Amyotrophic Lateral Sclerosis (ALS), also known as Lou Gehrig’s disease, is a fatal neurodegenerative disease that is caused by the progressive loss of function in the upper and lower motor neurons in the CNS. One of the most famous personalities to be affected by this disease was Stephen Hawking! On average, about one in 50,000 people each year is diagnosed with ALS. The countries with the highest incidence rate of ALS are Uruguay, New Zealand and the United States.
i. Age Of Onset: 64 years in average
ii. Types of ALS:
There are two different types of ALS, sporadic and familial. Most of the cases are sporadic(90–95%), however about 5–10% of the cases are inherited (familial). The age of onset of familial ALS is at least a decade earlier than sporadic cases.
iii. Early symptoms include:
· Muscle twitches in the arm, leg, shoulder, or tongue
· Tight and stiff muscles (spasticity)
· Muscle weakness affecting an arm, a leg, the neck, or diaphragm
· Slurred and nasal speech
· Difficulty chewing or swallowing
With the progression of the disease, muscle weakness and atrophy spreads to other parts of the body. Individuals may develop problems with moving, swallowing (dysphagia), speaking or forming words (dysarthria), and breathing (dyspnoea). Death usually occurs due to respiratory failure following the loss of ability to breathe on their own.
iv. Genetic disposition:
The first gene associated with the cause of ALS (familial) is the superoxide dismutase gene, SOD1, two decades ago. SOD1 is a major antioxidant protein and a mutation in this gene could cause cytotoxicity. Since then, various other genes have been discovered which are mentioned in the following table:
v. Cellular Pathophysiology:
Dysregulation of axonal transport and the axonal compartment play a critical role in the pathophysiology of ALS. Several pathways may be responsible for the impaired axonal transport. Some of the most important mechanisms involve defective mitochondrial function or energy depletion, disruption of kinesin function by tumour necrosis factor, and excitotoxic damage by glutamate. Defective axonal transport causes an accumulation of neurofilaments, mitochondria, and autophagosomes in degenerated motor neurons. This leads to further hindrance of axonal transport and eventual motor neuron death.
Although several studies have been conducted to develop effective therapeutic interventions, ALS remains a progressive and incurable disease. Riluzole is the only FDA-approved drug that improves the survival of ALS patients whereas several other pharmacologic agents, apparently promising when tested in animal models, have failed when translated into clinical practice. The application of AI in the drug discovery process would shift the majority of the research in silico, thus increasing the effectiveness of the drug produced and also speed up the process.
I. TARGET IDENTIFICATION:
Two novel therapeutic targets were chosen for ALS- Epidermal Growth Factor Receptor (EGFR) and Matrix Metalloproteinase 9 (MMP9). Though these targets are conventional for cancer, recent studies have shown that these two targets also play a major role in ALS development.
The EGFR expression and activation are essential for the development of astrocytes and neurons in the developing CNS, but, in the adult differentiated CNS, EGFR expression in these cells is reduced, and its activation is absent in the adult brain. However, when the brain goes through a stressful situation such as oxidative stress, the EGFR expression is upregulated.
EGFR adds to the pathophysiology of ALS via:
1. Reactive astrocyte pathology (JAK-STAT, MAPK, ERK pathway)
2. Inhibiting autophagy (mTOR pathway)
MMP9 is activated by pro-inflammatory cytokines (TNF-α) that induces nerve damage.
It is a promising target for the following reasons:
· Studies have shown significant delay in fast muscle denervation in KO mice.
· Removal of the MMP9 gene in patients increases the lifespan by 25%
· MMP9 acts early in the disease pathway in cases with mutant SOD1 gene.
· Reduction of MMP9 levels due to knock-out or removal of the gene appeared safe for humans with no significant side effects.
· Even partial reduction in expression level is sufficient to confer benefits.
I. HIT IDENTIFICATION:
A cumulative of 15148 small molecule hits for EGFR and 4264 small molecule hits for MMP9 based on IC50 values were identified from ChemBL and BindingdB databases.
II. HIT PROPERTY VALIDATION:
The HIT data obtained from the databases were fed to the Property Predictor Generator on the BoltChem platform to create a property predictor model based only on IC50 values. This model would be capable of predicting the IC50 values of known and unknown molecules.
To validate the model, the smiles of known inhibitors (with known IC50 values) of both EGFR and MMP9 were given as input into the model and the predicted IC50 values were compared with the known ones.
The discrepancies between the predicted and known IC50 values were within the accepted range, thus validating the model.
III. SUBSTRUCTURE GENERATION:
The substructure generation process involves the generation of a framework based on which the novel lead molecule would be generated. Several mechanical and geometrical properties (for example, degrees of freedom of the molecule) are considered in this experiment to choose the important fragments or substructures for lead molecule generation.
For EGFR, an input of 6480 data points (all points within natural log IC50 values of -2 to 5) resulted in 2982 substructure generation.
For MMP9, an input of 4264 data points (all points within natural log IC50 values of -4.605 to 5.707) resulted in 1269 substructure generation.
IV. MERGING SUBSTRUCTURES
To generate dual-target lead molecules, the substructures generated for each target were merged to create molecules that could bind to both targets.
An input of about 200 substructures for each of the targets generated about 34000 substructures that are a combination of rationales of both the targets.
V. LEAD GENERATION
Once the framework of the lead molecule has been generated, the next step is the create the lead molecule itself. To generate the molecules, first, we need to build a generator that would create the lead molecules within the required IC50 value range. The number of molecules to be generated per rationale can be given as input in BoltChem.
Based on this generator, the lead molecules would be generated.
VI. NOVELTY CHECK
Once the lead molecules are generated, the smiles are screened over public databases such as ChemBL, PubChem, etc. to avoid duplicity and ensure novelty based on the fingerprint score or Tanimoto score.
Once the in silico process is done, the designed drug is needed to be generated in vitro to further the process of drug discovery. To be able to synthesize the lead molecule in vitro, retrosynthesis is done. Retrosynthesis is the synthesis pathway that is designed backwards by breaking the product into small known compounds.
Thus, shifting the drug designing process in silico can essentially reduce the amount of time required for creating a new drug from years to days which can accelerate the whole drug discovery process. Once synthesized in vitro, it can straight away go to the pre-clinical stage, thus saving at least 5 years in the process. The small molecule drug discovery platform, BoltChem saves time in drug designing and increases the effectiveness of the drug-designed manifolds.