Fragments of fetal DNA float in maternal blood. This discovery of circulating fetal DNA opened the way for non-invasive prenatal testing which, for instance, makes it easier to detect Down syndrome, also known as trisomy 21. It might also be replicated to other fields like cancer detection, tells Dennis Lo, who pioneered this research.
Ultrasound of a 13-week-old fetus. Photo credit Zephyr / SPL / Cosmos - DR
When a woman becomes pregnant, she would naturally wish to know about the health of the fetus she is carrying. One important determinant of the well-being of the fetus is its genome, encoded in DNA and carried by its chromosomes. For many years, obtaining this genetic information would require invasively sampling fetal cells, e.g. using a process called amniocentesis that obtains fluid surrounding the fetus. Unfortunately, such invasive sampling carries risks to the fetus and could result in miscarriage. Hence, for decades, many scientists around the world have been working on methods that are non-invasive towards the fetus, e.g., through the taking of a blood sample from the pregnant mother. One approach that had been attempted since the late 1960s involved trying to look for fetal cells that might have entered into the maternal blood. I had started working on approaches of trying to detect such circulating fetal cells since I was a medical student in Oxford in the late 1980s. I had continued this effort subsequently as a doctoral student, research fellow and faculty staff in Oxford until late 1996. Unfortunately, such cells are very few in numbers and many groups, including my own, had failed in developing practical and reliable non-invasive prenatal testing methods using such cells.
DISCOVERY OF CIRCULATING FETAL DNA
1997 was the year that Hong Kong had returned to Chinese sovereignty and my wife and I thought that it might be a good time to go back home, which would also give us more time to look after our ageing parents. I knew very well that after I had returned to Hong Kong, I would need to develop a new research direction. Interestingly, three months before I left Oxford for Hong Kong, I saw two papers in a journal called Nature Medicine, both co-authored by Philippe Anker and the late Maurice Stroun from the University of Geneva, in which they used aberrations in a class of genetic markers to demonstrate the presence of cell-free tumour DNA in the blood of cancer patients. The ‘cell-free’ part of the findings was most important. Our blood consists of the blood cells, i.e., red blood cells, white blood cells and platelets, which are bathed in a cell-free fluid called plasma. Conventionally, when scientists carry out DNA testing, they would focus on the white blood cells, which contain DNA within their nuclei. Looking for cell-free DNA in plasma, on the other hand, is rather counter-intuitive, as such a search implies somehow that the DNA is floating outside cells. Upon reading the two papers, I thought that the scenario of a tumour growing in a cancer patient had some similarities to a fetus growing within the womb of its pregnant mother. Hence, I wondered whether a fetus would also release its DNA in cell-free form into the blood of its pregnant mother. Furthermore, in my medical career, I had not yet seen a tumour as big as a 8-lb baby! Hence, I was cautiously optimistic that there was high likelihood that I should find cell-free fetal DNA in maternal plasma, at least in late pregnancy.
I next needed to decide on a method for extracting the DNA from the plasma of pregnant women. As I was transitioning my career from Oxford back to Hong Kong at that time, I had decided to use something that was very simple, namely, just by boiling the plasma for a few minutes. This was like cooking instant noodles, an activity that I sometimes did in my room when I was fed up with college food in Oxford. To prove that a DNA signal that I would see was from the fetus, I had decided to use a sequence that was present on the Y chromosome that was only present if a fetus was male. Excitingly, I was indeed able to see a Y chromosome signal in some of the boiled plasma samples from pregnant women. On checking, those signals were only present in pregnant women carrying male fetuses, and absent in those carrying female fetuses. These data were the first proof that there was indeed fetal DNA present in maternal plasma.
I would next like to know what was the concentration of such circulating fetal DNA and how it would change during the course of pregnancy. I had decided to use a new technology that was just becoming available then, a method called real-time quantitative polymerase chain reaction (PCR). The PCR is a method for amplifying DNA, thus creating a bigger signal that one can more readily detect. Real-time PCR is a variant of PCR in which one is continuously monitoring the amplification process so as to obtain quantitative information, e.g. for concentration measurement. The problem was that a machine for carrying out real-time PCR, a new technology then, was rather expensive, costing some 90,000 Euros. Interestingly, my new boss at The Chinese University of Hong Kong, Professor Magnus Hjelm, had a house near London and had invited me for a Christmas party at his house a few weeks before I was about to start my new job in Hong Kong. I had decided to ask Professor Hjelm to support me to buy such a real-time PCR machine. It took a bit of courage for me to ask because I had only met him once in person when I had my interview for the Hong Kong post. Luckily for me, Professor Hjelm immediately agreed to my proposal. This was probably the best Christmas present that I had ever received in my life! The machine was thus installed within a few months of my arriving in Hong Kong. Using this machine, I showed that some 3 to 10% of the total DNA in a pregnant woman’s blood plasma was from the fetus. Such concentrations were surprisingly high, bearing in mind the large difference in body size between a fetus and its mother. I also documented the variation in circulating fetal DNA concentration as pregnancy progresses. All of such information is critical for the subsequent use of circulating fetal DNA for non-invasive prenatal testing as one needs to decide how early in gestation would we have sufficient concentrations of fetal DNA for a robust test.
