The Language of Life is a book about the future by Dr. Francis Collins, published in 2010. Dr. Collins was head of the government’s Human Genome Project that won the race with a private enterprise to sequence the human genome. Because he won, access to the results is freely available. Had he lost, well that’s just too scary to contemplate. Collins acknowledges Saint-Exupéry’s words: ‘Your task is not to foresee the future, but to enable it’. The Human Genome Project has done just that. Dr. Collins now heads the NIH.
A subtitle for the book is suggested by the first chapter title: The Future Has Already Happened, speaking of course about the technology that will enable our future. The Language of Life is about the future of individualized medicine and its promise. But to realize this future promise, patients and providers need to become informed regarding the genetic basis of disease and its therapeutic treatment. Continuing research needs to be well-funded, in order to get the human genome, and the genomes of pathogens, to give up their secrets. The government needs to make enlightened policy. At odds with the suggested subtitle, this is a very different future from our present. Now our government, we the future patients, and the medical providers all need to play catch-up with the enabling technology.
Walter Gretsky’s advice was ‘Skate to where the puck is going to be’. Knowledge of our personal genomes will enable us to see ahead and plan ahead, so as not to accede to the inevitability programmed into uninformed lives. By being informed, we’ll be beating the grim reaper to the puck of life at critical junctures where life choices or therapeutic interventions must occur to prolong life. Winning the contests ensures being able to skate for many more years than otherwise would have been the case.
The Language of Life is highly technical, and what follows is a lengthy set of notes on what I learned. Real life medical situations, including some encountered by members of his own extended family, offer glimpses of what will soon be possible and even routine.
Cells and DNA
Our bodies contain ~400*(10^12) cells of many different types. Each cell contains essentially the same DNA as all the rest. But each different type of cell may use different genes and different gene transcription to perform its function, as would be expected for widely different types such as liver, muscle, and brain cells. Each time a cell divides, its entire genome must be copied. The copies are not always faithful; mistakes creep in. External environmental effects such as radiation and cigarette smoking can increase the chance for error. One class of errors results in cancer; the cells reproduce endlessly with turn-off switch disabled.
~50% of the human genome has been introduced over the eons via DNA parasites that insert their own sequences in our DNA. Once introduced, they can make copies of themselves and insert these randomly across the genome. A small percentage of these parasitic sequences have been found useful and thus were preserved and fixed by natural selection. Others have neutral effect on the organism and so remain as clutter.
The human DNA has only ~0.4% variability across all living humans. This commonality is the basis for our understanding that all modern humans descend from a progenitor population of ~10,000 individuals living ~100kya in East Africa. Other animals also have large similarities to human DNA, so most of what we learn about them will have application to humans also. One way of locating the important sequences in the vast universe of DNA is to compare the genomes of different species to detect sequences of bp that match. For such matching sequences to have been preserved over tens of millions of years implies that these have been the most constrained by evolution and thus serve the most important life functions.
Large-scale studies have begun to sequence the complete genomes of more than a thousand individuals across the world to obtain a clearer view of our genetic diversity (the variation present in the 0.4% of us that differs). The Encyclopedia of DNA Elements project (ENCODE) is attempting to identify all functional elements of human DNA and how they work together to turn genes on and off in different tissues. Other studies research model organisms. One will systematically ‘knock out’ each gene in the mouse genome to understand the purpose of each gene. Since 95% of mouse genes are matches to the human genome, the benefits of the study will be directly applicable to us.
We are all mutants. Everyone, in their 0.4% of variable DNA, has encoded risk for future disease. Help is here, in the form of a new medical paradigm. Each person can have his DNA sampled and interpreted. Such analysis provides the single most direct and important source of information regarding health and risk for illness for ourselves and our close relatives. An indirect path to the same goal is to compile a family health history. The author and colleagues have initiated a family health history initiative, to encourage and facilitate compilations of family health history data.
The new paradigm is important not only for our own longevity and well-being, but also for the financial condition of our society. The U.S. now spends 2 trillion dollars a year on health care, using the old paradigm of waiting for disease to strike, diagnosing the symptoms, and prescribing corrective action. But disease is neither random nor unavoidable. A better way is to deal with the underlying causes directly in order to prevent illness and its prohibitive cost. Our health is our responsibility, achieved through appropriate personal choices. The medical community does not have all the answers. We can assist them.
When Genes Go Wrong
We and all life inherit our DNA. Humans are diploid, which means our chromosomes occur in pairs (except for the sex chromosomes), so that we carry two copies of most of the genes in our instruction manual, one from father and one from mother. Mistakes in a gene may lead to disease or death. Some diseases, such as cystic fibrosis (CF), sickle cell anemia, and Huntington’s disease are caused by a single defective gene. These are simplest to understand. They are also the most dramatic, because they predetermine rather than merely suggest the existence of subsequent disease.
There are two types of single defect disease, recessive and dominant. A recessive disease (trait) requires both parents to have the defective gene. A dominant disease requires only a single parent to have the defective gene. CF is recessive. Huntington’s is dominant, as is neurofibromatosis and QT syndrome. Parents harboring a defective gene for a recessive disease are called carriers, because the disease can only occur in subsequent generations. Children of two carriers of a recessive disease have a 1 in 4 chance of being affected and a 50-50 chance of becoming a carrier. Children of a parent affected by a dominant disease have a 50-50 chance of being affected.
Incomplete penetrance describes the situation when defects in multiple genes are required for disease. Then, each individual defect will only predispose the person to disease, not predetermine that the disease will occur. An example is the BRCA1 gene. An inherited defect bestows women with an 80% probability of breast cancer and a 50% risk for ovarian cancer. The defect bestows only a modest risk of cancer for males. Most diseases are polygenic, involving multiple genetic risk factors, each bestowing only low risk. Then, disease only occurs in the presence of several factors, accompanied by environmental influences.
