Phelan-McDermid Syndrome [PMS; under the Autism Spectrum Disorders (ASD) umbrella] is a rare genetic condition caused by deletion or de novo pathological variation of the long arm terminal end of chromosome 22 (location q13.3; also referred to as 22q13.3 deletion syndrome) within the Shank3 gene ^[1][47]. Shank3 encodes for multidomain scaffolding Shank family proteins critical for postsynaptic function of glutamatergic synapses. Shank3 deficiencies also account for 1–2% of all ASD cases, making it one of the most prevalent ASD mutations. This loss results in deficits in synaptic function, plasticity, and clinically presents with neural malformations, hypotonia, developmental delays, delayed/absent speech, autistic-like behavior, and dysmorphic facial features ^[2][48].

Clinical: Initial pilot clinical investigations into insulin and IGF-1 for the treatment of PMS utilized intraperitoneal injections of IGF-1 to treat nine children (aged 5–15) over a 3-month treatment period, with a double-blind, placebo-controlled, crossover design. Compared to placebo, children treated with IGF-1 showed significant improvement in social impairment and restrictive behaviors ^[3][49]. This early success led to investigations of intranasal (IN)-insulin on the development and behavior of PMS patients. In a randomized, double-blind, placebo-controlled clinical trial of 25 children (1–16 years) treated with daily IN-insulin sprays for a period of 6 months, IN-insulin improved developmental functioning by 0.4–1.4 months per 6 month period, with significant cognitive and social improvement in children <3 years of age, where PMS patients usually show delays in developmental growth ^[4][50].

Insulin resistance is present in 52% of BD patients and is associated with a chronic course, non-responsiveness to treatment, adverse brain changes, further cognitive impairment, and an increased mortality rate ^[5][51].

Clinical: It was hypothesized that in stable euthymic individuals with bipolar disorder I/II (BD I/II), IN-insulin would enhance hippocampal-dependent neurocognitive function. Initial double-blind, placebo-controlled crossover trials evaluating the effect of IN-insulin on adult individuals with BD I/II (n = 62) utilized an adjunctive insulin treatment of 40 IU (n = 34) or placebo (n = 28) four times daily for a period of eight weeks. Neurocognitive function and outcome were assessed with neurocognitive battery testing; significant improvements versus placebo in executive function were seen with several tests. The authors noted a need to phenotype and or/genotype subjects based upon other possible pre-existing risk factors [e.g., apolipoprotein E (ApoE)] for further possible testing ^[6][52].

SZ patients have been shown to have cognitive deficits in domains of attention, verbal memory, and executive function, with severity of impairment correlated to the extent of impact on real-life functioning. Additionally, alterations in brain insulin signaling pathways (particularly of PI3K/AKT) have been shown in SZ patients.

Clinical: Initial single-dose, double-bind, placebo-controlled trials were undertaken to evaluate the effect on verbal memory and attention deficits of IN-insulin in nondiabetic adult individuals (n = 30) with DSM-IV diagnoses of SZ or schizoaffective disorder and stabilized on another antipsychotic agent for at least one month. Participants performed cognitive tasks before and after receiving 40 IU IN-insulin or placebo. Despite decreased serum insulin levels, no significant differences in cognitive performance were observed ^[7][53]. Further repeated-dose randomized, double-blind, placebo-controlled studies (n = 45) were conducted over an 8-week period to assess the possible benefits of IN-insulin, with subjects receiving either 40 IU IN-insulin 4 times per day (n = 21) or placebo (n = 24). A battery of cognitive testing was used as a measure of psychopathology and cognition at weeks 0, 4, and 8. No beneficial effects were seen following IN treatment on psychopathology or cognition by any measures of testing ^[8][54]. Additional testing on the effects on body metabolism were done using whole body dual-energy X-ray absorptiometry and nuclear magnetic resonance (NMR) spectroscopy, for which 39 of 45 study participants were evaluated (insulin = 18, placebo = 21). Both measures failed to show a benefit of IN-insulin on any major metabolic outcomes, despite increased pro-opiomelanocortin (POMC) expression ^[9][55]. It has been concluded through these trials that although insulin signaling is impaired in the brains of SZ patients, centrally available insulin through IN delivery may not be able to overcome these impairments, leading to the observed lack of effect on cognitive, psychopathological, or whole-body metabolic measures.

Clinical: Pilot trials to assess the therapeutic potential of IN-insulin on cognitive function and mood in adults with MDD have been done. Patients (n = 35) were given either 40 IU IN-insulin or placebo 4 times a day for a period of 4 weeks, after which they were switched to the other treatments (placebo or insulin, respectively) for another 4 weeks following a 1-week washout period. Neurocognitive tests were conducted at four time points: baseline phase 1 (week 0), endpoint phase 1 (week 4), baseline phase 2 (week 5) and endpoint phase 2 (week 9) ^[10][56]. No significant changes were seen in any of the cognitive or psychiatric measures used ^[10][56].