I was also interested to know what happens to the circulating fetal DNA after the baby was born. This is important because there are a number of publications demonstrating that a small number of fetal cells can persist within a woman’s body even after she has given birth to a baby. I would like to know if such persistence would also hold true for cell-free fetal DNA. I had therefore performed a study whereby I took blood samples from pregnant women from the time of Caesarean section onwards until 2 hours afterwards. Interestingly, I found that cell-free fetal DNA was cleared extremely rapidly following delivery. Indeed, most women would no longer had detectable fetal DNA in their plasma at 2 hours after delivery. This information is important because it demonstrates that cell-free fetal DNA would not persist in maternal plasma and so would not affect the results of non-invasive prenatal testing carried out in a second pregnancy.
FROM SCIENCE TO CLINICAL APPLICATIONS
With such background information, I was then ready to use circulating fetal DNA for prenatal testing. The first series of applications were all based on the detection of fetal DNA sequences that were inherited from the fetus’s father, and which were absent in the genome of its pregnant mother. One example is the detection of Y chromosomal sequences that a male fetus has inherited from its father. Detecting such sequences would allow one to determine the sex of the fetus, an application that is useful for the prenatal testing of sex-linked diseases, such as haemophilia and certain forms of muscular dystrophy. The disease genes of such disorders are present on the X chromosome. A male fetus only has one copy of the X chromosome (as compared to two in the case of a female fetus) and so is more susceptible to such disorders. Another example is the detection of a mutation that a fetus has inherited from its father.
However, the most common reason why a pregnant woman would opt for prenatal testing is to screen for fetal chromosomal abnormalities, such as Down syndrome. Down syndrome is most commonly caused by a fetus having an extra copy of chromosome 21. To detect fetal Down syndrome using circulating DNA is more challenging that detecting a fetus’s sex or whether it has inherited a mutation from its father. This is because the latter applications are qualitative analyses, amounting to a ‘yes’ or ‘no’ answer on whether the fetus has inherited a Y chromosome or a particular mutation. On the other hand, for detecting Down syndrome, one has to accurately measure how many copies of chromosome 21 that the fetus possesses. This is particularly challenging when the pregnant mother also possesses her own chromosome 21 which would also release DNA sequences into her own plasma. Hence, within the plasma of a pregnant woman, there is a mixture of DNA molecules released by the fetus’s chromosome 21 and the mother’s own chromosome 21. More difficult still, the fetus’s chromosome 21 sequences are the minority component in maternal plasma. Hence, over a period of 10 years, I had developed a number of different methods for achieving this goal. Eventually, in 2007, we had shown that by counting circulating DNA molecules one at a time, one could detect a slight increase in DNA molecules from chromosome 21 that is present in the plasma of a pregnant woman carrying a fetus with Down syndrome. In 2008, we had published a report showing that the use of a method whereby one randomly sequenced millions of DNA molecules in maternal plasma, one could detect fetal Down syndrome with very high sensitivity and specificity. This latter method becomes one that is now used in many countries around the world. We completed the first large scale clinical trial of this method in 2011. Within a few months, this approach was launched commercially in the USA. To date, non-invasive prenatal testing, now generally called NIPT, is available in dozens of countries around the world, and millions of pregnant women are tested annually. For example, some 4 million NIPTs are performed in China alone for 2017. As a result of the launch of NIPT, the number of conventional invasive prenatal tests have greatly reduced in numbers in many countries.
The types of chromosomal aberrations that can be detected using NIPT have grown from Down syndrome, to include those involving a number of other chromosomes, including chromosome 18, chromosome 13 and the sex chromosomes. Furthermore, a number of research groups have also shown that the approach can be applied to detect aberrations involving just a sub-region of a chromosome.
It is important to bear in mind that when used for detecting chromosomal aberrations, NIPT is used as a screening (albeit a highly accurate one), rather than as a diagnostic test. Hence, following an abnormal NIPT result, a confirmation step involving an invasive test, e.g. amniocentesis, would be necessary. Such a step is needed because NIPT can yield false-positive results. One reason for false-positivity is a statistical one as many laboratories use a result that is statistically significantly higher or lower than that from a healthy pregnant woman carrying a normal fetus as indicative of an ‘abnormal’ result. A second reason for false-positivity is a biological one in that the fetal DNA in maternal plasma is released by the placenta. In some cases, it is possible that the placenta may have a population of cells that have a chromosomal aberration, when the cells within the main body of the fetus is actually chromosomally normal.
On the other hand, it is also possible for NIPT to deliver a false-negative result, in that the fetus has a chromosomal aberration that is not seen in the NIPT result. One reason for false-negativity is that a maternal plasma sample contains a level of fetal DNA too low for the DNA test to work robustly. This is the reason why certain NIPT providers would first measure the concentration of fetal DNA in a maternal plasma sample before reporting an NIPT result. Another reason for false-negativity is that the placenta is chromosomally normal while the body of the fetus carries a chromosomal aberration.
