Relationship Between Angiotensin-Converting Enzyme Gene Polymorphism and Body Composition, Functional Performance, and Blood Biomarkers in Advanced Cancer Patients

An intriguing line of investigation into the determinants of nutritional and functional characteristics of patients with chronic diseases (1, 2), which present with various degrees of cachexia and/or sarcopenia, has focused on a particular polymorphism in the human genotype: the insertion/deletion in angiotensin-converting enzyme (ACE) gene. Located on chromosome 17, the ACE gene contains a RFLP, which is characterized by the presence (insertion “I” allele) or absence (deletion “D” allele) of a 287 bp sequence at intron 16 in the Caucasian population (3). Given that each gene has two alleles, there are three resulting combinations: insertion/insertion (II), insertion/deletion (ID), and deletion/deletion (DD). For reasons that cannot yet be fully explained, the presence of the I allele gives rise to lower ACE activity in serum (3) and tissues (4), whereas those individuals with the DD allele typically have higher circulating and endogenous ACE levels (5).

ACE plays a critical role in the renin-angiotensin system by catalyzing the conversion of the inactive angiotensin I to angiotensin II, which is the physiologically active form of the hormone. Acute and chronic exposure to angiotensin II in animal models was associated with weight loss, anorexigenic effects on the central nervous system, enhanced protein breakdown in skeletal muscle mass, and increased peripheral oxygen consumption (6–8). Angiotensin II levels are elevated in patients with advanced stages of chronic diseases such as cirrhosis, congestive heart failure, and chronic obstructive pulmonary diseases, where body wasting or cachexia are of significant clinical concern (8, 9). Whether a direct link or connection exists among the polymorphic alleles of the ACE gene and the nutritive, functional, and blood biomarkers associated with the development of cancer cachexia is only speculative at the present time. Therefore, more information must be obtained to characterize these outcome measures in patients according to their angiotensin-converting gene polymorphism (ACEGP). The aim of this study was to determine whether the ACEGP is associated with nutritional and performance characteristics, as well as with specific routine blood markers in advanced cancer patients (ACP).

An inception cohort of 172 ACP, ages 18 y or older, newly or recently diagnosed (within 6 mo of diagnosis), with stage III (inoperable) to IV non–small cell lung cancer and unresectable, metastatic gastrointestinal cancers were recruited from within the McGill University Health Centre. Recruitment took place between March 2006 and July 2007. After scientific review and ethical approval of the protocol by the McGill Institutional Review Board, patients within this population were considered eligible and their written informed consent was obtained before inclusion in this study.

The following domains were assessed at the time of patient accrual: (a) Demographic and clinical: age, gender, primary and secondary tumor sites, concurrent treatments (radiotherapy and chemotherapy), and use of medications (ACE inhibitors, anti-inflammatories, and statins); (b) Blood biomarkers: cell blood counts and differential, serum albumin, ACE, and C-reactive protein; (c) Nutritional assessments and body composition: patient-generated subjective global assessment (PG-SGA) for nutritional status (10) with weight, height, and its derived body mass index (BMI; ref. 11), body composition using Dual-energy X-ray Absorptiometry (DXA; Lunar Prodigy Advance; GE Healthcare) to obtain total fat mass (TFM), percent body fat, lean body mass, and appendicular lean mass; (c) Strength measurements: handgrip force measured by Jamar dynamometry (12); and (d) Genotyping: determination of the polymorphic alleles of the ACE gene.

All assessments, except for the DXA, were completed in a hospital setting (i.e., inpatient wards or outpatient clinics). A subgroup of patients (n = 64) who were willing and capable of attending a human laboratory center outside the hospital had a DXA scan done at the McGill Nutrition and Performance Laboratory.⁶

For the determination of the polymorphism status in the ACE gene, buffy coat serum and plasma from 7 mL blood sample was used to extract DNA using standard laboratory procedures. The quantity and quality of DNA was verified by absorbance. The DNA was genotyped for the ID polymorphism in the ACE gene using a three-primer system as previously described (13). This system results in amplification products of 525 and 155 bp when the insertion is present (ACE allele I) and a single 237 bp product for the deletion allele (ACE allele D). Randomly selected samples were repeated as quality control, and two observers independently scored the results. The serum ACE concentration was measured by spectrophotometric method (14).

