Finding successful therapies for the treatment of Alzheimer’s disease (AD) is one of the most challenging tasks existing for human health. Several drugs have been found and validated in preclinical studies with some success, but not with the desired breakthroughs in the following clinical development phases. AD causes multiple brain dysfunctions that can be described as a brain organ failure, resulting in significant cognitive decline. Aggregation of amyloid proteins and neuronal loss are the hallmarks of AD. Thus, one of the strategies to treat AD is to find a multifunctional drug that may combine both anti-aggregation and neuroprotective properties. Such a candidate could be chemically modified dendrimers. Dendrimers are branched, nonlinear molecules with multiple reactive groups located on their surface. Chemical modification of reactive surface groups defines the property of the dendrimers. In this chapter, I will discuss poly(propylene imine) dendrimers with the surface functionalized with histidine and maltose as an example of a multifunctional therapeutic drug candidate able to protect the memory of AD transgenic model mice.
- Alzheimer’s disease
Alzheimer’s disease (AD) is a complex neurological disease, which already in its earliest clinical phase is characterized by remarkable memory impairment. Multiple pieces of evidence suggest that in AD, memory impairment begins with dysfunction of synapses, a unique characteristic of nerve cells. Early neurochemical analyses of AD brain tissue revealed that the deficits in numerous neurotransmitters (including corticotropin-releasing factor, somatostatin, GABA, and serotonin) and the early symptoms correlate with dysfunction of cholinergic and glutamatergic synapses . In addition to the deficits of the transmitters, many other biochemical and morphological indicators suggest that in early AD, synapses are under attack as reviewed in . It has been shown that in biopsied AD cortex, there is a significant decrease in the numerical density of synapses in the brain and the number of synapses per cortical neuron . The amyloid cascade hypothesis, one of the widely accepted theories, suggests that progressive accumulation and aggregation of amyloid-β proteins (Aβ) could be the main cause of AD, which triggers AD neuropathology. Aβ proteins are the proteolytic products of amyloid precursor protein (APP), a type-I transmembrane protein which is highly expressed in neurons, known to regulate synaptic function and neurite outgrowth . There are two main alternative enzymatic pathways to process APP :
Non-amyloidogenic pathway, where APP is subjected to consecutive cleavage by α-and γ-secretases that cut APP within the Aβ fragment
Amyloidogenic APP pathway, where APP is subjected to cleavage by β-and γ-secretases generating Aβ, a mix of short peptides ranging from 38 to 43 amino acids in length able to form polymorphous aggregates, so-called oligomers, and fibrils 
APP processing is regulated by neuronal activity, and neuronal activity may favor β-secretase-mediated amyloidogenic cleavage of APP during which Aβ proteins are generated . It was accepted that after APP cleavage, Aβ peptides are first secreted, and then, extracellularly, soluble Aβ peptides aggregate into amyloid plaques. This extracellular Aβ, which is the main constituent of amyloid plaques, is thought to be toxic to the neurons. More recently, the intraneuronal Aβ has been demonstrated and reported to be involved in neuronal damage [8, 9]. It has been demonstrated that Aβ attacks synapses, small membranous protrusions that permit one neuron to pass a signal (electrical or chemical) to another neuron.
It has been shown that synaptic activity may affect Aβ secretion , and it has been hypothesized that synaptic activity may stimulate the generation of Aβ although why this occurs and whether Aβ might have a normal function in neuronal synapse have not been understood well . Strikingly, it has been shown that Aβ selectively binds to synapses when added to cultured neurons . Further, the level of Aβ is shown to be increased in synaptosomes in early AD . Immunoelectron microscopy and high-resolution immunofluorescence microscopy studies show that this early subcellular Aβ accumulation leads to progressive damage of neurites and synapses . Thus, synapses could be sites of early accumulation of pathogenic Aβ. It is believed that soluble Aβ oligomers rather than monomeric or fibrillar Aβ are the main neurotoxic species. However, a structure of neurotoxic Aβ oligomers and the nature of their effects on synapses are not identified .
Despite advances, the efforts to target neurotoxic Aβ oligomers in the brain are confounded by high polymorphism of amyloid structures . Oligomer specific antibodies may interact mainly with a specific type of Aβ conformers against which these antibodies were produced . Therefore, to target polymorphic Aβ oligomers, a cocktail from several antibodies might be required. Another way to modulate Aβ aggregation could be via establishing H-bond interactions  to favor the formation of less toxic Aβ species .
To fight a brain disease such as AD pathology, both synapse protection and anti-amyloid modulation would be desired properties of a possible therapeutic drug. However, to protect synapses and to modulate Aβ aggregation, amyloid aggregation modulator and neuroprotective therapeutics have to be delivered to the synapse. One way to deliver both therapeutic molecules is to use a compound which may carry both molecules simultaneously. Such multifunctional compound could be a dendrimer.
