Diseases that run in families usually have genetic causes. Some are genetic mutations^[1] that directly cause the disease if inherited. Others are risk genes^[2] that affect the body in a way that increases the chance someone will develop the disease. In Alzheimer’s disease^[3], genetic mutations in any of three specific genes can cause the disease, and other risk genes either increase or decrease the risk of developing Alzheimer’s.

Some genetic mutations or variants interact with other genetic alterations that lead to Alzheimer’s disease. In some cases, gene alterations can interact with Alzheimer’s-causing genetic variants in a way that proves beneficial; they actually suppress the pathological brain changes the other mutations would normally lead to. These protective gene variants can drastically slow or prevent cognitive decline. In two recent^[4] case reports^[5] on familial Alzheimer’s disease, mutations delayed Alzheimer’s symptoms for decades.

I am a neurologist and neuroscientist^[6] who has spent my career studying Alzheimer’s disease and dementia both in the laboratory and in the clinic. Determining how genes affect brain chemistry is vital to understanding how Alzheimer’s disease progresses and devising interventions to prevent or delay cognitive decline.

The amyloid hypothesis

In the early 1990s, scientists proposed the amyloid hypothesis^[7] to explain how Alzheimer’s disease develops. The first neuropathological changes detected in the brain of Alzheimer’s disease patients were the formation of amyloid plaques^[8] – clumps of protein pieces called beta-amyloid. Other changes in the Alzheimer’s brain, such as the accumulation of another type of abnormal protein called neurofibrillary tangles, were thought to develop later in the course of the disease.

Beta-amyloid begins to accumulate in the brain up to 15 years^[9] before symptoms emerge. Symptoms correlate with the number of neurofibrillary tangles^[10] in the brain – the more tangles, the worse the cognition. Researchers have tried to determine whether preventing or removing amyloid plaques from the brain would be an effective treatment.

Imagine the excitement of the scientific community in the 1990s when researchers identified three different genes causing familial Alzheimer’s disease – and all three were involved with beta-amyloid.

The first was the amyloid precursor protein^[11] gene. This gene directs cells to produce the amyloid precursor protein, which breaks down into smaller fragments, including the beta-amyloid that forms amyloid plaques in the brain.

The second gene was termed presenilin 1, or PSEN-1^[12], a protein needed to cut the precursor protein into beta-amyloid.

The third gene, presenilin 2, or PSEN-2^[13], is closely related to PSEN-1 but found in a smaller number of families with familial Alzheimer’s disease.

These findings added strength to the amyloid hypothesis explanation of the disease. However, uncertainty and opposition to the amyloid hypothesis^[14] have developed over the past several decades. This was in part tied to a recognition that several other processes – neurofibrillary tangles, inflammation and immune system activation – are also involved in the neurodegeneration seen in Alzheimer’s.

The hypothesis also^[15] got significant pushback^[16] after many clinical trials^[17] attempting to block the effects of amyloid or remove it from the brain were unsuccessful^[18]. In some cases, treatments had significant side effects. Some researchers have come up with strong defenses^[19] of the hypothesis. But until a clinical trial based on the amyloid hypothesis could show definitive results, uncertainty would remain.

Genetic discoveries with treatment implications

The vast majority – more than 90%^[20] – of Alzheimer’s cases occur in late life, with disease prevalence increasing progressively from age 65 and up. Such cases are mostly sporadic, with no clear family history of Alzheimer’s.

However, a relatively small number of families have one of the three known genetic mutations that cause the disease to be passed down. In familial Alzheimer’s^[21], 50% of each generation will inherit the mutated gene and develop the disease much earlier, usually from their 30s to early 50s.

In 2019 and 2023, researchers identified changes in at least two other genes that markedly delayed the onset of disease symptoms in people with familial Alzheimer’s disease mutations. These mutated genes were found in a very large family in Colombia whose members tended to develop Alzheimer’s symptoms by their 40s.

