.4 | 469 | 0.11 | 0.28 |
Urea nitrogen (mg/dl) | 0.23 | 0.034 | 25.1 | 0.243 | 14.2 | 74.3 | 466 | 0.01 | 0.08 |
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
Albumin (g/dl) | 0.61 | 0.046 | 3.2 | 0.015 | 2.0 | 4.6 | 445 | −0.12 | −0.03 |
SAPc (IU/l) | 0.78 | 0.047 | 34.0 | 0.727 | 11.0 | 108.8 | 445 | −0.14 | −0.09 |
ALTd (IU/l) | 0.39 | 0.074 | 47.7 | 0.938 | 11.0 | 254.5 | 445 | 0.19 | 0.32 |
Calcium (mg/dl) | 0.39 | 0.058 | 9.2 | 0.025 | 8.0 | 11.4 | 445 | −0.04 | −0.07 |
Cholesterol (mg/dl) | 0.47 | 0.061 | 123.9 | 1.605 | 58.0 | 348.0 | 445 | −0.12 | −0.06 |
Chloride (mmol/l) | 0.28 | 0.053 | 120.9 | 0.115 | 114 | 144 | 445 | 0.20 | 0.11 |
Creatinine (mg/dl) | 0.25 | 0.050 | 1.3 | 0.013 | 0.8 | 3.1 | 445 | −0.27 | −0.03 |
Glucose (mg/dl) | 0.39 | 0.053 | 83.8 | 0.716 | 44.0 | 222.0 | 444 | 0.04 | 0.16 |
Potassium (meq/l) | 0.25 | 0.041 | 4.3 | 0.016 | 3.4 | 6.7 | 530 | −0.21 | −0.22 |
Sodium (meq/l) | 0.16 | 0.036 | 153.8 | 0.142 | 148.5 | 181.0 | 530 | −0.17 | −0.34 |
Phosphorus (mg/dl) | 0.20 | 0.033 | 5.0 | 0.032 | 3.1 | 7.8 | 530 | −0.06 | 0.06 |
Triglyceride (mg/dl) | 0.13 | 0.034 | 45.8 | 2.350 | 6 | 614 | 444 | −0.09 | 0.03 |
Total protein (g/dl) | 0.38 | 0.054 | 6.7 | 0.024 | 5.3 | 8.4 | 469 | 0.11 | 0.28 |
Urea nitrogen (mg/dl) | 0.23 | 0.034 | 25.1 | 0.243 | 14.2 | 74.3 | 466 | 0.01 | 0.08 |
Values in bold are statistically significant (P < 0.05).
Gender correlations to hematology: negative values represent males' higher means.
AOD—age at death correlations to hematology: negative values represent negative correlation.
SAP, serum alkaline phosphatase.
ALT, alanine transferase.
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
O2 saturation (%) | 0.589 | 0.035 | 83.31 | 0.312 | 53.8 | 96.2 | 629 | −0.033 | −0.137 |
pCO2 (mmHg) | 0.452 | 0.047 | 45.65 | 0.191 | 32.0 | 64.5 | 629 | −0.181 | −0.114 |
Total CO2 (meq/l) | 0.425 | 0.045 | 23.24 | 0.087 | 14.4 | 29.7 | 629 | −0.364 | −0.260 |
HCO3−(meq/l) | 0.398 | 0.044 | 21.63 | 0.087 | 12.8 | 29.4 | 629 | −0.357 | −0.251 |
pO2 (mmHg) | 0.396 | 0.036 | 58.79 | 0.379 | 35.0 | 92.5 | 629 | 0.083 | −0.033 |
Base excess (meq/l) | 0.272 | 0.044 | −3.89 | 0.094 | −15.7 | 7.4 | 629 | −0.351 | −0.248 |
pH (units) | 0.253 | 0.037 | 7.29 | 0.002 | 7.09 | 7.40 | 629 | −0.251 | −0.164 |
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
O2 saturation (%) | 0.589 | 0.035 | 83.31 | 0.312 | 53.8 | 96.2 | 629 | −0.033 | −0.137 |
pCO2 (mmHg) | 0.452 | 0.047 | 45.65 | 0.191 | 32.0 | 64.5 | 629 | −0.181 | −0.114 |
Total CO2 (meq/l) | 0.425 | 0.045 | 23.24 | 0.087 | 14.4 | 29.7 | 629 | −0.364 | −0.260 |
HCO3−(meq/l) | 0.398 | 0.044 | 21.63 | 0.087 | 12.8 | 29.4 | 629 | −0.357 | −0.251 |
pO2 (mmHg) | 0.396 | 0.036 | 58.79 | 0.379 | 35.0 | 92.5 | 629 | 0.083 | −0.033 |
Base excess (meq/l) | 0.272 | 0.044 | −3.89 | 0.094 | −15.7 | 7.4 | 629 | −0.351 | −0.248 |
pH (units) | 0.253 | 0.037 | 7.29 | 0.002 | 7.09 | 7.40 | 629 | −0.251 | −0.164 |
Values in bold are statistically significant (P < 0.05).
