Journal Volume 3 - June 2008
Next Generation Molecular Diagnostic Assays for MS and Other Demyelinating Diseases
Eric M. Eastman, Chief Science Officer, and Douglas Bigwood, SVP of Biostatistics and Analysis, DioGenix, Inc.
This article will discuss new approaches to the diagnosis of demyelinating diseases such as Multiple Sclerosis (MS) and Transverse Myelitis (TM). I will introduce you to some concepts that will be new to many of you, although you may have read about them in recent newspaper or magazine articles or heard stories about them on NPR (see references below). Although some of these concepts may seem complicated, I will do my best to define each concept as we go and put them into the proper context.
Although I will focus on issues relating primarily to MS, many of the issues discussed here are relevant to other demyelination diseases including TM, neuromyelitis optica (NMO), optic neuritis (ON), and acute disseminated encephalomyelitis (ADEM) – see disease descriptions at the TMA website, www.myelitis.org. All of these diseases result from damage to the central nervous system (“CNS” – brain and spinal cord) resulting from malfunctions in the body’s own immune system.
The immune system is the body’s major defense system and is made up of a large number of specialized cells that circulate in our blood and reside in other tissues. These cells act as sentinels and are responsible for our ability to fight infections by recognizing specific proteins from invading microbes (bacteria or viruses) as foreign and potentially harmful. When a foreign invader is detected, the immune system produces “antibodies”, proteins that selectively bind to these foreign proteins and target the microbes for destruction. At the same time, other cells are deployed to hunt down the microbes, wherever they are, and destroy them. These cells are also involved in healing wounds and response to other types of tissue injury. This process is called “inflammation”.
All the demyelinating diseases listed above involve inflammation of the CNS. MS is further complicated by the fact that it is also an autoimmune disease. In autoimmune diseases like MS and systemic lupus erythematosus (SLE or lupus), the immune system malfunctions and wrongly thinks certain normal proteins produced in healthy tissues are foreign proteins and they become the target.
In MS, the immune system attacks proteins contained in the myelin sheath, which is the protective coating on the nerves in the CNS. Normally, the myelin sheath acts like the insulation around electrical wires and is important for the proper transmission of nerve signals from the brain to other parts of the body. In MS, anti-myelin antibodies direct a “friendly-fire” inflammatory attack on the myelin sheath, leading to demyelination. This can lead to permanent nerve damage slowing or blocking critical nerve transmissions that control muscle coordination, tactile and vision sensation, bladder function and strength. Even in the absence of anti-myelin antibodies, certain cells from the immune system will invade the brain and spinal cord and cause direct tissue damage.
Since all the demyelination diseases listed above involve changes in the immune system and the inflammatory response is a major driver of the disease process, the information we generate studying MS should help elucidate the underlying causes of TM and other demyelinating diseases. Furthermore, this information will provide the framework for the development of diagnostic tests for all these diseases.
Diagnosis of Multiple Sclerosis
MS is a debilitating autoimmune disease that attacks the central nervous system causing demyelination of nerves. There are 25,000 to 30,000 new cases of MS diagnosed each year in the United States while estimates of patients presenting with clinical symptoms that could be MS are 5 to 10 times that number per annum. At present, effective treatment of this disease is hampered by the lack of “clinical assays” (diagnostic tests) capable of providing actionable information about diagnosis, prognosis, disease segmentation and response to treatment. Unfortunately, the diagnosis of MS is somewhat subjective, based on the experience of the practitioner and the severity of disease at the time of presentation. Many patients suffer for months, if not years, before they get a definitive diagnosis.
Figure 1 illustrates the current process used to diagnose patients presenting with symptoms of MS.
Figure 1: Current MS Diagnostic Process
Patients with symptoms of MS are often subjected to a battery of tests that must both support the diagnosis of MS and exclude common mimics of the disease. These tests are expensive, require multiple visits to the clinic and some are invasive and painful, such as spinal taps for CSF testing. Patients with early signs of demyelination undergo this sequence of testing in order to stratify their risk of having MS. The risk of MS is usually expressed in rough percentages ranging between 20 and 80% depending on symptoms and MRI results. Worse yet, the outcome of this diagnostic workup results in a significant healthcare dilemma: given the benefits of treating MS at an early stage (Kappos et al., 2007), patients must decide whether they will start costly, invasive treatments even when uncertain about the diagnosis.
Patients must elect to either:
Why is MS hard to diagnose and treat?
Here are a few issues that make MS (and other demyelination diseases) so difficult to diagnose and treat:
Symptoms commonly shared among these diseases include blurred or double vision, slurred speech, muscle weakness and fatigue, partial paralysis, numbness, unexplained pain, incontinence, etc.
These facts present serious challenges for both patients and their physicians. Thus, there is a critical need for better methods to diagnose, prognose, and monitor patients with MS and other demyelination diseases.
Most existing diagnostic tests measure only one biomarker. As used here, a biomarker is any biological analyte found in the blood, other body fluid, or tissue of a patient that can be objectively measured and correlates with specific biological processes. These simple, single-analyte tests tend to be inaccurate and are seldom definitive. Good examples are the PSA test for prostate cancer and anti-nuclear antibody (ANA) tests for autoimmune disease. The current PSA test has a very low specificity. Only about 25% of individuals with a positive PSA test actually have prostate cancer based on biopsy results and about 15% of individuals with a negative test do have cancer (Paul et al., 2005). Almost everyone has some ANA in their blood. Of those with high levels of ANA, only about half actually have some type of autoimmune disease, such as SLE, Sjogren’s syndrome or rheumatoid arthritis. High ANA levels can also result from viral infections; certain liver, lung, intestinal and skin diseases; and can even result from taking certain drugs, including some blood pressure and anti-convulsant drugs. Thus, the ANA test alone is not an accurate diagnostic test for autoimmune disease.