A woman in the family^[22] carrying a mutated PSEN-1 gene did not have any cognitive symptoms^[23] until she was in her 70s. A genetic analysis showed that she had an additional mutation in a variant of the gene that codes for a protein called apolipoprotein E^[24], or ApoE. Researchers believe the mutation, called the Christchurch variant^[25] – named after the city in New Zealand where the mutation was first discovered – is responsible for interfering with and slowing down her disease.

Importantly, her brain had a great deal of amyloid plaque but very few neurofibrillary tangles. This suggests that the link between the two was broken and that the suppressed number of neurofibrillary tangles also slowed down cognitive loss.

In May 2023, researchers reported that two siblings in the same large family^[26] also did not develop memory problems until their 60s or late 70s and were found to carry a mutation in a gene that codes for a protein called reelin. Studies in mice suggest that reelin has protective effects against amyloid plaque deposition^[27] in the brain. In these patients’ brains, as with the patient who had the Christchurch variant, there were extensive amyloid plaques but very few neurofibrillary tangles. This observation confirmed that the tangles are responsible for the cognitive loss and that there are several ways to “disconnect” amyloid and neurofibrillary tangle accumulation.

Finding medicines that might mimic the protective effects of the Christchurch variant or the reelin mutation could help delay Alzheimer’s disease symptoms for all patients. Since the vast majority of nonfamilial Alzheimer’s manifests after age 70 or 75, a 10-year delay in the emergence of first symptoms of Alzheimer’s could have a massive effect in decreasing the prevalence of the disease^[28].

These findings demonstrate that Alzheimer’s can be slowed and will hopefully lead to additional new therapies that can someday not only treat the disease but prevent it as well.

Starts and stops

Despite over 20 years of doubts and therapy failures, the past several years have seen positive results from three different treatments – aducanumab, lecanemab and donanemab – that remove amyloid plaques and slow loss of cognitive function to some extent. Although there is still discussion of how much slowing of decline is clinically significant, these successes provide support for the amyloid hypothesis. They also suggest that other strategies will be needed for optimal treatment.

The U.S. Food and Drug Administration’s 2021 approval of the first antibody treatment for Alzheimer’s, aducanumab, sold under the brand name Aduhelm^[29], was controversial. Only one of the two clinical trials testing its safety and effectiveness in people yielded positive results. The FDA approved the drug on the basis of that single study through an accelerated approval process^[30] in which treatments meeting an unmet clinical need can receive expedited approval.

The second antibody, lecanemab, sold as Leqembi^[31], was approved in January 2023 via the same accelerated approval pathway. It was then fully approved^[32] in July 2023.

The third antibody, donanemab, completed a successful phase three clinical trial^[33] and is awaiting more safety data. When that is submitted to the FDA, the agency will consider the drug for approval.

Have you ever wondered why certain medications don’t seem to work as well^[1] for you as they do for others? This variability in drug response is what pharmacogenomic testing hopes to explain by looking at the genes within your DNA.

Pharmacogenomics, or PGx^[2], is the study of how genes affect your response to medications. Genes are segments of DNA^[3] that serve as an instruction manual for cells to make proteins. Some of these proteins break down or transport certain medications through the body. Others are proteins that medications target to generate a desired effect.

As pharmacists^[4] who see^[5] patients who^[6] have stopped multiple medications because of side effects or ineffectiveness, we believe pharmacogenomic testing has the potential to help guide health care professionals to more precise dosing and prescribing.

How do PGx tests work?

PGx tests^[7] look for variations within the genes of your DNA to predict drug response. For instance, the presence of one genetic variant might predict that the specific protein it codes for is unable to break down a particular medication. This could potentially lead to increased drug levels in your body and an increased risk of side effects. The presence of another genetic variant might predict the opposite: It might predict that the protein it codes for is breaking down a medication more rapidly than expected, which may decrease the drug’s effectiveness.

