Serum Insulinlike Growth Factor 1 as Possible Marker for Risk and Early Diagnosis of Parkinson Disease

With a prevalence of about 2% in the population older than 60 years, Parkinson disease (PD) is one of the most common neurodegenerative disorders. Parkinson disease affects the whole brain; however, the typical symptoms (bradykinesia, rigidity, and resting tremor) are caused by a progressive loss of dopaminergic neurons in the substantia nigra (SN). In view of a steadily increasing life expectancy, new strategies are warranted to protect neurons from early degeneration.¹ Therefore, much effort is put into determining markers to identify individuals at risk to develop PD before the first clear motor symptoms allow clinical diagnosis.²

Many symptoms, such as depression, loss of olfactory function, rapid eye movement sleep behavior disorder, or mild motor abnormalities (eg, unilaterally reduced arm swing) may antecede the first clear motor symptoms by several years; however, none of these premotor symptoms is specific for the development of PD.³ Functional neuroimaging of the presynaptic dopaminergic system in the basal ganglia using positron emission tomography or single-photon emission computed tomography may display reduced dopaminergic function in early disease stages⁴ and even in some people with putative premotor symptoms.⁵ However, because of the high cost and limited availability, functional neuroimaging is available to a small number of people only. Transcranial B-mode sonography (TCS) is a widely available and inexpensive technique that enables the detection of SN hyperechogenicity as a sensitive and fairly specific marker for the early diagnosis of PD.⁶ Moreover, it has been demonstrated⁷ that SN hyperechogenicity affects approximately 10% of healthy individuals and is more commonly found in those with putative premotor symptoms of PD, such as depression, loss of olfactory function, or mild motor abnormalities. Additionally, SN hyperechogenicity was found^{7 - 8} to be related to dopaminergic dysfunction as measured by fluorodopa F 18 positron emission tomography in approximately 2 of 3 clinically affected yet healthy people. However, no neuroimaging technique allows the diagnosis of PD in the premotor state with certainty.³ Elderly people are more likely to show reduced motor function and a variety of the nonmotor symptoms associated with PD; this presents a challenge in the differentiation between normal aging and early PD.⁹ Therefore, other biological markers are needed to help in the identification of people at risk to develop PD.

Insulinlike growth factor 1 (IGF-1) is deeply involved in many processes concerning growth and vitality. It is mainly secreted from the liver after stimulation of the somatotropic axis. Serum IGF-1 levels are reduced with increasing age.^{10 - 11} Several studies^{12 - 15} have shown that, in patients with PD, serum IGF-1 levels are higher than those in a healthy control population. In earlier studies,¹⁶ increased serum IGF-1 levels in PD were attributed to dopaminergic medication, which is known to induce IGF-1. However, a recent conceptional study¹² demonstrated that serum IGF-1 levels are also elevated in patients who have never received dopaminergic medication. It may therefore be hypothesized that increased serum IGF-1 in early PD might be related to the disease process.^{12 - 13}

Based on the results of this previous study¹² we set out to (1) determine whether increased serum IGF-1 level is a potential marker for early PD in an independent cohort, (2) assess the association between serum IGF-1 level and motor function in healthy people, and (3) investigate whether serum IGF-1 level may help in the identification of individuals putatively at increased risk for PD. In addition, IGF-1 binding protein 3 (IGFBP-3) was evaluated.

We evaluated 139 people (99 men and 40 women) without diagnosis of a neurodegenerative disorder (defined as healthy ; aged 50-80 years; mean [SD] age, 60.6 [6.6] years) and 15 consecutive patients of our outpatient department with newly diagnosed idiopathic PD (mean [SD] age, 69.0 [8.3] years; 9 men and 6 women) who had never received dopaminergic medication. All healthy controls were participants in a large epidemiologic study⁹ addressing the early diagnosis of PD (Prospective Validation of Risk Factors for Parkinsonian Syndromes [PRIPS] study). Parkinson disease, according to the UK Brain Bank Criteria,¹⁷ was excluded in these healthy people by a movement disorders specialist. In both groups we did not include individuals with disorders or medication interfering with serum IGF-1 levels, ie, diabetes mellitus, body mass index greater than 30 (calculated as weight in kilograms divided by height in meters squared), pituitary or thyroid disease, acute or chronic inflammatory diseases, or cancer, as well as use of β-blocker, corticosteroid, neuroleptic, or hormone replacement medications. In addition, individuals with insufficient temporal bone windows to perform TCS were not included. None of the participants followed a specific diet. All participants gave written informed consent, and the study was approved by the local ethics committee.