The most common cause (over 50%) of CF is a three-letter deletion (CTT) in gene CFTR, of unknown function on chromosome 7. CFTR codes for a large protein consisting of 1,460 different amino acids. The deletion causes the loss of amino acid 508, phenylalanine, dooming double recipients of defective CFTR to a shortened and very difficult life. Some 1,000 mutations of CFTR have been discovered that are capable of causing CF to some degree. Various of these alleles will spare some or most organs. Even after inheriting the primary allele, some recipients will manifest milder disease than others due to ever-present contributing factors such as normal variations in other pathways. As of 2009, the genetic causes of ~2,000 medical conditions have been learned.
Sickle cell anemia is due to another defective recessive gene, one producing hemoglobin. Here, the defective gene has natural selection on its side, because people who inherit just one copy have some protection to childhood malaria. But those with two copies will experience serious shortening of expected life span.
Once the genetic cause of a medical condition is determined, can a cure be far behind? Yes, unfortunately. Gene therapy, the implantation of replacement good genes in cells, either via infection using modified, inactivated viruses, or via inserting cells with good genes into bone marrow, is currently beset with great difficulty. Even if a massive virus attack succeeded in evading the body’s immune system, it is still very problematic that the resulting infection could insert enough replicable good genes to make a difference. Designer drug therapy, informed by DNA, is the best intermediate hope. One such designer drug for combating CF is looking promising in trials. It corrects the salt transport defect in airway cells. Sometimes, molecular analysis of disease will suggest an existing drug that can intervene, as is the case with use of losartan to help Marfan syndrome patients by binding to the TGF-beta protein whose overabundance contributes to aorta enlargement.
Screening is the gateway to the new paradigm, beginning with the parents before conception, and then again with the baby at birth (or prenatal if the preconception testing revealed potential for a genetic condition). Inherited disorders account for 5-10% of pediatric hospital admissions. The March of Dimes currently recommends screening for 29 conditions at birth. Each year, 4,000 infants are diagnosed with one of these disorders. Some states go even beyond this recommendation, and all states screen newborns for phenylketonuria (PKU), hypothyroidism, galactosemia, and sickle-cell anemia. CF screening is being added in many states now that treatment regimens are promised. A measure of progress in the new paradigm of preventive medicine will be the length of the list of conditions that can be screened.
About 0.1% of births involve diseases that could have been detected by recessive trait carrier screening of the parents, including Tay-Sachs, sickle-cell, CF, spinal muscular atrophy (SMA), fragile X syndrome. Currently, such tests are often done after a pregnancy is begun. But pre-conception testing preserves more options. Tay-Sachs was the first disease for which pre-conception carrier testing was available and widely used. Options to carriers are adoption, artificial insemination by a non-carrier, or prenatal testing with options for termination or advance preparation to deal with a child with special needs. Because the disease is so terrible, use of screening has been high and the number of Tay-Sachs births have been virtually eliminated. Sickle-cell screening has not been so successfully adopted because of cultural issues and lack of an effective prenatal test. Carrier screening is not always 100% predictive. For CF, 90% of carriers are detected because the test currently only looks for the most common 23 alleles of CFTR. Some tests are quite expensive. The SMA test costs several hundred dollars. A consensus whether to make fragile X test available has not been reached yet.
Difficulties will arise when screening gets ahead of itself, detecting conditions whose significance or mitigation is not yet fully understood. Such incomplete knowledge will be vexing for health care professionals and unnerving for parents. As costs drop, complete genome screening at birth will be available perhaps in the next decade, giving us a soft version of the scene from the 1997 movie GATTACA where one’s entire medical future is displayed at birth. But the benefits far outweigh any other considerations. Many detected conditions require no significant medical intervention, but rather just appropriate diet. PKU patients must live on an extremely low protein diet. The most common condition benefiting from screening may be a predisposition to obesity. 60-70% of body weight is determined genetically; several of the predisposing genes are known. The earlier diets are used to resist the effects of these genes, the healthier the life.
Maternal screening for birth defects are widely used: ultrasound for anatomic abnormality; blood tests for chromosomal disorders such as Down syndrome and neural tube defects. These tests do not yet provide definitive diagnoses, and the resulting uncertainties need to be carefully considered when providing results and options to the patient. As the techniques advance, direct screening of fetal DNA in maternal blood will make such screening more direct and predictive. Until then, fetal cell samples can be obtained via amniocentesis (fetal risk 1/400) and by chorionic villus sampling (CVS, fetal risk 1/200).
Pre-implantation Genetic Diagnosis (PGD) is available to couples choosing in vitro fertilization (IVF). At the 8-cell stage after fertilization, a cell biopsy can be performed on each embryo and analyzed, informing selection of which embryo to implant. Unfortunately, when carried to extremes, these methods can lead to ‘designer babies’. The U.S. has no standards yet for this testing, but the U.K. in 2006 authorized PGD to test only for inherited breast, ovarian and bowel cancer risk, limiting use to detection of single, highly penetrant mutant genes. Without such statement of principles, the U.S. is becoming increasingly in danger of sliding down the slippery slope into those practices fictionalized in the futuristic GATTACA. Already, sex selection is practiced here using PGA. A particularly ethically-challenging use is to select for a donor baby whose tissue could help save the life of a child previously born with a life-threatening disease. A successful case is cited where stem cells from the umbilical of a PGA newborn were transplanted to the bone marrow of an older sibling born with Fanconi anemia.
Those of us born without benefit of complete start-of-life screening may want to know if we have any genetic ‘time bombs’ ticking inside us, particularly if one’s family history has some disease represented. There are virtually no conditions for which heredity does not play some role. If we decide we want to know, genetic testing companies such as 23andMe can provide answers.