Gender correlations to hematology: negative values represent males' higher means.
AOD—age at death correlations to hematology: negative values represent negative correlation.
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
O2 saturation (%) | 0.589 | 0.035 | 83.31 | 0.312 | 53.8 | 96.2 | 629 | −0.033 | −0.137 |
pCO2 (mmHg) | 0.452 | 0.047 | 45.65 | 0.191 | 32.0 | 64.5 | 629 | −0.181 | −0.114 |
Total CO2 (meq/l) | 0.425 | 0.045 | 23.24 | 0.087 | 14.4 | 29.7 | 629 | −0.364 | −0.260 |
HCO3−(meq/l) | 0.398 | 0.044 | 21.63 | 0.087 | 12.8 | 29.4 | 629 | −0.357 | −0.251 |
pO2 (mmHg) | 0.396 | 0.036 | 58.79 | 0.379 | 35.0 | 92.5 | 629 | 0.083 | −0.033 |
Base excess (meq/l) | 0.272 | 0.044 | −3.89 | 0.094 | −15.7 | 7.4 | 629 | −0.351 | −0.248 |
pH (units) | 0.253 | 0.037 | 7.29 | 0.002 | 7.09 | 7.40 | 629 | −0.251 | −0.164 |
Variable | h2 | h2 SE | Mean | Mean SE | Min | Max | N | Gendera | AODb |
O2 saturation (%) | 0.589 | 0.035 | 83.31 | 0.312 | 53.8 | 96.2 | 629 | −0.033 | −0.137 |
pCO2 (mmHg) | 0.452 | 0.047 | 45.65 | 0.191 | 32.0 | 64.5 | 629 | −0.181 | −0.114 |
Total CO2 (meq/l) | 0.425 | 0.045 | 23.24 | 0.087 | 14.4 | 29.7 | 629 | −0.364 | −0.260 |
HCO3−(meq/l) | 0.398 | 0.044 | 21.63 | 0.087 | 12.8 | 29.4 | 629 | −0.357 | −0.251 |
pO2 (mmHg) | 0.396 | 0.036 | 58.79 | 0.379 | 35.0 | 92.5 | 629 | 0.083 | −0.033 |
Base excess (meq/l) | 0.272 | 0.044 | −3.89 | 0.094 | −15.7 | 7.4 | 629 | −0.351 | −0.248 |
pH (units) | 0.253 | 0.037 | 7.29 | 0.002 | 7.09 | 7.40 | 629 | −0.251 | −0.164 |
Values in bold are statistically significant (P < 0.05).
Gender correlations to hematology: negative values represent males' higher means.
AOD—age at death correlations to hematology: negative values represent negative correlation.
Erythroctyes
All 4 erythrocyte variables were significantly (P < 0.05) heritable, and low correlations (P < 0.05) to gender were found for all 4 variables. Erythrocyte count, hemoglobin, and packed cell volume correlated positively with female gender, whereas mean cell volume correlated positively with male gender. Mean cell volume had low positive correlation (P < 0.05) with age at death, whereas erythrocyte count and hemoglobin had low negative correlation (P < 0.05) with age at death (Table 1).
Clinical Chemistry
All 14 clinical chemistry variables were significantly (P < 0.05) heritable, and low correlations (P < 0.05) with gender, of mixed male–female rank, were found for 10 of the variables. Only serum calcium, glucose, phosphorus, and urea nitrogen were independent of gender. Low correlations (P < 0.05) with age at death were found for alanine transferase, chloride, glucose, potassium, sodium, and total protein; correlations were negative for potassium and sodium (Table 2).
Acid–Base Metrics
All 7 venous variables were significantly (P < 0.05) heritable. Low correlations (P < 0.05) with gender were found for all variables except oxygen saturation. Males had generally higher values for gender correlations. Low negative correlations (P < 0.05) with age at death were found for all variables except pO2 (Table 3).