Motor function was assessed in all PD patients and healthy subjects by a masked movement disorders specialist (J.G. or K.B.), using the motor part of the Unified Parkinson's Disease Rating Scale, Part III (UPDRS-III). To date, no validated scale for quantitative assessment of motor function in healthy elderly people is available. Therefore, we used the UPDRS-III, which can be regarded as the current standard of practice because it has been commonly used in recent epidemiologic studies^{18 - 21} in healthy elderly individuals. A UPDRS-III score of 0 points was defined as normal motor function (M−); scores higher than 0 points were defined as altered motor function (M+).

To date, no clear criteria for the classification of healthy persons at risk for PD are available. In the present study, we used 2 independent criteria to determine a person's at-risk status. First, he or she was expected to exhibit altered motor function, as measured by the UPDRS-III (M+). Because the UPDRS-III is not specific for motor abnormalities caused by dopaminergic dysfunction, such as in PD, the presence of dopaminergic dysfunction was defined as the second obligatory criterion. Functional neuroimaging was not possible in this large cohort of healthy individuals for ethical reasons and costs; therefore, SN hyperechogenicity, as measured by TCS, was used as an indicator for dopaminergic dysfunction. In the further analysis, persons at risk for PD by these criteria are referred to as M+/SN+ (Figure 1).

The TCS examination was performed by an experienced examiner (J.G. or D.B.), according to a standardized protocol, using an ultrasound system equipped with a 2.5-MHz ultrasound transducer (Siemens Sonoline Elegra; Siemens, Erlangen, Germany).²² The examiner was masked to the results of the motor examination and laboratory assessments. The echogenic area at the anatomic site of the SN was visualized within the hypoechogenic butterfly-shaped midbrain, subsequently from both sides. The image was frozen and magnified 2- to 3-fold, and the echogenic area of the SN was encircled manually and measured planimetrically. Substantia nigra hyperechogenicity was defined as an enlarged SN area of echogenicity above the 90th percentile of the whole cohort of healthy participants on at least 1 side (0.20 cm²).²²

Blood samples were collected from all participants after overnight fasting, using serum tubes. Samples were immediately centrifuged and serum was stored at −70°C until analysis. Serum samples of IGF-1 and IGFBP-3 were assessed in the accredited reference laboratory of the Department of Pediatric Endocrinology, University of Tübingen Children's Hospital, using validated institutional assays. Serum IGF-1 level was measured using a radioimmunoassay with polyclonal anti-rabbit antibodies and recombinant human IGF-1 standards.^{23 - 24} Serum IGFBP-3 level was measured using a radioimmunoassay with polyclonal anti-rabbit antibodies and native human IGFBP-3 standards.^{23 ,25}

Statistical analysis was performed (SPSS 17.0; SPSS Inc, Chicago, Illinois) and included demographic data (age, sex, and disease duration), UPDRS-III scores, area of SN hyperechogenicity, serum IGF-1 level, serum IGFBP-3 level, and the molar ratio of IGF-1 to IGFBP-3. Descriptive statistics are given as mean (SD). After Kolmogorov-Smirnov testing for normal distribution, nonparametric testing was performed for group comparison (Mann-Whitney test) and for correlation analysis (Spearman rank correlation). Results were considered statistically significant at P < .05. For subgroup analysis, Bonferroni correction for multiple comparisons was performed. As shown in Figure 1, healthy controls were grouped according to UPDRS-III results in persons with (M+) and without (M−) impaired motor function and then further divided into those with (SN+) and without (SN−) SN hyperechogenicity as an indicator for dopaminergic dysfunction. Analysis was performed in 4 subsequent steps to address the study's goals: (1) comparison of patients with PD and healthy controls; (2) correlation of serum factors and motor function in healthy controls and comparison between M+ and M− participants; and (3) comparison of individuals at risk for PD (M+/SN+) with patients with PD and with those with impaired motor function but without signs of dopaminergic dysfunction (M+/SN−).

