Activated PI3Kδ syndrome, an immunodeficiency disorder, leads to sensorimotor deficits recapitulated in a murine model

Abstract

The phosphoinositide-3-kinase (PI3K) family plays a major role in cell signaling and is predominant in leukocytes. Gain-of-function (GOF) mutations in the PIK3CD gene lead to the development of activated PI3Kδ syndrome (APDS), a rare primary immunodeficiency disorder. A subset of APDS patients also displays neurodevelopmental delay symptoms, suggesting a potential role of PIK3CD in cognitive and behavioural function. However, the extent and nature of the neurodevelopmental deficits has not been previously quantified. Here, we assessed the cognitive functions of two APDS patients, and investigated the causal role of the PIK3CD GOF mutation in neurological deficits using a murine model of this disease. We used p110δ^E1020K knock-in mice, harbouring the most common APDS mutation in patients. We found that APDS patients present with visuomotor deficits, exacerbated by autism spectrum disorder comorbidity, whereas p110δ^E1020K mice exhibited impairments in motor behaviour, learning and repetitive behaviour patterning. Our data indicate that PIK3CD GOF mutations increase the risk for neurodevelopmental deficits, supporting previous findings on the interplay between the nervous and the immune system. Further, our results validate the knock-in mouse model, and offer an objective assessment tool for patients that could be incorporated in diagnosis and in the evaluation of treatments.

Keywords: APDS; ASD; Mouse; PID; PIK3CD; Primary immunodeficiency.

Fig. 1

Visuomotor assessment tasks. Visual representation of the saccade (a), tapping (b) and trajectory prediction (c) tasks performed by patients P1 and P2, and respective age-matched controls. Pro-tasks involved the execution of reflexive saccades (and tapping) towards a newly appeared target while anti-tasks required a saccade execution (and tapping) to the opposite side of the new target. In memory tasks, subjects waited for target omission to perform a saccade (and tapping). Trajectory prediction tasks involved the execution of a saccade and tapping towards the basket in which a moving ball would be expected to fall. The number and age distribution of control participants per task can be found in Table 1.

Fig. 2

APDS patients present with intact reflexive saccades but altered integration. a) Performance in the saccade and tapping tasks is presented as percentage of correct trials. Eye latency for the saccade tasks (b), and eye and hand latency for the tapping tasks (c–e) are presented in ms. f) Representative traces of the eye trajectories performed towards one basket, during the trajectory prediction task. P1, patient 1, P2, patient 2; C1, age-matched controls for patient 1, C2, age-matched controls for patient 2.

Fig. 3

Slight fine locomotion impairments are caused by the murine E1020K mutation. a) Example of automatically tracked trajectories showing the body position of a WT and a p110δ^E1020K mouse during the 30 min of the OF task. b-c) Quantification of the mean speed (n ​= ​15 per genotype) and total distance travelled (n ​= ​15 WT and 14 p110δ^E1020K) during the OF task, presented as median and quartiles (2-tailed t-test). d) The total time each mouse spent on the rotarod, over the course of 4 trials/day, is presented as median and quartiles (2-way repeated-measures ANOVA, n ​= ​15 per genotype). On the last day, the maximum rod speed was increased to 80 ​rpm. e-g) The Erasmus ladder was used to investigate locomotion pattern. The average time each mouse spent on the ladder, across 42 daily trials, is presented in (e) (2-tailed Mann-Whitney; data presented as mean ​± ​SEM). The distinct step types analysed are schematically represented in (f) and quantified in (g) (Mixed effects model; data is presented as daily mean percentage, n ​= ​15 per genotype). ∗p ​≤ ​0.05, ∗∗p ​≤ ​0.01, ∗∗∗p ​≤ ​0.001.