Statistical analysis. All statistical analyses were done using SPSS version 14.0 statistical software. One-way ANOVA (Tukey post hoc) and χ² tests were respectively used to compare means and percentages of the nutritional, functional, and biological characteristics across the three allele groups. To determine independent relationships between the ACEGP and the nutritional, functional, and blood biomarker characteristics, we did multiple linear regression analyses adjusting for the following potential confounders: gender, age, diagnosis (whether patients had a diagnosis of lung or gastrointestinal malignancies), current oncological treatment (whether patients were receiving radiotherapy or chemotherapy at the time of assessment), previous oncological treatments (whether patients had received chemotherapy in the 6 mo before assessment; none of our study patients had received surgery or radiotherapy), survival (presence/absence of death at 8 wk after assessment), and medications (whether patients were taking one of three types of medications: ACE inhibitors, anti-inflammatories, or statins at the time of assessment).

Patients across the three allele groups were homogeneous in terms of demographics, and clinical and treatment characteristics (Table 1). No statistical differences were found across groups (II, ID, DD) for radiotherapy, chemotherapy, and medication use when these variables were analyzed separately. Given that it is possible that individual study subjects could be receiving more than one of the treatments or medications, Table 1 cannot be used to explore how the groups compare in terms of combined radiotherapy, chemotherapy, and medication use. However, to assess the possible confounding effect of both current and previous oncological treatments (i.e., chemotherapy and/or radiotherapy) and medication use (i.e., ACE inhibitors, anti-inflammatories, or statins), we controlled for each of these variables in the multiple regression analysis. No confounding effect was found for either oncological treatments or medication use.

Bivariate analysis (Table 2) did reveal significant differences (P < 0.05) in weight, BMI, TFM, handgrip force, and serum ACE between the three allele groups. Both weight and BMI associated with the ID allele patient group were significantly lower compared with the II group. Patients with ID alleles had significantly lower TFM than patients with the DD allele. Handgrip force percentiles, adjusted for age and gender, and serum ACE levels were significantly higher in patients with the DD allele than patients with the II ACEGP. Interesting trends were observed for both WBC and C-reactive protein levels, which seemed to increase consistently with the presence of either the monozygous or heterozygous D alleles.

In the multivariate analysis (Table 3), where the II polymorphism was considered as the reference allele (15), higher ACE (P = 0.01) and WBC count (P = 0.06) levels were independently associated with the DD ACEGP. The latter group had almost a 7-pound greater handgrip force when compared with the II patients (P = 0.03). Higher hemoglobin (Hb) levels were found to be associated with both ID (P = 0.05) and DD (P = 0.07) alleles. When compared with the II alleles, patients with ID ACEGP presented with lower BMI (P < 0.01), TFM (P = 0.05), percent body fat (P = 0.05), and lean body mass (P = 0.07), as measured by DXA. No significant differences were observed for any measures recorded in the PG-SGA across the three alleles.

To the best of our knowledge, this is the first study to ascertain the effect of the ACEGP on nutritional and functional characteristics of ACP within a translational and multidimensional model. The model included tumor characteristics, concurrent tumor treatments, comorbidities, concurrent medications, and time to death as possible confounders of the effect of ACEGP on nutritional and performance characteristics of ACP. Finally, an inception cohort of newly diagnosed ACP was identified and recruited to study the clinical phenotypes of the ACEGP.

Our study has four major findings. First, patients with the ID allele showed significantly lower BMI, TFM, and percent body fat with a trend (P = 0.07) for lower lean body mass compared with those with the II allele. Second, patients with the DD allele presented with both higher levels of circulating ACE and greater muscle strength compared with II patients. Third, patients with II alleles presented with significantly lower levels of serum Hb. Lastly, none of three allele combinations correlated with differences in subjective measures of nutrition and performance as measured by the PG-SGA.

Differences in body composition. The variations found in body composition measures in relation to ACEGP are consistent with the findings of several studies. Katsuya et al. (16) found an increased BMI associated with ACE I allele in normal subjects. Kritchevsky et al. (15) studied a large cohort of well-functioning community-dwelling older adults and found increased adiposity, as well as more intramuscular thigh fat in patients with II ACEGP, but a similar BMI to individuals with ID-DD ACEGP. Normal subjects with II ACEGP were shown to have improved endurance capability and greater anabolic response during structured exercise training programs, and this may allow for relative sparing and/ or increase of fat stores (17). Montgomery and colleagues (17) hypothesized that at the base of this greater metabolic efficiency, there may be associations between the ACEGP and: (a) growth hormone gene, (b) alterations of appetite and tolerance to physical effort in the central nervous system, (c) modification of the absorptive capability of the gut, and (d) alteration of the mobilization and use of fatty acids via local adipose and skeletal muscle renin-angiotensin systems. ACE inhibition associated with I allele increases insulin sensitivity as manifested by greater suppression of non-esterified fatty acid flux in adipocytes, as well as higher glucose uptake and glycogen stores in skeletal muscle (17, 18). In addition, use of the ACE inhibitor, perindopril, has been shown to improve exercise capacity, equivalent to that reported after 6 months of exercise training, in a group of functionally impaired elderly (19).