Obtaining the genetic answers has been the difficult work of genetic research in the past decade. Before that, luck was the main ingredient for success: guess which genes might be involved in a condition, then look for variants of those in people with the condition. With polygenic conditions, a systematic approach is needed. Take a thousand persons with a condition, a thousand without the condition, sequence the 2000 genomes, and compare results. Both the major risk factors and their relative strengths will be revealed with no reliance on luck. Complete sequencing is not required; the search can be reduced to regions where human genetic codes are known to have differences called SNPs, single nucleotide polymorphisms (one letter misspellings). This restricts the problem from scanning 3 billion bp to only looking at 10 million bp, a 300-fold reduction. Also, the DNA has regions where only certain combinations of SNPs occur, forming a grouping called a haplotype. Once haplotype is known, the SNPs are predictable in that region. The HapMap project identified these regional dependencies and reduced the problem by up to another 40-fold. Together with great advances in sequencing technologies, the cost of the systematic search for genetic precursors of a disease in a 2000 person study is reduced from $10 billion in 2003 to under $1 million in 2006.
The first polygenic success with HapMap came in 2005 when variations in two genes involving the inflammatory pathway were associated with an 80% risk for macular degeneration, when combined with two environmental risk factors, obesity and smoking. This laid to rest any skepticism about the efficacy of HapMap and a deluge of results began to accumulate. Other clues to the nature of macular degeneration were already available; sufferers of rheumatoid arthritis have low incidence of macular degeneration, probably due to high doses of anti-inflammatory drugs that are prescribed. Both high dietary omega-3 fatty acids and regular anti-inflammatory medication may offer protection to those having the risk factors.
In the five years since, the genetic predispositions for a variety of ailments have been identified, including diabetes, heart disease, common cancers, asthma, stroke, obesity, high blood pressure, gallstones, atrial fibrillation. Very few of these genetic glitches produce garbled proteins. Most adversely affect the timing of genes turning on and off. And unlike the macular degeneration case, for most conditions, the influence of any single gene is low, in the 10-40% range. Some genes are implicated in multiple conditions. A single gene is implicated in type 1 diabetes (T1D), rheumatiod arthritis, and Crohn’s disease. A small segment of chromosome 9 is implicated in both type 2 diabetes (T2D) and coronary heart disease.
More than a dozen risk factors for T1D have been found, and more than 20 genes are now known to influence T2D. Those whose function are understood impact on pancreatic beta cells. In the T2D case, two of the genes point to pathways currently pursued by current T2D drugs. So there is some hope that further understanding of additional genes will point to more drug pathway types. But lifestyle has been shown to be the most effective mitigator of T2D tendencies. A study showed that exercise and diet reduced risk 71%, while the beta cell-targeted drug Metformin reduced risk only 31%.
There are three objective considerations that will influence our desire to know our genetic risks:
- the size of the risk (relative risk times baseline risk),
- the burden of the disease,
- the interventions that are available.
Thus the RBI rule states the benefit of knowing equals risk*burden*intervention. When considering all risk factors, some gene mutations bestow positive risk and some negative. All are considered when assessing total risk. For T2D, the absolute risk (percentage of population with disease) is 23%. The most significant gene influencing T2D has a relative risk factor of 1.4, giving a total risk of 32%. Thus the average person has about a 1 in 4 chance of T2D, but a person with the defective gene has his total risk raised to 1 in 3. The combination of several genetic risk factors raises the relative risk to 2 in 3. Multiplication of risks suffices at the current state of knowledge, although there are indications that multiple risks can act synergistically to raise risk even higher.
Studies conducted on sets of identical twins show most common diseases have heritable risk of ~50%. But gene studies have not yet found this much discrete risk. Some theories speculate where the missing risk resides. It may reside in very small quantities distributed over many of the common SNPs, so that only very large sample populations would reveal it. It might reside in a very few uncommon variants that have very low frequency in the population, again requiring massive sampling to detect. It may reside in synergies between variants whose combinatorial power we haven’t yet measured. It could reside in another class of mutations called copy number variants (CNV) where a portion of DNA is repeated many times. Some CNVs are being associated with increased risk for autism and schizophrenia.
The UK BioBank project leads the way in learning more about how the environment influences susceptibility to common diseases. Since we cannot change our genomes anytime in the foreseeable future, addressing environmental risks is our near-term backup. BioBank has a population size of 500,000, large enough to assess where much of the missing risk resides. It also has a modest amount of environmental risk assessment. Other countries are also planning large scale studies: Japan, Germany, Estonia, Iceland. The U.S. is not among them. The American Genes and Environment (AGES) study was planned several years ago, but its $400 million per year price has not been funded.
[WS Aside: It seems apparent that for many broadly characterized illnesses such as heart disease, both the lack of condition specificity and the incomplete genomic inputs are yielding statistical generalizations that are of no diagnostic or pharmacogenic value. The HapMap approach, looking only at common variants found in DNA, is apparently missing much of the risk factor that is expected to be coded in DNA. In the near future, as costs of complete genome sequencing fall below $1000, the costs of large scale population studies necessary to identify rare risk factors will become more easily fundable, and the real value of the information locked on one’s DNA can be realized. Meanwhile, there are many cases documented where life-saving value has been realized from the currently available limited risk analyses.
Early adopters will need to fund their own testing. And it is probably better that this is the case in the long run, because if health insurance companies get involved, the individual may well be rated based on detected risk, making us all in effect self-insured (in a pool of 1). Perhaps the government will get its act together and establish health care regulations that would prohibit such rating or exclusion practices. In any case, it will be a strong signal to consumers that genetic testing is mature when insurance companies are able to do the cost-benefit analyses to justify insurance coverage for obtaining such testing. This will be the case when the genetic test results have widely applicable clinical validity and clinical utility.]
Direct-To-Consumer (DTC) genetic testing has been available for a few years. There are initially three labs that have pioneered this work: 23andMe, deCODE, and Navigenics. But many more are starting to spring up, so that caveat emptor will as usual be the watch word for the cautious consumer. Until then, early adopters will be challenged to derive useful medical information. But the exceptions are powerful motivators. The CSO of deCODE received a genetic risk of nearly double the average for prostate cancer. While PSA and physical data were still in the normal range, family history for cancer was strong; his father died at 68 of prostate cancer. The man, aged 48, decided to undergo a biopsy, which detected cancer at Gleason stage 6. He then opted for a radical prostatectomy, which further revealed advanced cancer at Gleason stage 7, essentially an early death sentence. Testing allowed him to dodge this bullet.