An overview of demographic data, UPDRS-III ratings, SN echogenicity, and serum IGF-1 and IGFBP-3 levels in the healthy control and patient groups is given in Table 1; data on the subgroups are given in Table 2.

In the cohort of healthy participants, serum IGF-1 levels showed a moderately high inverse correlation with age (ρ = −0.35, P < .001); IGF-1 levels between men and women did not differ significantly (P = .19). The IGFBP-3 levels were higher in men than in women (P = .004), but this did not correlate with age (ρ = −0.17, P = .08). Therefore, all subsequent results were corrected for age (IGF-1) and sex (IGFBP-3).

The patients with recently diagnosed PD were significantly older than the healthy controls (P = .001). However, the serum IGF-1 level was higher in the patients compared with the controls (P = .004). Serum IGF-1 levels were highest in patients with PD who had low UPDRS-III scores; IGF-1 levels and UPDRS-III scores showed a high inverse correlation in this group (ρ = −0.77, P < .001). The IGFBP-3 levels (P = .80) and the molar ratio IGF-1 to IGFBP-3 (P = .07) showed no significant differences between patients and controls.

In the healthy controls, UPDRS-III scores did not correlate with serum IGF-1 (P = .63), IGFBP-3 (P = .66), or the molar ratio IGF-1 to IGFBP-3 (P = .45). Mean IGF-1 was not significantly different between healthy controls with (M+) and without (M−) abnormalities in motor function (P = .68). The IGFBP-3 (P = .79) level and molar ratio IGF-1 to IGFBP-3 (P = .53) also did not differ significantly.

Figure 2 shows raw data of serum IGF-1 levels and UPDRS-III scores, separating the healthy controls according to dopaminergic function as indicated by the SN echogenicity status. In participants without impaired motor function (M−), no significant difference in serum IGF-1 levels between the SN+ and SN− groups could be found (P = .61). However, in participants with altered motor function (M+), serum IGF-1 levels were significantly higher in the SN hyperechogenicity (M+/SN+) group than in the SN hyperechogenicity (M+/SN−) group (P = .009). There was no statistically significant difference in serum IGF-1 levels between PD patients and M+/SN+ participants (corrected P = .15). However, there was a significant difference between PD patients and M+/SN− participants (P < .001).

In the M+/SN− group, serum IGF-1 levels correlated negatively with UPDRS-III scores (ρ = −0.44, P = .01). In contrast, in the M+/SN+ group, IGF-1 levels correlated positively with UPDRS-III scores (ρ = 0.80, P < .001).

For serum IGFBP-3 levels, no significant differences between the 4 groups could be found (Table 2). The molar IGF-1 to IGFBP-3 ratio results were comparable to the IGF-1 results (Table 2).

This study set out to further validate serum IGF-1 as a putative biological marker for idiopathic PD and to assess whether it may be an additional marker for the identification of healthy individuals putatively at increased risk for PD.

We found that serum IGF-1 levels were significantly higher in patients with untreated PD at the time of diagnosis compared with healthy controls. These findings confirm serum IGF-1 as a potential marker for idiopathic PD in early disease stages. However, one needs to be aware that the sample size was small in the PD cohort and that the cohort was highly selected regarding medication and comorbidity. Therefore, the results may be only partially applicable to the general population but are still proof of the presence of high serum IGF-1 levels in very early stages of PD.

In the SN− participants, a clear negative correlation between the degree of functional impairment, as measured by UPDRS-III, and serum IGF-1 level was observed. Furthermore, a significant negative correlation with age was noted in this group. These findings support previous data reporting a general decrease of serum IGF-1 in elderly people and most likely reflect a decrease in general trophic and regenerative capacity of the brain and body related to normal aging.^{10 - 11} This might also explain the nearly parallel decrease of IGF-1 observed in patients with PD, although the absolute levels seem to be higher overall than those in the SN− group. One might therefore assume that the decrease of serum IGF-1 observed in association with motor function decline in both PD patients and SN− controls is a physiological rather than a pathological phenomenon.

