Brain functions and cognition on transient insulin deprivation in type 1 diabetes

Abstract

BACKGROUNDType 1 diabetes (T1D) is a risk factor for dementia and structural brain changes. It remains to be determined whether transient insulin deprivation that frequently occurs in insulin-treated individuals with T1D alters brain function.METHODSWe therefore performed functional and structural magnetic resonance imaging, magnetic resonance spectroscopy, and neuropsychological testing at baseline and following 5.4 ± 0.6 hours of insulin deprivation in 14 individuals with T1D and compared results with those from 14 age-, sex-, and BMI-matched nondiabetic (ND) participants with no interventions.RESULTSInsulin deprivation in T1D increased blood glucose, and β-hydroxybutyrate, while reducing bicarbonate levels. Participants with T1D showed lower baseline brain N-acetyl aspartate and myo-inositol levels but higher cortical fractional anisotropy, suggesting unhealthy neurons and brain microstructure. Although cognitive functions did not differ between participants with T1D and ND participants at baseline, significant changes in fine motor speed as well as attention and short-term memory occurred following insulin deprivation in participants with T1D. Insulin deprivation also reduced brain adenosine triphosphate levels and altered the phosphocreatine/adenosine triphosphate ratio. Baseline differences in functional connectivity in brain regions between participants with T1D and ND participants were noted, and on insulin deprivation further alterations in functional connectivity between regions, especially cortical and hippocampus-caudate regions, were observed. These alterations in functional connectivity correlated to brain metabolites and to changes in cognition.CONCLUSIONTransient insulin deprivation therefore caused alterations in executive aspects of cognitive function concurrent with functional connectivity between memory regions and the sensory cortex. These findings have important clinical implications, as many patients with T1D inadvertently have periods of transient insulin deprivation.TRIAL REGISTRATIONClinicalTrials.gov NCT03392441.FUNDINGClinical and Translational Science Award (UL1 TR002377) from the National Center for Advancing Translational Science; NIH grants (R21 AG60139 and R01 AG62859); the Mayo Foundation.

Keywords: Diabetes; Endocrinology; Insulin; Memory.

Conflict of interest statement

Conflict of interest: The authors have declared that no conflict of interest exists.

Figures

Figure 1. Study design.

Both study groups (nondiabetic [ND] and T1D) started the study with a fasting blood sampling. Participants with T1D were then started on a continuous insulin infusion titrated to maintain blood glucose between 5 and 6.7 mmol/l (Time 1). Following a 3-hour insulin infusion in the T1D group or the corresponding time course for the ND group, participants had cognitive testing. Following this, the participants were taken to radiology, and MR studies were performed. The insulin infusion was continued during both cognitive testing and MR studies. Then, blood sampling was performed, and immediately after the insulin infusion was discontinued in the participants with T1D. Four hours following after a period of insulin withdrawal (Time 2), or at the corresponding time course for the control participants, the cognitive testing (4–4.5 hours) and then MR studies were performed (5 ± 0.6 hours). Time 1 represents the insulin-treated period in T1D, and Time 2 represents the period following insulin deprivation in T1 D. Both times represent no intervention but specifically address time-related changes in the ND group.

Figure 2. Cognitive test results showing changes in participants with type 1 diabetes from treatment with insulin (Time 1) to the insulin-deprived state (time 2) in comparison with the same period in participants without diabetes.

(A) Compared with participants without diabetes (ND), participants with type 1 diabetes (T1D) showed significantly poorer attention and working memory based on the Wide Range Assessment of Memory and Learning-2 (WRAML-2) Number Letter subtest (*P < 0.001) and fine motor speed based on Delis-Kaplan Executive Function System (D-KEFS) Trail Making (*P < 0.02). (B) Baseline differences in myo-inositol (*P < 0.05), N-acetyl aspartate (NAA) (*P < 0.05), and cortical fractional anisotropy (FA) (*P < 0.05) between the T1D group and ND controls; none of them were significantly altered by insulin deprivation (data not shown). (C) Total adenosine triphosphate (ATP) levels that significantly decreased on insulin deprivation (Time 1 to Time 2) in participants with T1D (P < 0.04) and phosphocreatine (PCr) showed no significant changes. The ratio of PCr to ATP increased in the T1D group (P < 0.03).

Figure 3. Seed-based functional connectivity maps.

Seed masks were extracted from the FreeSurfer parcellation of the N27 brain template. Three seed regions reached statistical significance (P < 0.01) (left and right hippocampi and bilateral posterior cingulate cortex). The functional connectivity (FC) of these seed regions was higher (yellow) or lower (blue) compared with 6 different brain regions (Table 2). At baseline (left column), there was appreciably higher FC in nondiabetic (ND) participants relative to that in participants with type 1 diabetes (T1D) (A) between the left hippocampus and the right early visual area and (D) between the right hippocampus and bilateral early visual areas. (E) In contrast, there was significantly (P < 0.01) lower FC between the right hippocampus and right putamen in ND participants. There were no significant FC differences with the bilateral posterior cingulate cortex in the baseline state. Following insulin deprivation (middle column), there was substantially higher FC in ND participants relative to that in participants with T1D (B) between the left hippocampus and bilateral sensorimotor cortices but (G) lower FC between both posterior cingulate cortices and the right precentral gyrus. There were no significant FC differences with the right hippocampus in the postwithdrawal state. Finally, assessing changes between the postwithdrawal and baseline states in participants with T1D (right column), (C) there was considerably decreasing FC between the left hippocampus and bilateral sensorimotor cortices and (F) the right hippocampus and the left caudate/putamen following insulin withdrawal, indicating that insulin deprivation adversely affected FC between these regions in participants with T1D. No significant FC differences with the bilateral posterior cingulate cortex were found in participants with T1D between the baseline and postwithdrawal states. Note that the FC clusters for the bilateral sensorimotor cortex identified in B and C are similar but not identical (see Table 2).