Dendrimers are three-dimensionally branched, globular macromolecules built by a series of iterative steps from a small core molecule which defines the type of the dendrimer . They were first synthesized and described in 1978 , and since then dendrimers are in focus, due to their outstanding complexation properties. The most important features of dendrimers are controlled molecular structure, nanoscopic size, and high tunable availability of multiple functional groups at the dendrimer surface. Dendrimers are composed of three elements: a core branched dendron and terminal groups which could be used for dendrimer functionalization. The number of surface functional groups of the dendrimer depends on the degree of dendrimer branching (Figure 1). For example, PPI or PAMAM dendrimers of the second generation have 16 functional groups on their surface, the third generation has 32, and the fourth dendrimer generation has 64 functional groups. Strikingly, the number of terminal groups increases exponentially, while the size increases linearly. The terminal groups on the dendrimer surface can be used for surface modification and dendrimer functionalization. Such modifications could change dendrimers’ surface charge and, for example, reduce toxicity associated with a cationic surface charge as reviewed by Appelhans et al. . Dendrimers are most commonly synthesized using divergent or convergent different synthetic pathways . Importantly, the high tunability of dendrimers’ surface allows endless possibilities for dendrimers’ biomedical applications, for example, for pharmaceutical applications, the terminal groups can be functionalized with different active conjugates such as specifically targeting antibodies, drugs, metal ions or imaging agents, and more . Moreover, several research groups demonstrated that some types of dendrimers are able to cross the BBB [22, 26, 27, 28], showing their applicability for the research and possibly treatment of brain diseases.
In the present chapter, I summarize the experimental evidence showing that functionalized poly(propylene imine) dendrimers may provide multitargeting properties for dendrimers increasing their potential for the treatment of AD.
2. Amyloid aggregation and dendrimers
According to the amyloid cascade hypothesis, Aβ peptides are important players triggering the AD development. Multiple in vitro studies have demonstrated that the Aβ peptides can form fibrils and other aggregates called oligomers. The formation of insoluble Aβ fibril follows a nucleation-dependent polymerization mechanism (Figure 2) as described . The formation of soluble Aβ oligomers in vivo is largely unknown; it is believed that soluble Aβ oligomers may precede fibril formation  and are more toxic than mature Aβ fibrils .
In the search for drugs that would inhibit neuronal death in Alzheimer’s disease, one of the ways one can use is to find compounds that interfere with Aβ, cleaning the brain tissues from neurotoxic Aβ oligomers. It has been demonstrated that PPI dendrimers modified with maltose are capable of interfering with the amyloid formation in vitro [18, 28, 32, 33]. Amyloid fibril formation is usually monitored using the fluorescence dye thioflavin T (ThT). The dye becomes fluorescent when interacting with the ordered β-sheet structures characteristic for amyloid fibrils. With the fibril growth, the ThT fluorescence increases until its value reaches a plateau. Figure 3 demonstrates the sigmoid-shaped line corresponding to the ThT kinetics corresponding to the fibril growth of Aβ(1–40), where the lag (nucleation) phase is followed by the elongation phase and plateau; when all ThT molecules have intercalated into β-sheets of the amyloid fibrils, the aggregation kinetics of amyloids is reviewed .
PPI dendrimers modified with maltose may, in the case of Aβ(1–40) or Aβ(1–42), interfere with amyloid fibril formation in a concentration-dependent manner, indicating that maltose PPI dendrimers bind amyloid proteins . Figure 4 demonstrates the ThT fluorescent kinetics of Aβ(1–40) and Aβ(1–42) in the presence of maltose PPI dendrimers. As expected, Aβ alone forms the typical amyloid fibrils . However, when the maltose PPI dendrimers are present, the morphology of amyloid fibrils is altered, demonstrating binding of the dendrimers to Aβ [18, 28, 35, 36]. The electron micrograph shows the morphology of amyloid fibril in the presence of maltose PPI dendrimers. Fibril clumps were generated by incubating maltose PPI dendrimers with Aβ(1–40). As it has been suggested that dendrimers interact with Aβ thus, fibrils seem to be varnished by maltose dendrimers and clumped together, and importantly, no Aβ oligomers were observed in the presence of maltose PPI dendrimers . Thus it is reasonable to think that maltose dendrimers interacting with Aβ may form hybrid fibrils, shifting the balance between oligomeric and fibrillar forms of Aβ toward less toxic hybrid products.