Discussion
As in other mammals, erythrocyte, clinical chemistry, and acid–base profiles in cats are mostly heritable, implying a substantial degree of genetic variability. These profiles are designed as screening panels to reflect organ or metabolic dysfunction. The underlying principle is individual expression relative to a reference interval that separates “normal” from “abnormal.” Despite their origin and uses, many of these blood and serum variables can be relatively insensitive to covert changes until more serious or extensive damage has occurred. In addition, individual variation within reference intervals can at times denote pathological change, whereas individual variation outside of reference intervals can at times still reflect normality.
There exist also a number of well-known procedural challenges and caveats in the interpretation of these data. Reference intervals for each variable typically are established at each laboratory and usually are based on analytical survey of “clinically healthy” subjects (Ash 1980). Potential problems with this practice include 1) some outwardly normal individuals may harbor covert abnormalities; 2) the statistical structure of reference intervals can be based on percentiles, quartiles, standard deviations, or occasionally the lowest-to-highest data range within the sampled population, making comparative evaluations among laboratories more difficult; and 3) some variables do not exhibit Gaussian distribution, which must be considered during interpretation. It is readily apparent that a degree of deviation from reference is inevitable in routine metabolic profiling (Ash 1980; Wilkins and Hurvitz 1980; Rebar and Boon 1983).
Based on current technology, the most appropriate means of providing contiguity in health care is annual individual screening for baseline reference. In population studies, including biomedical research, it is important to realize that pooled data from different laboratories potentially introduce biases resulting from differing sampling conditions, analytical methods, equipment, and reference intervals. Therefore, even when optimal genetic controls exist, such as comparing data from monozygotic and dizygotic twins, other biases remain and need to be considered.
Given the numerous procedural and analytical difficulties that are known to exist in clinical chemistry, robustness is suggested by the relative consistency among heritability studies of these variables, across populations, within species, and across species. Nonetheless, physiological implications of the phenotypic expressions of variables with similar heritabilities among species may be very different. Several important examples may be considered.
Alkaline Phosphatase
Serum alkaline phosphatase expression was highly heritable (0.78) in our cat database. Alkaline phosphatases are a group of membrane-bound glycoproteins that are present in the body as multiple isoenzymes. Tissue expression of specific isoenzymes appears to result from posttranslational modification (Siraganian et al. 1989). Multiple isoenzymes in serum most probably reflect clearance from tissues (Kruse 1983; Siraganian et al. 1989). Distribution of alkaline phosphatase in reference populations typically is non-Gaussian, and there are gender and age effects (Lewis 1987). It has been reported that feline liver/bone/kidney alkaline phosphatase is 88–90% homologous to other mammals (Ghosh and Mullins 1995). Separate loci code for human liver/bone/kidney, placental, and intestinal isoenzyme. Placental/intestinal isoenzymes are homologous and located on chromosome 2, whereas the liver/bone/kidney isoenzyme locus, common to both mouse and humans, is located on human chromosome 1 (Ghosh and Mullins 1995).
Evaluations of different human test groups yielded alkaline phosphatase heritability ranging between 0.00 and 0.81 (Havlik et al. 1977; Colletto et al. 1981; Whitfield and Martin 1983, 1984; Kalousdian et al. 1987) and included different metholodology-associated outcomes within the same study (Kalousdian et al. 1987), thus emphasizing the importance of considering influences of data collection procedures. In humans and Siberian Husky dogs, extreme but benign overexpression of serum alkaline phosphatase has been described as a presumed recessive trait (Iancu et al. 1978; Dunn et al. 1979; Wilson 1979; Kruse 1983; Siraganian et al. 1989; Panteghini 1991; Lawler et al. 1996) that has not been reported in cats or other mammalian species. It has been recognized for many years that alkaline phosphatase expression in cat serum tends to be substantially lower than in dogs and humans, even during pathological states (Everett et al. 1977; McLain et al. 1978; Horney et al. 1992). Thus, smaller changes in serum expression of alkaline phosphatase in cats must be interpreted as having much greater physiological significance, compared with other species.
Glucose
The exaggerated hyperglycemic response of domestic cats to short-term stress has been recognized for many years (Kaneko 1980; Thibault and Roberge 1988). Momentary excitement in a sampling room, which is very common in cats, and short-term effects of unfamiliar cage confinement, both are sufficient to produce dramatically increased circulating glucose in cats (Thibault and Roberge 1988). In our study of cats, heritability of serum glucose expression was moderate (0.39) in a well-controlled environment. In one study of Finnish human twins, heritability of fasting glucose expression was only slightly greater, at 0.45 (Katoh et al. 2005). A study of fasting plasma glucose also was done with Brazilian twins, with heritability of glucose expression reported as 0.73 (Colletto et al. 1981). In post–glucose loading measurements conducted with twins at 5 American centers, heritability of plasma glucose expression postloading was 0.88 in a first examination and 0.52 in a second examination (Kalousdian et al. 1987).