The efficacy of genetic testing at the current time is itself being tested by the Multiplex Project, covering 1,000 people in Detroit. One personal observation of the author was that while testing commonly only revealed slight elevations of risk, having this information motivates people to make and maintain life style changes designed to mitigate the risk. Thus, two important roles current testing can assume is providing further validation of family history indications, and providing a sense of urgency to control environmental risk contributors.
Both the FTC and FDA will need to assume roles in regulating DTC genetic testing. The FTC can help weed out the plethora of scammers that will inevitably rise to the occasion. The FDA can set standards and maintain an information database of respected labs that will inform the consumer decision process.
[WS Aside: Any FDA regulation, however, must not turn into heavy-handed intervention into the consumer’s right to know. The medical establishment is very conservative and resists strongly intrusions onto whatever it deems its home turf. This same establishment has strong influence at the FDA. An NGO, the American College of Medical Genetics, has appointed itself spokesperson for the establishment in this regime, and apparently has the ear of the FDA. It appears to want to restrict genetic testing to be on a prescription-only basis. Such heavy-handedness must be fought. We have benefited so much from an open genome and court rejections of gene patents. We must say no to 3rd-party meddling and interference when we are so close to the finish line.]
Before anyone undergoes genetic testing for medical diagnostic purposes, there are several considerations.
- Risk for most tested conditions rises very little above the average. As a guide, testing for 20 conditions raises likelihood of finding at least one risk in the top 5% of the population.
- Testing is a poor substitute for family history analysis at the present. Both are required to reach state of the current art.
- Because of their rarity, current testing does not look for many factors currently known to impose significant risk.
- More and more risks will be identified in the future, so that repeated testing over time will deliver increased confidence that all important risks are being identified.
- Lab mistakes are possible; lab quality performance should be assessed before testing.
- Different analysts interpret results in different ways.
- Data supporting current analysis comes from populations with northern European origin and may not be correctly extrapolated to other populations.
- Consumers should not rely on labs to provide intervention suggestions (the I in RBI). Intervention advice and planning should remain in the hands of medical professionals.
- Reported results will not always be understandable and transparent, and may cause anxiety.
- One may need help to understand results and place them in perspective. One’s physician may not be well enough informed to assist.
- Elevated risk for disease can be used against one in applications for long-term care and life insurance. Care must be taken regarding whom one shares one’s data with.
- Run away from any company that tries to sell you something based on your result.
The book goes on to detail research on genetic risks for cancer. The important finding is that while cancer is a disease of the genome, most cancer is not inherited, but rather is due to mutations occurring during our own lifetimes. Thus cancer is largely environmentally initiated. Major new insights are multiplying from the Cancer Genome Atlas (TCGA) project, a U.S. effort to catalog all mutations involved in three common cancers: lung, brain, ovarian. A similar sequencing of the leukemia genome has produced further insights. Molecular classification of various tumors is progressing. For example, such classification has led to a new drug Gleevec (Novartis) has been shown very effective against very different tumor types, leukemia and gastrointestinal stromal tumor. Because the genes involved in each produce a protein of similar molecular structure, Gleevec works on both. Another drug was found effective for a form of leukemia and HIV.
Cancer tumors can all have slightly different genetic signatures, so that treatments may work well for some and not at all for others. By learning the genetic codes of the tumors, treatment can be directed to those patients that it can help. The FDA’s old way of massive random drug trials will rule out such drugs, as it did for the drug Iressa in 2005, removing availability of a life-saving treatment for those that it does actually help. In the near future, DNA-targeted treatments for most cancers are promised. It’s been a long time coming.
[Aside: Going forward, the FDA is in position to do harm to the promise of individualized medicine, and in the process kill people unnecessarily, unless it embraces DNA-targeted studies of drug efficacy. Rather than side-tracking such treatments based on old-think, they instead should be fast-tracked. In our complex world, the government is frequently over-matched and two steps behind. We need smarter people in charge and a culture change that is adverse to subverting science to win bureaucratic turf wars. Whatever happened to ‘do no harm’? What a shame if going forward we have to look to other countries on a continuing basis for information and cures denied here because of an archaic FDA.]
Our bodies are immersed in a sea of pathogens from birth. The skin provides a mechanical barrier to the pathogens. Once breached, the immune system has a large number of response mechanisms (aided since the mid 20th century by antibiotics, immunization, and anti-virals). A significant fraction of our 20K genes are involved in our immune response.
The immune system can be compromised by viruses such as HIV. HIV, present in chimpanzees for a long time, jumped to humans around a century ago. Its genetic information is encoded in RNA, not DNA. It brings along its own enzyme for converting RNA to DNA, which is inserted in host T-cells and replicates itself, gradually destroying the immune system over a decade. HIV also frequently morphs, not giving the body’s defenses time to adapt to one variant before changing. Deployment of multi-drug anti-virals such as HAART have changed the HIV landscape, from death sentence to chronic disease.
Molecular exploration of HIV mechanisms found the specific cell boundary proteins that are used for binding; one is CCR5. A sad fact of the HIV epidemic was the deaths of many hemophiliacs via transfusions of infected blood. But a percentage of hemophiliacs were unscathed. When their CCR5-producing genes were examined, they showed a 32 bp deletion that altered the protein produced. Since 32 is not divisible by 3 (the word length of the amino acid genetic selection code), the deletion resulted in a DNA frameshift, a change of such significance that the CCR5 protein simply goes missing. 1% of those with European ancestry have this HIV-proof mutation. 12-16% inherit only a single copy of the mutation and these are only temporarily resistant to HIV. The mutation is known to have existed during the Bronze Age nearly 3000ya, but the initial benefit causing selection has not been identified.
A single case based on this evidence shows the potential for gene therapy to eliminate a disease from a human. A patient with both leukemia and AIDS needed a stem cell transplant to survive. The savvy attending physician sought a donor with the missing CCR5 protein. After the transplant, both leukemia and AIDS were eradicated from the patient. Other new drugs are being designed to bind to CCR5 protein to prevent HIV from penetrating the cell.
