Therapeutics such as enzymes, antibodies, and even transport proteins (e.g., albumin), which are mostly intended to link with molecular targets to be removed from the organism, do not really need to be released in the fluid or tissue to action. In fact, binding to the molecular target can be achieved regardless of the compartment. With this in mind, IT pseudodelivery of drugs is a novel concept to administer drugs to treat CNS conditions relying on the CSF-sink therapeutic strategy [60], by means of implantable DDS to put in touch therapeutics with molecular targets inside of the device, without delivering to the biological fluid (hence the name “pseudo”-delivery). 

Functional nanoporous materials are an important class of nanostructured materials because of their tunable porosity and pore geometry (size, shape, and distribution) and their unique chemical and physical properties. Progress in developing a broad spectrum of nanoporous materials has accelerated their use for extensive applications in biomedical fields [68]. Nanoporous membranes are natural or synthetic membranes that can be made from a variety of materials and can be fabricated in different configurations including pore size, surface coating, geometry, and pore distribution, providing unique mass transport characteristics that have numerous potential biological and medical applications that involve isolating, sorting, sensing, and releasing biological molecules. Nanoporous membranes are of great interest in drug delivery because they offer a secure delivery system for medications and stop bodily enzymes from breaking them down and because they can be tailored-made and fine-tuned for precise control of the rate of drug delivery or to exquisitely adjust the selective molecular permeability [69,70]. While a few years ago there were technical challenges for the successful application of nanoporous membranes to controlled drug delivery applications—including the need for biocompatibility, the reduction of risk of infections, and the reduction of risk of biofouling [71] most of these challenges have already been overcome and solutions are now being optimized [72,73]. Nanoporous membranes can be used as stand-alone DDS or assembled into complex DDS.

IT pseudodelivery is the first DDS to be endowed with nanoporous membranes acting on the CNS [72,74,75]. Devices for IT pseudodelivery of drugs look similar to intrathecal pumps as they also have a subcutaneous reservoir and an intrathecal catheter accessing the CSF. However, they are not necessarily endowed with electromechanical pumps. The mechanism of action depends on the use of nanoporous membranes enabling selective molecular permeability [72,75]. On one side, the membranes do not allow crossing of drugs, but on the other side, they allow crossing of the target molecules present in the CSF. Target molecules bind drugs inside the system, thus being trapped or cleaved and eliminated from the CNS (a short simulation illustrating the mechanism of action can be found as a Supplementary Materials). Drugs are not released from the reservoir to the organism, and they can be replaced as needed percutaneously through self-sealing septa in the reservoir.

Not every target molecule or drug is suitable to be targeted/used via pseudodelivery. For a disease to be suitable to be treated using IT pseudodelivery, three conditions must be met:

• A target molecule should be present in the CSF (soluble). This should be identified as potentially “toxic” or “pathogenic” and involved directly (aggregating proteins) or indirectly (mediators) in the physiopathology of the disease.
• A drug acting specifically on the target molecule is needed. This can be an antibody, an aptamer, an enzyme, or any other compound that has specificity over the target molecule and either binds or cleaves the target molecule.
• A significant size difference should exist between the target and drug molecules. While other physicochemical features may also play a role (such as electrostatic charge), the size difference is the main feature driving the selective molecular permeability through nanoporous membranes.

While the development of this therapy is still in the preclinical stage, it offers promising advantages over traditional routes of delivery. Being target-selective provides advantages over other CSF clearance systems since the level of other proteins —not involved in disease pathogenesis—would be preserved. It also provides important advantages over “standard” peripherally administered drugs, including the following: 1. Acting continuously, on the CSF directly, is expected to be much more effective than acting peripherally. 2. Immunoisolation of drugs impedes immune responses, fully avoiding immunologically mediated side effects reported with biological drugs systemically administered [74,75].

In contrast, potential adverse effects related to the intrathecal system implantation and functioning should be taken into consideration, with expected local complications similar to those seen with intrathecal pumps, such as CSF leak, hemorrhage, and infection, along with device-derived problems such as CSF flow obstruction or even device disconnection [74,75].

The field of disease-modifying therapies for NDD is one of the hottest topics in medicine nowadays. Despite a myriad of studies, no effective disease-modifying treatment is available at the present for most of these conditions [76] while the first disease-modifying therapies for AD have been recently approved with some controversy regarding their efficacy and safety [77,78]. However, much knowledge has been accumulated regarding the molecules and cellular pathways involved in the pathogenesis of NDD that can become valuable targets for future therapies. Different classes of therapeutics are suitable to be used via intrathecal pseudodelivery in the treatment of NDD. Table 1 summarizes the most relevant NDD, their known molecular targets, and the therapeutic agents proposed to be applied through this route, based on previous evidence on the drugs’ mechanism of action. There is little research testing IT pseudodelivery in these conditions yet, hence this list should be considered just as therapeutic hypotheses today.