Fig. 4

p110δ^E1020Kmice exhibit subtle changes in burying and grooming patterns. a) Representation of the OF arena parcellation into corner, outside and inside areas. b) Total time spent on each OF area (2-tailed Mann-Whitney; data presented as mean ​± ​SD). c-d) Total time spent and number of entries performed in each EPM area (2-way repeated-measures ANOVA; data are presented as median and quartiles). e) Total marble area buried during the MB task (2-tailed t-test). f) Marble disposition before the task, superimposed with the average percentage of buried area per marble (n ​= ​15 WT and n ​= ​13 p110δ^E1020K). g-i) Quantification of the total time spent grooming (g), total number of grooming bouts (h) and the average time interval between grooming bouts (i), during the grooming assay (2-tailed t-test; data presented as median and quartiles). j) Representative plot depicting grooming and rearing events for one mouse of each genotype. k) Time-binned plot with the average time spent grooming (dashed line; 2-way repeated-measures ANOVA) and the average number of grooming bouts (full line; mixed effects model) (data are presented as mean ​± ​SEM). G, grooming, R, rearing; ∗p ​≤ ​0.05, ∗∗p ​≤ ​0.01, ∗∗∗p ​≤ ​0.001, n ​= ​15 mice per genotype, expect for e) and f) (see above).

Fig. 5

Stimulus-dependent ladder exit and reversal learning are mildly affected in p110δ^E1020Kmice. a) Schematic of the Erasmus ladder goal-box with the time intervals between stimuli. b-d) Number of times individual mice left the goal-box with the air stimulus (b), the light stimulus (c), or before light cue presentation (d). e) ratio between light and air exits. f) Best-fit regression model between the data points used to plot (d) and (e). g) Picture of the Y-Maze, with dashed squares representing the possible locations for the hidden platform, either on the right or left arm of the apparatus. h) Total number of correct arm choices for both genotypes, during each phase of the Y-Maze (data presented as median with interquartile range). i) Step function with the cumulative median and interquartile range for the number of correct arm choices during all acquisition and test trials. j) Percentage of correct arm choices for each genotype over the four days of acquisition and the day of test (data presented as median with interquartile range) k-l) Similar to (i) and (j) but for the reversal I phase. 2-tailed Mann-Whitney, except for f). ∗p ​≤ ​0.05, n ​= ​15 mice per genotype, except for reversal phases where n ​= ​13.

Fig. 6

Social behaviour is largely preserved in p110δ^E1020Kmice. a) Example of automatically tracked body positions during test phase (novel mouse on the left). b) Total time individual mice spent in each chamber of the apparatus during test phase (mixed effects model; data presented as median with interquartile range). c) Before and after plot of the total time each individual mouse spent on the novel mouse chamber during baseline (BL) and test (T) (2-way repeated-measures ANOVA). d) Body position heatmap depicting position frequency per 2.5 ​mm bins (novel mouse on the left). e) Median and quartiles with the ratio between the time each individual mouse spent near the social cup over the time it spent near the empty cup (2-tailed t-test). f) Representative plot with the duration, in seconds, of each transition into the empty (light bars) or novel mouse (dark bars) cup area. g) Median and quartiles with the percentage of transitions each individual mouse made to the novel mouse or empty cup (2-way repeated-measures ANOVA); h) Median and quartiles of the ratio between the time individual mice spent exploring the novel mouse cup over the time spent in the whole novel mouse chamber (2-tailed t-test). ∗p ​≤ ​0.05, ∗∗p ​≤ ​0.01, ∗∗∗p ​≤ ​0.001. n ​= ​15 mice per genotype.

Fig. 7

LD1 and LD2 features separate WT from p110δ^E1020Kmice. a) Absolute contribution of each behavioural variable, in percentage, to linear discriminants 1 and 2, resulting from the LDA. b) LDA plot featuring the 10 best contributors to LD1 and LD2. Negative and positive contributions are represented by black and white bars, respectively. Each dot represents one mouse, with pink dots representing WT mice and green dots p110δ^E1020K mice. n ​= ​15 mice per genotype. EL, Erasmus ladder, EPM, elevated-plus maze, G, grooming, M, marble burying, OF, open-field, R, rotarod, SI, social interaction, YM, water y-maze. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)