In addition to using genome data from the host human, it is necessary also to understand the genomic data from the disease organism. This helps us determine which parts of the organism tend to be invariant under the frequent mutation activities that confer drug resistance, so that these areas of greatest vulnerability can be targeted in designing vaccines for the long term eradication. Malaria, TB, and influenza are organisms being studied from this perspective.
The Human Microbiome Project is cataloging results of such studies, in effect describing a human super-organism consisting of our bodies’ 400 trillion cells and the additional 600 trillion largely symbiotic microbial cells that accompany each of us. It is suspected that disruptions to our microbiome may be implicated in human diseases such as various inflammatory ailments (eczema, ulcers, gingivitis, Crohn’s, etc.). A difficulty with studying many such organisms is they will not grow in a lab. Thus genetic classification may be important in unlocking the secrets of our microbiome. Even tendencies to obesity seem to have microbial basis. The intestinal flora of obese and non-obese individuals are seen to be quite different. That this difference might be causal rather than derivative is shown in mice, where transfer of obese mouse intestinal microbes to a lean mouse caused the lean mouse to gain weight.
A potential negative of individuals knowing their genetic risks for diseases may be that those with low risks may become too relaxed about healthy lifestyle choices. In spite of that, complete genome sequencing leading to predictions of disease susceptibility will become commonplace in a few years. Such genome knowledge may lead to improved prediction of vaccine response and medication effectiveness, in turn leading to individual-targeted dosage of both vaccines and medications. Knowledge of our microbiome will lead to more effective treatment and even detection of early warning signs of trouble. These improvements will be particularly important in increasing controllability and eventual eradication of infectious disease.
Genetics is the sole cause of very little disease. Even when genetics is substantially implicated, other agents may be involved. For example, Parkinson’s disease has 13 known genetic risk factors, but even those without such risk can acquire Parkinson’s through unwise life choices. Drug addicts can become poisoned with MPTP, a substance that is toxic to the dopamine-producing cells in the human brain. Early tests on rats failed to show this, but later tests on primates did.
Brain cellular complexity is mirrored by genetic complexity of its mechanisms. For example, one gene controls for 38 thousand unique proteins. Of all human organs, the brain is also most susceptible to external influence. Thus, even the brains of identical twins are profoundly different.
Neuro-degenerative diseases such as Hungtington’s, Parkinson’s, and Alzheimer’s diseases are heritable. Development of therapeutic drugs is still in its infancy.
All four forms of major mental disorders are known to have heritable factors: the schizophrenic, bi-polar, major depressive, and autistic disorders. Twin studies confirm this finding. An identical twin of a schizophrenic has a 50-50 chance of the disorder; a fraternal twin has only a 15% chance. An identical twin of a bi-polar person has a 60% chance of having the disorder; an identical twin of a major depressive person has a 40% chance of being so also; in autism, this concordance can be as high as 90%. However, the associated DNA mutations apparently reside in DNA ‘dark matter’, as yet undeciphered. Each disorder may be a common face for a variety of underlying diseases, or have a unique molecular DNA basis. Part of the difficulty might be that these disorders impact the probability of having offspring, so the mutations that give rise to them are necessarily ephemeral and will disappear from the population in a few generations. Supporting this hypothesis, CNVs have been associated with schizophrenia and autism, some large enough to duplicate or delete entire genes. Real progress must await complete DNA sequencing of a large number of people with such disorders to catch the current culprits before they vanish, only to reappear again in some other family tree.
There are possible genetic bases for various behavior disorders, such as situational depression, alcoholism, and nicotine addiction. The dynamics of these conditions suggest that genes are not a static pre-determinant, but operate over a lifetime as both cause and consequence of our actions, constructing and reconstructing the body and its machinery in response to biological timetables and experience (Matt Ridley: Nature vs. Nuture). The genetics of personality disorders can have two components: susceptibility to the disorder itself, and susceptibility to the conditions that entail from the disorder, such as lung cancer for smokers and cirrhosis for drinkers.
A genetic basis for personality types is also on the table. Robert Cloniger, in a twin study, found four personality characteristics with strong heritable basis: novelty-seeking, harm avoidance, reward dependence, persistence. Three other characteristics may be marginally influenced by genetics and appear only in adulthood: self-directedness, cooperativeness, self-transcendence (spirituality). Most likely, many genetic contributions are necessary to establishing a trait, so prediction will be elusive.
A genetic basis for criminality has been sought for decades, but the surest predictor is having a Y-chromosome, which confers a 16-fold likelihood of incarceration. A mutation on the MAOA gene on the X-chromosome has been seen to influence aggressiveness, but most variants have no effect unless childhood mistreatment was also present. Even if a genetic link to criminality is established, it is unlikely to be valid support for a criminal defense, just as having a Y-chromosome is not a defense for performing a criminal act.
Male fidelity may have a genetic basis, as established by vole studies. Prairie voles mate for life, montane voles have no attachment to any female. Extrapolating from voles, human V1AR (arginine vasopressin receptor gene) is implicated. Males with two copies of a risk allele had 34% marital crisis episodes in a year, vs 15% incidence for males with normal allele.
There appears to be a genetic link to homosexuality. If an identical male twin is exclusively homosexual, there is a 20-30% probability the other twin will be gay. This is a ten-fold increase over the baseline incidence of male homosexuality, 2-4%. The other current biological clue is that birth order confers increased probability: a male’s probability of being homosexual increases 30% for each older brother, giving the appearance that a maternal immune reaction to the Y-chromosome may be involved.
Difficulties in identifying heritable components of human intelligence probably indicates that hundreds of genes are involved. In large studies where environment was controlled, performance on IQ tests was estimated to be 50% genetic and 50% influenced by non-heritable factors.
There has been no credible evidence for a heritable disposition toward spirituality.