In our study, differences in body mass were found to be associated with a particular ACEGP rather than with ACE serum levels. We do not believe that this finding is a result of methodologic limitations in measuring ACE serum levels or body composition. Serum ACE concentration was measured by a well-validated method (14). In addition, as expected, when the levels of serum ACE were compared across the three genotype classes, a definite trend was found between the groups, with the presence of the DD genotype having a statistically significant additive effect on serum ACE compared with the II genotype. With respect to the determination of measures of body composition, our group has confirmed the precision or test-retest reliability of DXA for assessing body composition in ACPs and have reported coefficients of variation of 1.56% for fat mass and 0.79% for fat-free mass (20). Furthermore, we have attempted to statistically control for all possible confounders of the observed differences in BMI and body composition outcome measures, thus eliminating alternate explanations for the reported relationship between these measures and ACEGP. Therefore, the results of our analysis indicate that the body mass differences in the ID group, as determined by multivariate analysis, were not explained by age, gender, type of diagnosis, concurrent and previous oncological treatments, and proximity to death (as expressed by presence/absence of death within 8 weeks from assessment).

Previous investigators (21) have commented that the ACE ID polymorphism is in a noncoding region of the gene, and it is therefore unlikely that it is directly responsible for changes in phenotype. However, all of the studies done in Caucasian populations have indicated that the majority of the polymorphisms in the candidate region of the gene are tightly linked with the ID polymorphism. Therefore, our data may support the presence of an unknown functional polymorphism(s) linked to the ID polymorphism, which are associated with body composition differences that are independent of ACE serum levels.

Differences in strength and functional performance. In this study, patients with DD ACEGP presented with both higher levels of circulating ACE and greater forearm muscle strength compared with II patients. The elevation in serum ACE could reflect a higher production in angiotensin II, thereby explaining the greater force production in the upper body. Again, these findings are consistent with most studies, which examined the relationship between ACEGP and strength performance. Subjects with the homozygous DD allele showed greater strength enhancement as a consequence of resistance exercise training programs (17). Associated with the D allele is the higher production of ACE and angiotensin II, which not only has been shown to induce cardiac and smooth muscle hypertrophy resulting in a possible risk of developing hypertension but may also play a role in the exercise-induced hypertrophy of skeletal muscle, specifically of the type-II fast-twitch fibers (17). These have been advocated as possible mechanisms to explain the greater strength gains associated with this allele (22). In the present study, the greater strength production at a time when lean muscle mass is decreasing would suggest that the ability to recruit an adequate number of α-motor units is paramount to maintain force production. Thus, skeletal muscle quality (force output/muscle mass) and not absolute skeletal muscle quantity could be an important trait to consider in our ACP with the DD allele.

We did not observe any significant differences in strength between DD and ID or between II and ID. Seminal studies looking at ACE serum levels (3) or changes in body composition associated with ACE genotype (15, 17) have shown similar characteristics in study subjects with ID and DD genotypes compared with those with II genotype. Therefore, our multivariate analysis compared study subjects with one or two D alleles to those with the II genotype. Our results for both univariate and multivariate analyses confirmed significant differences in serum concentration of ACE between DD and II groups, but not between II and ID or between ID and DD. Similarly, significantly greater muscle strength was found in the DD compared with the II group, with the former clearly demonstrating a codominant allele effect for ACE serum concentration. The ID patients, who did not present with significantly different concentrations of serum ACE compared with both II and DD, did not differ in muscle strength compared with those two groups.

Differences in serum levels of Hb. Patients with II allele presented with lower levels of Hb compared with patients with ID or DD genotypes. The association of anemia with the I allele is consistent with the work of Yaren and colleagues (23), who found that patients with the II and ID ACEGP had more frequent anemia at the time of their diagnosis of non–small cell lung cancer. Different fields of medicine have shown that ACEGP is involved in erythropoiesis through different mechanisms. For instance, the renin-angiotensin system stimulates the hematopoietic progenitor cell and induces the growth of early erythroid progenitors (24, 25). ACE inhibitors and angiotensin II receptor antagonist might suppress erythropoiesis (26, 27). In peritoneal dialysis patients, genotype II is associated with both lower ACE levels and higher EPO dose requirements (28, 29). Our findings, which support an association between lower levels of ACE and lower serum Hb levels, indicate an additional contributing factor to anemia in ACP, which is normally not mentioned or investigated.