As IN-insulin reverses anxiety-like behavior in rodents, and rats deficient in brain insulin receptors exhibit anxiety and depressive-like behavior ^[11][12][13][57,58,59], IN-insulin has been explored in a few clinical trials for its ability to modulate anxiety and fear.

Clinical: In a clinical trial to assess IN-insulin’s ability to modulate HPA axis response to psychosocial stress, 26 male participants received a single dose of 40 IU IN-insulin prior to social stress testing. Plasma cortisol, saliva cortisol, heart rate, and blood pressure were measured at resting, baseline, and in response to testing. All measured parameters increased significantly in response to stress, and IN-insulin was able to significantly diminish saliva and plasma cortisol but did not affect heart rate or blood pressure under stress conditions. The authors concluded that a single IN-dose effectively lowers the stress-induced HPA axis response ^[14][60].

Recently, IN-insulin was administered in a double-blind, placebo-controlled differential fear-conditioning paradigm in 123 healthy participants. As social anxiety is the most common type of anxiety disorder, threatening social experiences were simulated with pictures of neutral faces followed by electric shocks. This paradigm was conducted in four phases over 3 days: acquisition (day 1), extinction (day 2), reinstatement and re-extinction (day 3). Participants were given either 160 IU IN-insulin or placebo on Day 2, 45 min prior to fear extinction. Participants given IN-insulin showed a significantly greater decrease of skin conductance response in comparison to placebo, though no other parameters changed significantly (fear startle, expectancy ratings) ^[15][61].

Neurodegeneration in AD has not only been associated with an accumulation of Aβ and tau, but also with metabolic alterations in insulin signaling. AD patients have lower levels of insulin in the CSF and higher plasma insulin when compared to healthy controls, indicating peripheral insulin resistance in AD patients ^[16][62]. This brain insulin (and also IGF-1) resistance has been further demonstrated in the hippocampal formation and cerebellar cortex of AD patients, showing significantly decreased signaling in insulin receptor/IGF-1 receptor-IRS-PI3K pathways, in the absence of diabetes ^[17][63]. Impaired insulin signaling in the brain results in neuroinflammation, apoptosis, oxidative stress, and the overexpression of Aβ and tau, as reviewed by Nguyen et al. ^[18][64]. Therefore, restoring insulin levels in the CNS through non-invasive IN administration has been explored as a plausible therapeutic approach for AD cases, both in animal models and in human patients.

Preclinical: Several studies have investigated the effects of IN-insulin in WT and AD mouse models, showing dissenting results on cognition and signaling that were based on either acute or chronic insulin administration. For example, Gabbouj et al. showed that a single dose of IN-insulin in WT and APP/presenilin-1 (PS1) mice increased glucose uptake in the ventral brain and hippocampus of WT mice when compared to APP/PS1 mice, but did not improve spatial memory using the Morris Water Maze Test ^[19][65]. Nevertheless, using the same AD mouse model, another group demonstrated that 6 weeks of IN-insulin treatment improved cognitive deficits and brain insulin signaling while reducing Aβ production and plaque formation ^[20][66]. Similarly, long-term (2–6 weeks) IN-insulin administration in an AD rat model improved cognitive function and reduced tau hyperphosphorylation, inhibited microglial activation, and ameliorated neurogenesis deficits ^[21][22][23][67,68,69].

Clinical: The first randomized study of acute IN-insulin (20 IU and 40 IU) resulted in improved verbal memory in patients with AD or MCI but only in the absence of the apolipoprotein E (APOE) ε4 allele ^[24][25][70,71]. Given that ε4 negative AD patients had previously shown signs of insulin resistance ^[16][62], rethisearchers  study suggested that cognitive effects in response to intranasal insulin may indicate disrupted insulin metabolism in AD and MCI patients that do not carry the ε4 allele. Similarly, an acute, double-blinded, randomized, placebo-controlled, crossover study of a rapid-acting insulin analog (glulisine) has also shown diminished therapeutic response of IN-insulin in APOE ε4 carrier AD patients ^[26][72]. Another randomized placebo-controlled study showed that daily IN-insulin (20 IU) over 21 days improved cognition, measured by retention of new information after a delay, in patients with early stage AD ^[25][71]. However, their group reported improvement as “percentage change from placebo”, which could have affected the conclusions if there was a decrease in the placebo control group. Another group found no benefit on cognition when using Humulin-R as intranasal insulin (60 IU) administered four times daily over 48 h, although they did not stratify by APOE ε4 carriers; patients were supplemented with vitamin D2, and their dose may have been too high (IN-insulin response for memory is ∩-shaped), all of which could have potentially masked an effect ^[27][73]. Indeed, longer IN-insulin (20 IU) administration for a period of 4 months has shown improvements in delayed memory, and preservation of general cognitive abilities in patients with MCI or AD ^[28][74]. Patients in both insulin dose groups showed preserved glucose uptake in several areas of the brain, and the effects seemed to be stronger in AD compared to MCI patients ^[28][74]. In a subsequent study, this group also showed that there are important sex differences stratified by APOE ε4 carrier status. For instance, men that did not carry the APOE e4 allele showed cognitive improvement, while for women, cognitive outcome was the poorest when they were administered a high insulin dose (40 IU) ^[29][75]. The authors concluded that men were more sensitive to IN-insulin and have more pre-existing insulin abnormalities in the CNS than women ^[29][75], showing the importance of including both sexes in clinical trials.