Understanding how the brain functions is a long term goal. But even our attempts to understand human consciousness have come up way short. Some posit that the human brain may not be powerful enough to analyze itself. Ongoing projects such as the Allen Brain Atlas are at work to systematically catalog patterns of gene expression in the brain.
Currently, psychiatrists refer to the DSM, a 1000 page Diagnosis and Statistics Manual for mental disorders (see Wikipedia). This subjective manual is due to be replaced when molecular classification of mental disorders matures.
The search for heritable contributors to aging is the modern search for the fountain of youth. We hope to rewrite Shakespeare’s last scene of all: ‘second childishness and mere oblivion, sans teeth, sans eyes, sans taste, sans everything’.
Aging rate seems related to organism complexity. Bacteria seem to have endless lifetimes when nutrients are available. Studies on simpler organisms, yeast, worms, flies, mice, have found effects of single genes can increase lifespan 5-fold, although application of this knowledge to humans is a long way off.
More limitations appear as life moves up the complexity scale. A natural limiting mechanism may be increased copy mistakes as cells repeatedly divide, reinforced by natural selection in ensuring that aging organisms do not compete for scarce resources with those better able to reproduce successfully. Nature programs humans to live long enough to support their offspring through their reproductive period, but not much longer.
The most significant known contribution to longer life is reduced caloric intake. As the calories decrease, insulin signaling decreases and the organism approaches a hibernation state of low metabolic activity. However, this reduction cannot proceed to malnutrition and maintain its longevity effects.
In yeast, genes producing sirtuins are up-regulated by loss of caloric intake. Artificial up-regulation might be possible to increase longevity. Red wine is known to activate sirtuins via the resveratrol molecule. Two drugs with greater potency than resveratrol are in clinical trials. The FDA does not recognize the category of longevity drugs, so the trials must target reduction of diabetes and heart disease; longevity will be noted as a side effect.
William Harvey noted in 1657: ‘It has been found in almost all things, that what they contain of useful or applicable nature is hardly perceived unless we are deprived of them, or they become deranged in some way’. So it is with looking for genetic clues to normal processes such as aging. There are known genetic bases for rare disease that disrupt the aging process. Rare diseases like Werner’s syndrome and Cocayne syndrome that accelerate aging have been determined to interfere with DNA repair machinery. Thus it seems that DNA integrity is correlated to longevity.
The age-accelerating disease progeria (HGPS) is caused by an SNP within sperm, in the gene coding for the protein lamin A. It is thought to perhaps be found in older men whose sperm have reproduced many times. Lamin A is responsible for maintaining the shape of the nucleus under cell division. It has a tail appendage, a tag, that allows it to target the nucleus for its activity. Before it becomes active, it must lose its tail. The HGPS mutation prevents the tag from being eliminated, so many cell divisions cause an excess of tags to build up in the nucleus, interfering with DNA integrity.
Cells have special machinery to keep the ends of chromosomes from fraying during repeated cell divisions. The tips of chromosomes are padded with telomeres, repeated sequences of 6 bases: TTAGGG. Without repair, these repeated sequences get shorter and shorter, finally exposing the chromosome itself to damage at which point the cell dies. The enzyme telomerase serves to re-extend the telomere repeat sequences. Stem cells and cancer cells have turned-on their telomerase genes. Most other cells no longer produce telomerase and eventually must die.
Longevity is clearly heritable. It was long observed that people that lived well and avoided catastrophic illness or accident could make it to 85 or so with ordinary genes. To get beyond that requires a special trait that seems to run in families. Now we are in a position to move beyond conventional wisdom. Studies limited to people over 70 have shown that more than 50% of proclivity to longevity is genetic. However, the search for the DNA components of this heritability is still in its infancy. It has been shown that adults with longer telomeres on their white blood cells live longer, and this trait is heritable. It has also been observed that people who score higher on an optimism scale have longer telomeres.
Longevity without good health is not to be wished for. Alzheimer’s is a scourge on the elderly that has a genetic basis. Five percent of cases are known to involve a mutation on either the amyloid gene or one of the enzymes that processes amyloid into amyloid-beta. Twin studies show that for the other 95%, genetics carries 70% of the risk for late-onset Alzheimer’s. The gene APOE has three alleles in the human population. The ‘E4’ allele carries significant risk: 3-fold risk for a single allele carrier, 8-fold for a double allele carrier. Carrier’s who experienced head injuries, such as boxers, had even greater risk, showing that environment may be implicated. The evidence is clear, but because there are no present effective interventions, there is no motivation for anyone to inquire about their genetic risk. And even carriers of the risk allele are by no means assured of succumbing to Alzheimer’s.
A study, Risk Evaluation and Education for Alzheimer’s Study (REVEAL), evaluated adult children of Alzheimer’s patients. They were divided into two groups, those who wanted to know their risk and those that did not. Those who were advised of their results and were at risk experienced no undue anxiety. Many attempted life changes, including not putting planned activities off to later years. They also watch their own mental performance more closely, looking for first signs of a problem. The study practices extensive genetic counseling and education, something to consider before casually offering the testing to the general public. Of the three individuals (genetic scientists) who first had their genomes sequenced, two chose not to have their E4 result revealed. The other peeked and found himself at risk.
We all have the ‘death gene’. How can we postpone its activation? We can avoid preventable chronic illness, reduce caloric intake, look for new drugs for stimulating sirtuins, and until then drink red wine. Alas, our generation will not be the one to find the fountain of youth.
Good news has arrived for progeria patients, children that age 7 times faster than normal. The drug farnesyl transferase inhibitor (FTI) has performed well in clinical trials, improving cardiovascular health in the participants. FTI may have application to the general population as well, for new studies show that we all gradually accumulate the same toxic protein that results from the defective lamin A gene.
The Right Drug at the Right Dose for the Right Person
DNA analysis promises to inform us about the efficacy of drugs based on our own personalized genome, putting out to pasture the ‘one size fits all’ approach to therapeutic medicine. The Dx-Rx paradigm is on the way; first will come the genetic test (Dx) and then the prescription (Rx).