For example, citalopram is an antidepressant^[8] broken down by a protein called CYP2C19. Patients with genetic variants that code for a version of this protein with a reduced ability to break down the drug may have an increased risk of side effects.

Currently, there are over 80 medications with prescribing recommendations^[9] based on PGx results, including treatments for depression, cancer and heart disease. There are commercially available PGx tests that patients can have sent directly to their doorstep with or without the involvement of a health care professional. These direct-to-consumer PGx tests collect DNA from either a saliva sample or cheek swab that is then sent to the laboratory. Results can take anywhere from a few days to a few weeks depending on the company.

Some companies require a consultation^[10] with a health care provider, often a pharmacist or genetic counselor, who can facilitate a test order and discuss any medication changes once the results come back.

Limitations of PGx testing

PGx testing will not be able to predict how you will respond to all medications for several reasons.

First, most PGx tests do not look for every possible variant^[11] of every gene in the human genome. Instead, they look only at a limited number of genes and variants strongly linked to specific drugs. PGx tests can predict how you will respond only to medications associated with the genes it tests for.

Some drugs are broken down in very complicated pathways entailing multiple proteins and byproducts, and the usefulness of PGx testing for them remains unclear. For example, the antidepressant bupropion^[12] has three major pathways involved in its breakdown and forms three active byproducts that can interact with other drugs or body processes. This makes predicting how you will respond to the drug much more challenging because there is more than one variable involved. In many cases, there also isn’t conclusive data to confidently predict the general function of a protein and how it would affect your response to a drug.

The applicability of PGx test results is additionally limited by a lack of diversity of study participants^[13]. Typically, populations of European ancestry are overrepresented in clinical trials. An ongoing research initiative by the National Institutes of Health called the All of Us Research Program^[14] aims to address this issue by collecting genetic samples from people of diverse backgrounds.

Another limitation of direct-to-consumer PGx tests is that they can predict drug response based only on your genetics. Lifestyle and environmental factors^[15] such as your age, liver or kidney function, tobacco use, drug interactions and other diseases can heavily influence how you may respond to medication. For example, leafy greens with high amounts of vitamin K can lower the effectiveness^[16] of the blood thinner warfarin. But PGx tests don’t take these factors into account.

Finally, your PGx results may predict that you may respond to medications differently, but this does not guarantee that the medication won’t have its intended effect. In other words, PGx testing is predictive rather than deterministic.

Risks of PGx testing

PGx testing carries the risk of not telling the whole story of drug response. If variations within the gene are not found, the testing company often assumes the proteins those genes code for function normally. Because of this assumption, someone carrying a rare or unknown variant may receive inaccurate results.

It may be tempting for some people to see their results and want to change their dose or discontinue their medications. However, this can be dangerous. Abruptly stopping some medications may cause withdrawal effects. Never change the way you take your medications without consulting your pharmacist and physician first.

Sharing your PGx test results with all the clinicians involved in your care can help prevent medication failure and improve safety. Pharmacists are increasingly trained in pharmacogenomics and can serve as a resource to address medication-related questions or concerns.

PGx tests that are not authorized by the Food and Drug Administration cannot be clinically interpreted and therefore cannot be used to inform prescribing. Results from these tests should not be added to your medical record.

Benefits of PGx testing

Direct-to-consumer PGx testing can empower patients to advocate for themselves and be an active participant in their health care by increasing access to and knowledge of their genetic information.

Patients’ knowledge of their PGx genetic profile has the potential to improve treatment safety. For example, a 2023 study of over 6,000 patients in Europe found that those who used their PGx results to guide medication therapy were 30% less likely^[17] to experience adverse drug reactions.

Most PGx test results stay valid throughout a patient’s life, and retesting is not needed^[18] unless additional genes or variants need to be evaluated. As more research on gene variants is conducted, prescribing recommendations may be updated.

Overall, genetic information from direct-to-consumer PGx tests can help you collaborate with health care professionals to select more effective medications with a lower risk of side effects.