Interestingly, IN administration of 40 IU long-acting insulin analog (detemir) improved peripheral insulin resistance during treatment, which was associated with improved verbal memory in AD patients that carry APOE ε4, unlike what had been previously reported with rapid acting insulin analogs ^[30][76]. Detemir also improved visuospatial and verbal working memory in all study participants that were administered 40 IU, but it increased peripheral insulin resistance in the APOE ε4 negative patients during treatment ^[30][76]. These differences were explained by the affinity of regular insulin to insulin receptors as well as the pharmacodynamics of the different formulations, given that long-acting insulin analogs result in greater cumulative exposure because of their increased half-life, while rapid-acting insulin mimics the post-prandial release, reaching a higher acute peak ^[30][76]. However, when comparing effects between regular insulin and detemir, it was found that detemir’s long-term efficacy decreased, while regular insulin continued to show improvements on memory in MCI and AD patients ^[31][77]. Although there were technical limitations that affected the end-point interpretations of the first phase 2/3 multisite randomized double-blind clinical trial, their results suggested an advantage in the AD Assessment Scale, improved Aβ42/40 ratios, and improved Aβ42/total tau protein in the insulin group ^[32][78]. They further assessed effects in this cohort and found that 12 months of IN-insulin reduced white matter hyperintensity volume progression, which is correlated with cognition in their studies ^[33][79].

Preclinical: IN-insulin has received recent attention in PD research due to the commonality of its dysregulation in AD, PD, and other ND or metabolic disorders, and has been explored preclinically in 6-OHDA lesioned rodent models of PD. Rats received 400 µg IN-insulin (~11.5 IU) daily for 2 weeks, beginning 24 h post-lesion, which significantly ameliorated motor impairments (improved locomotor activity) and protected dopaminergic neurons against 6-OHDA neurotoxicity without affecting body weight or blood glucose levels ^[34][80]. In another study, 6-OHDA PD rats were given a single lower dose of 3 IU IN-insulin as pretreatment either 4 days, 2 days, or 30 min before surgery. Following lesioning, rats were again given 3 IU IN-insulin 5 days/week for 4 weeks; even with this significantly lower dose from the previous study (~25%), IN-insulin still rescued 6-OHDA motor deficits in motor behavioral battery testing and protected against dopaminergic neuron loss ^[35][81]. A subsequent study, using an even lower dose (2 IU) treatment for 6 weeks, found that 6-OHDA affected not only dopaminergic neurons in SN, but also insulin signaling in the ipsilateral hippocampus through pathway disruption from SN, and IN-insulin was able to restore the function of this injured insulin signaling pathway. IN-insulin was again shown to improve motor behavior, and at this very low dose also protected SN dopaminergic neurons and restored levels of pAKT and pGSK3β in the ipsilateral hippocampus affected by 6-OHDA ^[36][82].

Clinical: Initial clinical trials for IN-insulin in treatment of PD and parkinsonian-type multiple system atrophy have begun in recent years. In a double blind, placebo-controlled pilot study, eight PD patients (n = 8, 2 women) and six healthy age-matched participants self-treated with 40 IU IN-insulin daily for 4 weeks; IN-insulin treated PD participants had a significant word count increase, decreased severity of parkinsonism and improved motor score. IN-insulin did not affect any of the other parameters measured for cognition, gait, or depression ^[37][83].

Though conflicting results make it difficult to understand the exact role of the IGF-1/AKT pathway in Huntington’s Disease (HD), high plasma levels of IGF-1 have been correlated to cognitive decline in HD patients, and preclinical data has shown the protective benefits of IGF-1 in striatal neuronal culture and in R6/1 mice.

Preclinical: A pilot IN-IGF-1 study has been carried out using a YAC128 mouse model of PD, expressing human mutant Huntingtin (mHtt) with ~128 CAG repeats. Six-month old male YAC128 and wild-type mice were given 35 µg IGF-1 with an alternate-day IN dosing paradigm for two weeks. In motor testing, mice treated with IN-IGF-1 demonstrated reduced motor impairment and improved locomotor activity (rotarod and open field), significantly increased cortical and striatal PI3K/AKT/mammalian targets of rapamycin (mTOR) signaling, reduced phosphorylation of mHtt, reduced markers of CNS metabolic abnormalities, and increased cortical (but not plasma) levels of IGF-1 ^[38][84].