Case Study: A 12yo girl was diagnosed with acute lymphocytic leukemia (ALL). The prescribed chemotherapy involves the drug 6-MP (mercaptopurine). Because the medical center was at the cutting edge, they knew that 1 in 300 people cannot metabolize 6-MP due to a missing enzyme. A DNA test confirmed the patient was at risk, and the dosage was reduced appropriately, perhaps saving the girl’s life.
One size fits all statistics: A correct diagnosis, followed by prescribing the appropriate drug at the appropriate dosage, will only positively benefit 70-80% of patients. Some of the remainder may suffer a toxic reaction, a violation of ‘First, do no harm’. Virtually no drug gets perfect marks. In the US each year, 2 million hospitalized patients suffer serious adverse drug reactions, and 100,000 of these are fatal, making adverse drug reactions the 5th leading cause of death in the US. And these are only the hospital cases for which FDA reporting is mandatory.
Perhaps the largest reason for these bad outcomes is bad handwriting, making computerized record keeping a very high priority. Human error extends to the patient, particularly one taking a multitude of drugs. Underlying illness of liver and kidneys that compromises metabolism is also a significant contributor. Interactions between drugs is yet another factor. Yet, even if all these factors are addressed, a significant number of bad outcomes will still be experienced, due to a patient’s DNA. Enter the field of pharmacogenics, the study of how drug response is influenced by genome.
A drug may be administered in a prodrug delivery vehicle, relying on an enzyme to convert it into the active-state drug. The drug relies on other enzymes that convert it into an inactive metabolite so that it can be excreted. The drug must itself interact with some biological component in the body, the receptor, to have its prescribed effect. Each of these process steps is encoded by genes, so it is easy to see how different reactions are possible in different people. Depending upon the activity levels of the steps, the result may range from ineffective to toxic, so that people need various dosages depending on their genes.
With ALL, the right dosage for someone with a non-functioning variant of enzyme thiopurine methyltransferase (TPMT) is much less than for one with a normal TPMT, since this enzyme converts the 6-MP drug to an inactive metabolite; without it, the body cannot excrete the 6-MP. But this enzyme malfunction is not only of concern to the rare ALL patients. It is needed also for converting the rheumatoid arthritis drug azathioprine, which in normal dosage is toxic to those with defective TPMT.
The individual response to the anti-clotting drug clopidogrel is variable, due to some people having a low-functioning CYP2C19 enzyme. These patients need a higher than normal dosage.
Antidepressants are difficult to prescribe, because the condition is vaguely defined, and because the drugs take a long time to reach effectiveness. A genetic approach to selecting the right drug and dosage would have a big payback, but such is not yet a possibility.
It is the drug coumadin that may be the first prescribed under the Dx-Rx paradigm. In 2004, 31 million prescriptions were written for this anti-coagulant. Yet it is listed among the 10 drugs with the largest number of serious adverse effects each year, and it ranks first on U.S. death certificates in deaths attributed to adverse drug effects. Proper dosage is a nightmare for physician and patient, and blood levels must be constantly monitored. Individual response to the drug can vary ten-fold. More than half this variability is due to factors other than physical characteristics of the patient: bmi, gender, age, … Forty percent of the hidden variability in coumadin response has been traced to variations in genes CYP2C9 (converts coumadin to inactive metabolite) and VKORC1 (regenerates vitamin K clotting factor). The FDA may some day advise physicians to refer to patient genome data to properly prescribe coumadin dosage.
Abacavir blocks replication of HIV. About 6% of recipients exhibit hypersensitivity. A single gene variant, HLA-B*5701, was shown to be responsible; it’s a group of genes coding for proteins involved in immune response. The FDA dragged its feet, requiring yet another study, but in 2008 finally recommends genetic screening before prescribing abacavir.
Statins cause muscle damage in about 2% of patients. This has been traced to a variant of gene SLC01B1, which encodes a liver transporter that mediates statin uptake. A carrier of the variant has a 4-fold risk of muscle toxicity; a double carer has a 16-fold risk, and should be discouraged from taking statins.
Trastuzumab is a monoclonal antibody prescribed to treat breast cancers that express HER2 receptor. It can only be prescribed if the tumor has been genetically tested to ensure HER2 is being expressed. More complicated is the test for appropriateness of cetuximab and panitumumab, directed against growth factor EGFR. Some patients did not realize benefit, due to their tumors also carrying an activating mutation for oncogene KRAS, which is downstream of EFGR on the signaling pathway. For them, shutting down EGFR has no effect.
For over 30 years, breast cancers that test positive for estrogen receptor have been treated with Tamoxifen, but not all patients experienced a good outcome. This was traced to a low-activity variant of enzyme CYP2D6, which converts the Tamoxifen prodrug into its active form. No recommendations have yet been issued, but a test for the CYP2D6 variant should be considered before prescribing Tamoxifen.
There are several obstacles to the pharmacogenomics revolution:
- Studies to detect rare adverse reactions would require tens of thousands of individuals, in order to obtain useful data. Rather, we need to use the entire prescribed base as the study, by setting up an effective system of reporting adverse effects, together with collecting a sample from the adversely-reactive patient for genomic analysis. Lack of such reporting is a major flaw in U.S. medicine.
- Motivation is lacking for for-profit drug companies to conduct studies regarding people who should not be prescribed their drugs. Such studies should be sponsored by the NIH.
- The FDA requires several independent studies before requiring genetic testing prior to drug prescription. Their reluctance acts as an inhibitor.
- Health care providers drag their feet at the extra expense of the Dx-Rx paradigm.
- The ‘test DNA only when sick’ approach suffers from the logistics of testing; treatment often needs to be started immediately, but test results can take weeks. The only remedy is for people to sequence their genomes ahead of sickness.
The Future Still Yet To Come
Gene therapy and stem cell therapy are in store in the future. Gene therapy happens ex vivo or in vivo. Ex vivo therapy involves removing cells such as bone marrow, adding the gene therapy vector, culturing the cells in the laboratory, and then reinfusing the treated cells. In vivo therapy involves injecting the gene therapy vector directly into the tissue to be treated.