It is interesting that the cat population had lower heritability for fasted serum glucose expression than several human twin populations, but the human populations also revealed widely differing within-species results. These observations suggest again that study design and data collection procedures are very important considerations for the outcomes of heritability studies within populations, even if these populations are selected and structured to favor rigorous procedures. In addition, it is understood widely by veterinary clinicians that interpretation of elevated serum fasted glucose expression in domestic cats depends substantially on physiological status and the immediate environment at the time of sampling. Thus, results of serum glucose determinations in cats are subject to a somewhat different range of possible interpretations than are the same results in humans.
Creatinine
Serum creatinine expression is a function of spontaneous breakdown of skeletal muscle. Creatinine production is constant, resulting from unidirectional nonenzymatic dehydration of skeletal muscle creatine (First 2003). Elevations of serum creatinine most frequently are associated with renal compromise but can be associated also with changes such as state of hydration or declining lean body composition. Because many of the changes in serum creatinine occur during advanced life, it is useful but also potentially confusing to consider heritability evaluations of older subjects. In a study of elderly Danish twins (Bathum et al. 2004), elevated values predicted shorter overall survival, as expected, and heritability of serum creatinine expression ranged from 0.00 in men to 0.44 in women. In another large study of British twins between ages 18 and 72 years, heritability of serum creatinine expression was 0.37 (Hunter et al. 2002). Serum creatinine heritability was 0.25 in our study of cats, which included all adult ages and both genders, and there was no correlation to age at death.
Among mammalian species, serum creatinine has a non-Gaussian distribution, tends to occur over a relatively narrow numerical range among healthy individuals, and tends to be interpreted similarly among species. Thus, the comparative data might suggest that the environmental component in cats might be of somewhat greater importance than the environmental component in humans. However, such an interpretation would need to be taken very cautiously because of variation in the heritability results among human populations. These differences suggest that the human populations that were selected and evaluated likely were subject to differing demographic influences, whereas our cats all were maintained for life under uniform environmental conditions. It is interesting in this respect that results of a study of young adult twins suggested that fetoplacental factors were more important to expression of serum creatinine than were genetic and maternal factors alone (Gielen et al. 2005).
Cholesterol and Triglyceride
A final important example is the large difference in heritability between the serum lipid fractions total triglyceride (0.13) and total cholesterol (0.47) in cats. It has been shown in diet restriction trials across species that circulating total cholesterol is less influenced than is circulating total triglyceride (Liepa et al. 1980; Masoro et al. 1983; Choi et al. 1988; Van Liew et al. 1993; Bodkin and Hansen 1995; Cefalu et al. 1997; Verdery et al. 1997; Edwards et al. 1998; Lane et al. 1998; Kealy et al. 2002), suggesting that expression of serum cholesterol might be less subject to environmental influence than serum triglyceride.
Plasma lipids are known to change significantly with age (Greenfield et al. 2004), although cross-sectional heritability studies usually do not evaluate longitudinal changes (Snieder et al. 1999). Genetic and environmental effects may exert influence at different ages, either due to changing magnitude of the effect or to age-related changes in gene expression (Snieder et al. 1999). In most studies of humans, more than 50% of the variation in serum total cholesterol and triglyceride is genetically mediated (Snieder et al. 1999), which is compatible with our findings for expression of serum cholesterol but not triglyceride in cats.
There exists significant species difference between cats and humans with respect to implications of levels of serum lipid expression. Despite the relatively high heritability of serum cholesterol expression in cats (0.47) and humans (0.35–0.72, summarized by Heller et al. 1993), cardiovascular disease secondary to athersosclerosis is a very common complication of chronic hypercholesterolemia in humans but has not been reported in cats (Fox et al. 2005). On the other hand, hepatic lipidosis and associated hepatic failure in cats is a well-known and serious complication of dysfunctional triglyceride metabolism that seems to occur uniquely in cats (Scherk and Center 2005).
The foregoing comparative examples suggest that epidemiological factors that operate within and among specific populations can have substantial impact on outcomes of heritability studies. Therefore, comparing species heritabilities needs to be approached very cautiously. Our population, from which these data were derived, was maintained in a stable and healthy state over nearly 4 decades for the purpose of conducting nutritional research. Better populations might be structured for specific genetic purposes, and our results may be used as a guide for design and interpretation of future experiments.
References
Author notes
Corresponding Editor: Stephen J. O'Brien