Dendrimers’ intrinsic toxicity is an important issue in relation to their potential biological applications . It was observed that unmodified PPI dendrimers have high intrinsic toxicity for cells [38, 39]. It was hypothesized that this toxicity could be related to the dendrimer capacity of establishing strong interactions of electrostatic nature . It has been demonstrated that dendrimers with a surface decorated by polysaccharides, such as maltose or maltotriose, confer less toxicity [41, 42]. The charge of the dendrimer covered by polysaccharides is close to neutral; thus the interaction of dendrimer with other biomolecules is driven by hydrogen bonds, which is less strong; therefore, dendrimers covered by polysaccharides are less toxic [38, 39, 41].
In collaborations between the research groups of Dietmar Appelhans (Leibniz Institute of Polymer Research, Dresden, Germany), Josep Cladera (Autonomous University of Barcelona, Spain), and Isidro Ferrer in Barcelona (University of Barcelona, Spain), it has been shown that distinct PPI dendrimers with electroneutral maltose shell, with cationic maltose or maltotriose shell, were tested against amyloid toxicity in vivo and in vitro. The evaluation of the toxicity of Aβ in the presence of PPI maltose dendrimers showed that the dendrimers could significantly reduce the Aβ toxicity compared to Aβ alone .
Interestingly, only the electroneutral maltose dendrimers were able to reduce the toxicity of Alzheimer’s disease brain extracts in cultured SH-SY5Y neuroblastoma cells . Moreover, maltose PPI dendrimers with electroneutral or cationic surface penetrated the cytoplasm of cultured cells. Additionally, they penetrated inside the brain when administered to AD transgenic mice intranasally . These PPI maltose dendrimers were able to modify amyloid plaque load in the brains of AD transgenic animals, showing anti-amyloid potential for in vivo applications. However, the studied maltose PPI dendrimers could not reverse memory impairment in APP/PS1 mice following chronic administration. Strikingly, cationic maltose dendrimers were neurotoxic in vivo and caused cognitive decline in non-transgenic mice . Taken together, these results suggest that maltose PPI dendrimers require further optimization of biocompatibility.
3. Modified PPI dendrimers as potential multifunctional therapeutics for Alzheimer’s disease
As it has been mentioned at the beginning of the chapter, Alzheimer’s disease is a fatal neurodegenerative disorder. AD is characterized by a decade-long presymptomatic phase, and it is during the presymptomatic phase, before synaptic damage and neuronal loss, that therapies are most likely to be effective . Thus, a preventive treatment which could protect synapses and reduce the neurotoxicity of Aβ oligomers is one such strategy. Such successful drug candidates for AD treatment have to possess both anti-amyloidogenic and neuroprotective properties. Therefore, a modification of maltose dendrimers with a molecule with neuroprotective characteristics was the next logical step in search of the new drug candidate for the treatment of AD.
To further improve the pharmacological properties of maltose PPI dendrimers, it was decided to modify PPI dendrimers of the fourth generation with maltose and histidine. Maltose was used due to anti-amyloidogenic properties; histidine was added due to several reasons: it is selectively transported through the BBB . Histidine has chelating properties for Cu2+ ions . Thus these properties were considered to be important since Cu ion dyshomeostasis may play a detrimental role in AD progression , and importantly, histidine has been shown to have some neuroprotective capacity . After the modification, G4 PPI dendrimers modified with maltose and histidine were supposed to possess both anti-amyloid and neuroprotective properties simultaneously.
In vitro evaluations demonstrated that histidine-maltose PPI dendrimers could interact with Aβ. As maltose PPI dendrimers, G4 histidine-maltose PPI dendrimers did not prevent fibril formation but clump Aβ fibrils (Figure 5). Importantly, small oligomeric aggregates were not present in the studied suspensions in the presence of the dendrimers. Interestingly, the intensity of ThT was significantly decreased following the aggregation of Aβ probably due to the competition of the dendrimers with ThT for binding to Aβ(1–40) or due to change of structure, resulting in lower ThT fluorescence quantum yield [48, 49]. To test if G4 histidine-maltose PPI dendrimers could reduce the neurotoxicity of Aβ, primary neurons derived from wild-type mouse were treated with 1 μM Aβ(1–42) in the presence of the dendrimers at the ratio 1 to 1. As it was demonstrated by cell viability assay, histidine-maltose PPI dendrimers significantly reduced the neurotoxicity of soluble Aβ oligomers . Figure 6 shows the neuronal viability in the presence of the dendrimers and Aβ(1–42) oligomers as assessed by a lactate dehydrogenase (LDH) activity assay. 1 μM G4 histidine-maltose PPI dendrimers were added to primary neurons and incubated 24 h before the assay; as it was documented, the dendrimers alone were not toxic to the neurons. 1 μM recombinant Aβ(1–42) monomers, oligomers, and fibrils were added to primary neurons and incubated 1 h at 37°C in the presence and the absence of dendrimers. The results demonstrate that G4 histidine-maltose PPI dendrimers significantly reduced the toxicity of Aβ(1–42) for primary neurons.