Severe combined immunodeficiency (SCID) due to a mutation of the ADA gene was the target of the early attempts at gene therapy. It was a fatal disease, so risks were tolerable. The ex vivo approach was possible, because immune system cells can be harvested from bone marrow. It was thought that treated (normal ADA gene) cells would have a competitive advantage when reinfused.
The results were compromised by simultaneous drug therapy in the first test, but it appears not to have worked as well as hoped. There are three large challenges that are now appreciated.
- Delivery: DNA is a large, charged molecule. Transmission through both the cell and nuclear membranes is difficult. Initial attempts to use inactivated viruses as delivery vehicles has proved difficult.
- Function: Getting the corrected DNA to be transcribed into RNA is necessary for success. The viral vector must integrate itself with the correct chromosome. But such integrated genetic material is sometimes shut down by neighboring DNA.
- Immune Response: No matter how clever the bioengineers are at disguising the delivery virus, the immune system is even more clever, eventually recognizing the treated cells and destroying them.
In 1999, a volunteer patient died while undergoing in vivo gene therapy. This slowed research activity and progress. The second generation of therapy experimentation targeted the X-linked form of SCID. Of the 20 patients who underwent treatment, some seemed to benefit, but 5 developed leukemia because the delivery virus is thought to have activated an oncogene. Four of these cancer cases were successfully treated, and the lives of these patients has been extended by the therapy. A recent report from Italy details success treating 10 children with ADA-deficient SCID. All are reported cured by gene therapy.
Successful in vivo treatment has been pioneered with the congenital blindness condition called LCA. Results have been uneven, but many patients have had usable vision restored.
Going beyond delivery of normalized genes into existing cells, stem cell therapy allows us to deliver the entire normal cell. Stem cells are a type of precursor cell that is able to differentiate into different types of cells, given appropriate stimulus. The initial fertilized egg cell is a totipotent stem cell; it will divide to become every type of cell in the body and placenta. At the 100 cell fetal form, the cells have differentiated into two types, one type for the placenta and one for fetal cells. The fetal stem cells are called pluripotent because they will still divide to form all cell types in the body, but they no longer have the potential to form placental cells. Stem cells lose their pluripotency as the body develops, becoming merely multipotent. It is these adult multipotent stem cells that is the focus of stem cell therapy. Transplantation of bone marrow from donor to patient enables the stem cells to repopulate the patient’s blood-forming and immune systems. Some hope of further differentiation of adult stem cells raises the possibility of other uses, such as repair of heart muscle after a heart attack.
For its maximum benefit, stem cell therapy must use pluripotent cells. Thus there is great interest in fetal stem cells, as well as great medical ethics debates. The use of IVF embryos that are not implanted (now just frozen and ultimately discarded) is the favored approach. Federal funding is disallowed for research involving embryos created specifically for that purpose. With parental consent and absent inappropriate incentives, extra IVF embryos can be used in federally-funded research.
The initial target of embryonic cell research is spinal cord regeneration. Trials in humans are commencing. Treatments for diabetes and Parkinson’s are also logical targets. But the three challenges that gene therapies face also apply to some extent to stem cell therapy.
Cloning, or somatic cell nuclear transfer (SCNT), was considered to regenerate the totipotent cell identical to the fertilized cell from which the person originated. Such cells would avoid the major problem of rejection. Ethics would prohibit implantation of such cells in a uterus. The SCNT success inspired Japanese researcher Yamanaka in 2006 to develop a method of turning a skin cell into a pluripotent cell. In 2008, pluripotent cells were derived from cells from a single human hair. Yamanaka reasoned that the cytoplasm of an egg provides some finite set of signals able to reprogram a skin cell. He discovered just 4 genes that are able to transform a skin cell into a pluripotent stem cell. Subsequent research has been able to do this trick with just a single gene. The induced pluripotent stem cells (iPS) are being generated from a wide variety of individuals carrying a variety of inherited diseases, to aid in disease research. iPS treatment will raise self-healing to a new level.
Tempered optimism is best, though, because one of the genes used to create an iPS is an oncogene, raising the specter of inducing tumors with such treatment. On the upside, a mouse experiment illustrates the potential future synergy between iPS and gene therapy. The following is not science fiction. A mouse with sickle-cell anemia had iPS cells derived from skin cells, then treated via ex vivo gene therapy to eliminate the sickle mutation. They were then differentiated into bone marrow stem cells and reinfused into the mouse. The mouse was cured.
Where is this going? The first law of technology states the consequences of a new technology are always overestimated in the short term and underestimated in the long run. [WS Aside: The same short-sighted human syndrome is at play with global threat predictions. Considering global warming or all other predictions of dire consequences of human behavior on Earth, the consequences for the short term are always overblown, but the ultimate threat scenarios in the distant future are never fully recognized. In this case, the crying wolf too often in the short term, with no wolf materializing, seriously jeopardizes our ability to muster the collective human will to address the long-term issues.]
To realize the full potential of our new medical paradigm, we need:
- Research: We must do better by funding fundamental medical research or we will lose a generation of researchers.
- Electronic Records: Individualized medicine is not possible without electronic records.
- Good Policy: The average time from discovery to implementation of medical interventions is 24 years. Government oversight is responsible for a lot of the foot-dragging.
- Education: Medical practitioners need to become aware of, and then to embrace individualized medicine. Dr. Collins claims anyone who reads his book almost certainly knows more about the subject than their primary care physician.
- Ethical Decision Making: Endless ethical arguments and roadblocks slow down medical progress. The Presidential Bioethics Commission in the U.S. has no authority and is overrun by the politicization of the appointment process. We need a more authoritative model such as the UK’s Human Fertilization and Embryology Authority.
Family History Compilation:
(e.g. click on ‘human cortex’ and search on MAPT (microtubule-associated protein tau). Click on section 80561119 from donor 2898. Blue stain shows where MAPT is being expressed. Pan and zoom; zoom in to see individual neurons.