Differences in subjective measures of nutrition and performance. Despite clear differences in body composition and strength, no significant differences in subjective measures of nutrition and performance as recorded by the PG-SGA were identified. Several reasons may account for these findings. First, we may have not selected the appropriate measure to determine the translational value of ACEGP. For example, most studies have determined the effect of ACEGP in response to exercise. It is possible that the ACEGP may play a role in the response to specific treatments such as rehabilitation, nutritional interventions, or a combination of both. Second, objective rather than subjective measures of performance, such as changes in cardiovascular capacity, could have better captured variations across ACEGP. Finally, it is possible that changes occurring with advanced cancer may override the possible effect of ACEGP on the subjective measures reported in the PG-SGA.

The current study has some limitations. First, this study may be underpowered and unable to detect statistically significant differences, especially for those variables with a wide range of SDs. This may be the case for the inflammatory markers, such as WBC and C-reactive protein, which showed clear positive trends across the three allele combinations, with DD ACEGP showing the highest plasma levels for both markers. However, we neither identified nor considered these particular inflammatory markers among our primary outcome variables (compared with nutritional and performance measures). For the nutritional measures, such as the one reported in the PG-SGA, which did not present with significant differences across the three allele combinations, confidence intervals seemed to be symmetrically distributed around the null value. This statistical observation suggests a lower probability of finding significant differences with larger samples.

Fewer patients were assessed for body composition using DXA because they were either unwilling or unable to travel to our human laboratory testing center. Most of our findings using the DXA measurement are consistent with the literature, particularly for the difference in fat mass across the different allele combinations. The strong trend toward the presence of lower lean body mass in patients with the ID alleles represent a novel finding from our study and needs to be confirmed in larger data sets.

In the present study, only upper limb muscle strength was measured using Jamar dynamometry. The Jamar handgrip dynamometer was found by our group to be a precise and consistent method for measuring strength in ACPs (20). Whether this single measure of isometric strength is representative of whole body strength is currently unknown. However, in a companion study from our laboratory, we have shown strong and significant correlations between measures of upper and lower muscle strength (r = 0.72; P < 0.05) and between Biodex and Jamar recordings of isometric dynamometry (r = 0.82; P < 0.05) in the same cohort of patients.⁷

The changes in body composition and muscle strength related to the ACEGP might have relevant clinical consequences. Considering the importance that clinicians place on providing better or more effective treatment strategies and supportive care in the ACP, a more comprehensive understanding of specific determinants of body composition, muscle strength, and blood biomarkers known to influence the health and well-being of the patient over their cancer trajectory would be of ultimate value. This study represents the first step to determine if selective measures previously shown to identify a negative prognosis and cachexia-like traits in a cancer population are associated with specific polymorphic alleles of the ACE gene. For example, by recognizing those patients who possess the ID genotype and have lower body composition measures of BMI, lean body mass, TFM, and percent body fat than those with the II genotype may help direct the physician to more closely monitor the ID patients for weight loss and perhaps to consider recommending specific nutritional and/or functional interventions (e.g., resistance training activities) earlier in the course of their cancer. Furthermore, the finding that lower Hb levels are associated with the II allele creates a potential dilemma for those patients prescribed ACE inhibitors. These medications have been studied for attenuating muscle catabolism secondary to higher levels of angiotensin II. The use of ACE inhibitors may potentially aggravate anemia while decreasing muscle loss in ACP. Clinical studies designed to investigate the role of ACE inhibitors in limiting or preventing muscle loss in cancer cachexia should also examine the potential negative effect that these medications have on Hb levels, particularly in patients who are genetically predisposed to have lower levels of serum ACE.

These preliminary findings need confirmation in large cohorts of patients and other primary tumor sites. Further studies are ongoing in our laboratory to explore the effect of ACEGP on key inflammatory markers associated with cancer cachexia. As an extension of this study, we will examine the effectiveness of physical rehabilitation alone or in conjunction with nutritional interventions in ACPs grouped according to their ACEGP. These studies may ultimately prove the clinical role of ACEGP in both assessing and treating ACP for ameliorating or maintaining their nutrition, performance, and quality of life.

No potential conflicts of interest were disclosed.

We thank Jean-François Théberge for his invaluable technical support and Drs. Vickie Baracos, Marina Mourtzakis, Mazen Hassanain, and Lawrence Joseph for their critical commentary of earlier drafts of this manuscript.