In the subgroup of SN+ individuals, we did not observe a similar decrease of serum IGF-1 in association with age or functional decline. In contrast, higher UPDRS-III scores were related to increased serum levels of IGF-1. The M+/SN+ (ie, at-risk) participants showed serum IGF-1 levels in a range that was similar to levels in patients with early PD and higher than in all other subgroups of healthy participants (M−/SN−, M−/SN+, and M+/SN−). This might lead to the assumption that M+/SN+ people share altered functioning with patients with mild PD, which might be an underlying dopaminergic deficit potentially indicating a pre-PD status.

Furthermore, the absence of UPDRS-III values higher than 7 points in the M+/SN+ group implies a potential association with early PD, as one might assume that, in this group, individuals with higher scores have already developed the full manifestation of PD, ie, criteria for PD diagnosis are met.¹⁷ Clinical follow-up of the present cohort is needed to evaluate which of the still healthy people will eventually develop PD.

The pathophysiological changes underlying the relatively increased IGF-1 levels in early PD and at-risk individuals are not yet understood. Because neither growth hormone^{12 ,15} nor IGFBP-3¹³ levels show consistent changes in PD, it seems unlikely that the increase of IGF-1 level might be caused by primary activation of the somatotropic axis. Medication and comorbidity were excluded as potential sources of the IGF-1 increase in this study. Our data imply that serum IGF-1 may be a dynamic marker that is sensitive to pathological changes emerging near the time of conversion to PD (Figure 2).

Animal studies^{26 - 27} on PD showed that IGF-1 has a profound neuroprotective capacity in the brain. Data from individuals with pathogenic mutations in the leucine-rich repeat kinase 2 gene leading to PD demonstrate that as-yet asymptomatic mutation carriers show an increased dopamine turnover in the nigrostriatal system.²⁸ Therefore, one may hypothesize that increased serum IGF-1 levels in early PD may reflect a compensatory effort to protect dopaminergic neurons from early degeneration¹² and/or may be a result of an increased dopamine turnover as compensation for already diminished dopaminergic function.

On the other hand, one may also consider that the increase of serum IGF-1 levels observed in patients with PD as well as in M+/SN+ individuals may reflect chronically increased IGF-1 levels. Because serum IGF-1 levels are partially genetically determined,^{29 - 30} one may speculate that serum IGF-1 levels in the upper normal range may constitute a susceptibility factor for later development of PD. Evidence of this hypothesis comes from animal studies³¹ demonstrating that reduced IGF-1 signaling is related to longevity and increased resistance to oxidative stress, as well as from studies^{32 - 33} reporting that patients with Alzheimer disease show increased serum IGF-1 levels. These findings suggest that an increased IGF-1 level might increase the susceptibility to degenerative processes, thus contrasting with the neuroprotective effect of IGF-1. The specific role of IGF-1 in the pathophysiologic mechanism of PD still needs to be determined.

It may be argued that all measured serum IGF-1 values in healthy persons and in patients with early PD are within age-related normal ranges. However, our results show major group differences in serum IGF-1 levels. This implies that new cutoff values for levels of serum IGF-1 might be needed for future evaluation of idiopathic PD.

Furthermore, the examined cohorts are not representative of the elderly population regarding comorbidity and medication. However, at this step, exclusion of people with such interfering factors seemed to be necessary to prove that serum IGF-1 level is a potential biomarker for PD. Furthermore, we did not evaluate the impact of nutrition and exercise levels in this cohort. However, we believe that there were no major differences between the subgroups: they were all clinically healthy, were not following a specific diet, and were not using any medication. Future studies will have to answer the question of applicability of these results in the general population. The diagnostic value of the molar ratio of IGF-1 to IGFBP-3 also needs to be evaluated. In the present study, the molar ratio showed slightly more robust results in the separation of M+/SN+ participants from the other controls than did the IGF-1 levels alone and may therefore reduce the influence of somatotropic axis activation as an interfering factor in future studies.

In conclusion, an increased serum IGF-1 level is typical in early PD and has the potential to serve as an additional biological marker for the diagnosis of this disease in the very early—perhaps even in the premotor phase—of this disorder. It might therefore be a useful addition to a battery of screening instruments that allows identification of individuals at risk for PD as a basis for early neuroprotective intervention.²