In vivo evaluations demonstrated that chronic treatment with histidine-maltose PPI dendrimers of APP/PS1 mice prevented AD-related memory impairment . Figure 7 shows the results of the memory test after the treatment. APP/PS1 mice harbor two human genes: APP with the KM670/671NL, the Swedish mutation, and PSEN1 with the L166P mutation . In APP/PS1 mice, human Aβ increases with age, but Aβ42 is preferentially generated over Aβ40, and the expression of the human APP transgene is approximately 3-fold higher than the endogenous murine APP . For the treatment, APP/PS1 and wild-type mice were randomly divided into four groups, two groups (transgenic and wild type) were treated intranasally with histidine-maltose PPI dendrimers, and two groups (transgenic and wild type) were given intranasally phosphate saline. Administration lasted 3 months until animals reached the age of 6 months, the age when the first cognitive decline is detected . Memory evaluation tests were performed at the end of treatment using two object recognition tests in a VmazeR as described .
To understand a possible mechanism behind the memory rescue, the levels of pre- and postsynaptic markers in the brain of treated APP/PS1 mice were evaluated by Western blotting. Pre- and postsynaptic markers, such as drebrin and synaptophysin, play a crucial role in the synaptic plasticity and are downregulated in AD [53, 54]. Loss of synaptophysin correlates with cognitive impairments in AD patients and AD transgenic models [54, 55]; Psd95 knockout animals have impaired basal synaptic transmission and learning deficit ; transgenic animals lacking synaptophysin have reduced novel object recognition . Importantly, it has been shown that loss of synaptophysin immunoreactivity precedes amyloid plaque formation [58, 59]. Preventive treatment of AD transgenic mice with G4 histidine-maltose PPI dendrimers prevented a decrease in synaptic proteins compared to PBS-treated mice .
In contrast, G4 histidine-maltose PPI dendrimers did not change the level of these synaptic proteins in WT mice, indicating that, most likely, the level of their mRNA expression was not affected . Thus it is reasonable to think that the increased levels of pre- and postsynaptic proteins are more likely an effect of reduced synaptic loss in the treated AD transgenic animals (Figure 8). Thus a possible mechanism of memory protection in APP/PS1 could be the synapses were shielded by the dendrimers from toxic Aβ oligomers or the toxicity of Aβ oligomers were inactivated in the presence of the dendrimers.
4. Conclusions and perspectives
Dendrimers, which represent a type of 3D polymers, have been in the spotlight for three decades in biomedical and pharmaceutical research, and their chemistry and synthesis are continuously progressing by efforts from many research groups and companies. Although there are still many unclear problems in AD, in this chapter, functionalization of dendrimers dedicated to the prevention of memory decline in AD pathogenesis has been discussed. Based on the reviewed literature, PPI dendrimers have been shown to be useful in the way of the surface functionalization, which tuned their biochemical properties. Strikingly, the effect of the surface functionalization with histidine and maltose magnified exponentially neuroprotective properties of PPI dendrimers, resulting in an unprecedented outcome, such as memory protection in AD transgenic animals.
In this chapter, I have analyzed the functionalization of PPI dendrimers, which tuned the intrinsic properties of PPI dendrimers and converted them into a multifunctional drug candidate against Alzheimer’s disease. Modification of the dendrimer surface with maltose allowed dendrimers successfully to interfere with Aβ(1–42) by forming nontoxic hybrid glycofibrils. Modification of the dendrimer surface with histidine improved the ability of the dendrimers to cross the blood–brain barrier and resulted in synaptic protection. By reducing the level of soluble amyloid oligomers, on the one hand, and conferring synapse protection, on the other hand, the dendrimers were given multifunctionality against main features of AD, synaptic loss, and aggregation of Aβ. These observations, coming out of the studies on the interaction of dendrimers with amyloid peptides [18, 22, 28, 32, 42], carried out in vitro and in vivo, point toward a possible use of dendrimers (in particular functionalization of PPI dendrimers with histidine and maltose) as a multifunctional drug candidate against Alzheimer’s disease.
However, to find a successful drug against AD, other modifications of histidine-maltose PPI dendrimers might be required. For example, the ability to cross the blood-brain barrier, cell wall penetration, distribution in the specific tissue, and biodegradation could be tuned for a particular dendrimer application.
I acknowledge Dr. Dietmar Appelhans (Leibniz Institute of Polymer Research Dresden, Dresden, Germany) for the generation of histidine-maltose PPI dendrimers, collaboration, and valuable discussions.
The work is supported by MultiPark (Lund University), Vinnova, Swedish Research Council grants.
I thank Dr. Stefan Broselid for fruitful discussions and editorial help.