Monoclonal antibodies (mAbs) directed against misfolded proteins such as Aβ, Tau protein, or α-synuclein are a first choice when considering IT pseudodelivery, as they have been demonstrated to be effective when administered intravenously in many studies [118]. Moreover, mAbs targeting Aβ were approved very recently for the treatment of AD in humans (see Aducanumab [79] and Lecanemab [77]). mAbs is the only class of therapeutics with in vivo studies published via intrathecal pseudodelivery, which showed feasibility, good safety, and histological efficacy in animal models of AD [74,75]. Aptamers are an interesting class of compound that could replace antibodies in the near future, as they can also be used for therapeutic purposes within the pseudodelivery device. Compared to currently available mAbs, aptamers have some advantages such as a smaller size and mass, lower immunogenicity, greater replicability, and a greater level of control (high durability, sensitivity, and specificity) [80]. Similarly, antibodies and aptamers binding other pathogenic proteins such as Alpha-syn, Tau, TDP43, or mutant HTT might be of interest to treat other NDD via the pseudodelivery route, even if they failed when systemically administered for safety or efficacy reasons [58,85,86,96,97,103,105,115].

Other molecules binding pathogenic proteins can be of interest. For instance, human serum albumin (HSA) is a natural buffer of Aβ. A promising approach to AD prevention is to reduce the concentration of free Aβ by targeted stimulation of the interaction between HSA and Aβ. This approach can be implemented by pseudodelivering albumin alone [81] or in combination with agents increasing the affinity of HSA to Aβ through the action of HSA ligands [82].

Another therapeutic possibility is to act on the enzymatic dysfunction, a relevant example being the switch from the non-amyloidogenic pathway to the amyloidogenic one in AD [119]. In the same manner, compensating for the malfunctioning enzymes or even using different enzymes (from the family of membrane metallo-endopeptidases such as neprilysin and other Aβ cleaving enzymes [24]) inside the pseudodelivery device can be a smart option considering the high CSF throughput.

Protein conformation stabilization and aggregation inhibition that targets the upstream of the insoluble aggregate formation would be a promising approach toward the development of disease-modifying therapies for most NDD, particularly for polyQ diseases. PolyQ aggregation inhibitors of different chemical categories, such as intrabodies, peptides, and small chemical compounds, have been identified through intensive screening methods [116,117]. Among them, those with high molecular sizes are suitable to be used via IT pseudodelivery. The same approach could be used to inhibit the aggregation of Aβ, Tau, alpha-synuclein, SOD, and TDP43 [83,87,99,110,111,113,120]. In addition, clearing cofactors promoting protein aggregation, such as iron or tyrosine kinase, are an alternative way of inhibiting protein aggregation [121]. Interestingly, some nanomaterials such as polyoxometalates may also work as inhibitors of amyloid aggregation [84] and might be suitable to be used as therapeutic agents through this route.

Finally, another clear target in NDD are molecules involved in inflammation such as anti-TNF-α. According to several reports, anti-TNF-α agents may affect amyloidosis in inflammatory/autoimmune diseases, such as rheumatoid arthritis and familial Mediterranean fever [121]. Indeed, perispinal administration of the anti-TNF-α medication etanercept (a fusion protein produced by recombinant DNA) has been reported effective in cognitive improvement in one single case report [92], and similar results were obtained in animal studies [32]. Comparable results were noticed for infliximab, a chimeric monoclonal antibody already approved for the treatment of multiple autoimmune diseases such as Crohn’s disease, rheumatoid arthritis, and psoriasis. A study indicated that intracerebroventricular administration of infliximab reduced Aβ plaques and tau phosphorylation in APP/PS1 mice [93] and resulted in cognitive improvement in a human case [94], while recent research confirms the protective cerebral effects (reduced microgliosis, neuronal loss, and tau phosphorylation) of TNF-α inhibitors in a transgenic mouse model of tauopathy [114]. These results are encouraging, indicating that IT infliximab offers an alternative therapeutic approach for AD, and potentially for other neurodegenerative disorders whose pathogenesis involves TNF-α such as PD [101] and ALS [113]. Clinical trials for different conditions have shown a detrimental effect of TNF-α antagonists in advanced heart failure and anti-TNFs are associated with an increased risk of infection. Rare case reports of drug-induced lupus, seizure disorder, pancytopenia, and demyelinating diseases have been noted after systemic treatment with TNF-α antagonists [122,123]. Meanwhile, chronic dosing with a brain-penetrant biologic TNF-inhibitor induced hematology and iron dysregulation in aged APP/PS1 mice [95]. In this regard, IT pseudodelivery of anti-TNF-α agents may offer a safer route of administration.

Drugs targeting the complement component C5, CD19 on B cells, and the inter-leukin-6 (IL-6) receptor, have been used for the treatment of patients with refractory inflammatory CNS diseases. Particularly, Tocilizumab, a humanized, monoclonal antibody against the IL-6 receptor, has been tested for neurologic indications, such as neuromyelitis optica [124] or primary CNS vasculitis [125]. Tocilizumab has also been tested in ALS [112] and proposed in PD [100] and AD [91] As IL-6 is present in the CSF, monoclonal antibodies binding IL-6 directly—such as HZ-0408b [126]—via IT pseudodelivery might be an alternative route to target inflammation in NDD.

Lastly, a TREM2-activating antibody with a BBB transport vehicle enhances microglial metabolism in AD models [89] and tau pathology and neurodegeneration are associated with an increase in CSF sTREM2 [90]. However, some of these experiments can be interpreted as full-length TREM2 protecting rather than sTREM2 [26]. Therefore, while sTREM2 might be a suitable target via IT pseudodelivery in AD, more knowledge is needed to understand how, when, and in what cases this target might be of interest.