Beyond Down syndrome to single gene diseases and the fetal genome
In addition to detecting chromosomal aberrations of the fetus, my group has also developed technologies for performing NIPT for single gene disorders. Examples of such disorders include beta-thalassaemia (an inherited form of anaemia common in the Mediterranean region and parts of Southeast Asia), congenital adrenal hyperplasia (an inherited endocrine disorder), etc. One such technology involves the measurement of the relative dosage of a mutant versus a normal gene, while another technology involves the measurement of the relative dosage of a series of genetic markers that are in the vicinity of the gene. Such a series of genetic markers constitutes a ‘haplotype’ and so the latter method is also called relative haplotype dosage (RHDO) analysis. The RHDO method appears to be very robust, due to the synergistic effect conferred by having multiple genetic markers.
In an attempt to explore the ‘limit’ of NIPT, I was interested in seeing whether one could sequence the entire fetal genome from maternal plasma. I initially had no idea how one would best do this. However, one day, I went to see the movie “Harry Potter and the Half-Blood Prince” with my wife. The first part of this movie was in 3D. I remember seeing the Harry Potter movie title flying out towards me. My eyes were caught by the two vertical strokes in the “H” of the word “Harry”. I suddenly realised that the way to decipher the fetal genome non-invasively from maternal plasma would be to divide the task into two halves, the first involving deducing the half of the fetal genome that it has inherited from its father, and the second involving deducing the other half inherited from its mother. Within a few months of this realisation, my group had indeed sequenced a fetal genome from maternal plasma. This approach had subsequently been confirmed by a number of other groups.
ETHICAL, SOCIAL AND LEGAL ISSUES
The development of NIPT has brought with it a number of ethical, social and legal issues. For example, in mainland China, there are strict laws against the prenatal determination of fetal sex for non-medical reasons. Regrettably, a number of companies have been set up in Hong Kong, a special administrative region of China, which illegally transfer maternal blood samples out of China, for conducting NIPT for fetal sex in Hong Kong. Tighter regulation would be necessary to stamp out such activities. The achievement of non-invasive prenatal fetal whole genome sequencing has opened up new debates as to the potential impact of such a technology. Currently, such a technology is still too expensive to be implemented in clinical practice. However, with the continual reduction in DNA sequencing costs, it is likely that eventually costs would no longer be a major obstacle. On the other hand, there are ongoing concerns with regard to the identification of ‘incidental findings’ with such genomic analyses and the frequently encountered difficulty in interpreting the clinical significance, if any, of an identified DNA sequence change. Hopefully, with our continual increase in understanding of the functions of the human genome, our ability to interpret genomic findings would gradually improve in the future.
IMPACT TO OTHER FIELDS
As discussed above, I was inspired to look for circulating fetal DNA after I had read about the presence of cell-free tumour-derived DNA in the blood of cancer patients. The global success of NIPT has encouraged many workers in the cancer field to replicate such successes in their field. This is a key reason for the surge in interest in the ‘liquid biopsy’ of cancer, especially using blood as that liquid. Indeed, work from my group and others have shown that many of the technologies developed for NIPT, such as genomewide sequencing of plasma DNA, are equally applicable in oncology. However, most of the efforts in this field have focused on use of such technologies for the monitoring or investigation of subjects who have already been diagnosed with cancer. A much more challenging area is to use circulating DNA technologies for the screening of cancer in healthy individuals. My group had, in August 2017, published our data on over 20,000 individuals in the New England Journal of Medicine on using circulating DNA to screen for nasopharyngeal cancer (NPC). NPC is a type of head and neck cancer common in south China. Without screening, only 20% of NPC cases in Hong Kong are diagnosed in stages I and II. However, with circulating DNA screening, we could increase the proportion of cases diagnosed in stages I and II to 71%. Follow up of the NPC cases identified by screening has revealed a significant improvement in progression-free survival. The challenge now is to see if such benefit of circulating DNA screening can be generalised to other cancer types using technologies that are both sensitive, specific and affordable.
Another field in which circulating DNA technology has created new impact is transplantation. Shortly after I had detected Y chromosome DNA from a male fetus in the plasma of its pregnant mother, I wondered whether a male liver or kidney transplanted into a female recipient would also release Y chromosomal sequences into the plasma of the recipient. Subsequent research showed that my thinking was indeed correct and that a transplanted organ would release its DNA in the blood stream of a transplant recipient. Furthermore, it was later found that when a transplanted organ was being rejected, the level of donor-derived DNA in the plasma of the recipient would increase. This area has thus developed into a novel method for monitoring patients following transplantation.
Circulating DNA thus appears to be a treasure trove for molecular diagnostics. It now seems that different cells within the body would release DNA into the blood stream when cells die. As many pathologies are associated with cell death, circulating DNA has the potential for detecting and monitoring many different diseases. Circulating DNA has thus opened up a non-invasive window into our health.
Doctor and a professor of chemical pathology
At 53, Dennis Lo is a doctor and a professor of chemical pathology at Li Ka Shing Institute of Health Sciences, The Chinese University Hong Kong, Shatin, New Territories, Hong Kong SAR, China. He is one of the pioneers of a revolutionary new method of non-invasive prenatal diagnostics.