An emerging trend in human healthcare is to use panels of biomarkers (typically tens of markers) at the same time. Individual biomarkers in these panels are often associated with distinct biological processes involved in the disease. In these types of assays, subtle but characteristic changes in the pattern of biomarker expression are more important than changes in any single biomarker. Recent scientific advances in the fields of genomics (the identification of all genes in an organism’s genome and their function) and proteomics (the identification of all proteins that are encoded in an organism’s genome and their function) now make it possible to development next-generation “molecular diagnostic” (“MDx”) tests, as discussed in more detail below.
Another emerging healthcare trend involves pairing specific diagnostic tests with specific therapies to help insure that patients receive the drug(s) most likely to provide relief. This approach is often referred to as “companion diagnostics” or “theranostics” (therapy plus diagnostics).
Together, these trends make possible the dream of “personalized medicine”, also referred to as “individualized medicine” or “molecular medicine”. According to the Personalized Medicine Coalition (www.personalizedmedicinecoalition.org); “By employing new methods of molecular analysis to better manage a patient’s disease or predisposition towards a disease, personalized medicine aims to achieve optimal medical outcomes by helping physicians and patients choose the disease management approaches likely to work best in the context of a patient’s genetic and environmental profile”. As a result, healthcare professionals are now gaining access to new diagnostic tools to better deal with hard-to-diagnose and hard-to-treat diseases like MS and TM. Exemplifying the attention personalized medicine is getting, Senator Barack Obama (Illinois) introduced a bill titled “To improve access to and appropriate utilization of valid, reliable and accurate molecular genetic tests by all populations thus helping to secure the promise of personalized medicine for all Americans” (Bill S.3822, 109th Congress 2nd Session).
How will Personalized Medicine change how MS is diagnosed and treated?
It is generally accepted that MS results from the interplay of both environmental and genetic factors. Environmental factors include such things as 1) exposure to toxins that are ingested, inhaled or absorbed through the skin; 2) infectious disease agents such as bacteria and viruses; 3) radiation exposure; and 4) stress, etc. Genetic factors include both changes at the DNA or gene level, including gene mutations and the duplication or deletion of gene sequences; as well as the abnormal expression of genes (mutated or normal) involved in many complex biological processes such as DNA repair, cell replication, cell metabolism, inflammation and immune response. Anything that disrupts the normal functioning of one or more of these biological processes can lead to disease.
The human genome – the genetic blueprint – contains an estimated 20,000 to 30,000 individual genes. Genes contain the information needed to make proteins, which are the main functional constituents of all cells. Proteins act as enzymes, hormones, signaling molecules and antibodies required for the proper function of each cell and, in turn, each organ in the body. Although we used to think that each gene produced a single protein with one function, we now know that most genes actually can produce multiple versions of a protein and each version can have a different function. In this manner, the genome can actually produce hundreds of thousands of different proteins. Each of the roughly 50 trillion – yes trillion!! – cells in your body contains virtually all the same genes. So, what makes a liver cell different from a kidney cell or a muscle cell, brain cell, blood cell, etc. is determined by which genes are turned on (“expressed”) and to what level they are turned on in each type of cell. For example, red blood cells need to express the gene that makes hemoglobin, the protein that carries oxygen throughout the body. Other cells do not need to make hemoglobin to function so they do not turn this gene on. B cells, which are the cells in the immune system responsible for making antibodies to fight infections, express the genes that code for the various bacteria- and virus-specific antibody proteins – other cells do not.
What determines which genes are expressed in a given cell? One can think of the regulation of “gene expression” like having 20,000 to 30,000 molecular rheostats, similar to rheostats that might control lights or fans in your home. Each gene can be turned off but, if it is turned on, it can be continuously adjusted from low to high. It also can be turned up or down at any time. Each gene in the genome is controlled by a molecular rheostat. The sum total of the expression of all genes in a cell represents that cell’s “whole-genome expression profile” and determines which proteins are produced by the cell and, therefore, the function of that cell. This also determines whether that cell functions normally or abnormally.
Furthermore, individual proteins do not work in isolation, but rather they interact with other proteins in complex biological networks similar to intricate electrical circuits in computers. The complex system of biological networks must be carefully controlled for each cell, organ, and ultimately the entire body to operate normally. Damage to key genes or the improper regulation of key genes involved in these biological networks can result in the development of serious disease. When these changes can be reliably measured and shown to be specifically associated with a disease like MS, they may be useful as a “biomarker” in a diagnostic test.
In general, each biomarker will be associated with a specific biological network. Thus, a given biomarker will only be informative as a diagnostic biomarker if this biological network is affected in that patient. It will not be informative if the patient’s disease is a result of changes in other biological networks. Thus, the best way to diagnose these complex diseases is to measure multiple biomarkers, each associated with different biological systems that are known to be involved in the disease in question.
MS is difficult to diagnose because it can be caused by aberrant regulation of multiple biological networks. What is diagnosed as the same disease in different patients can involve different biological networks. As mentioned earlier, MS is actually a class of related diseases comprised of various disease subtypes. In addition, other diseases can look like MS (“MS mimics”) including other demyelinating diseases, such as TM, NMO, ON and ADEM. Unfortunately, the best diagnostic tests currently available are unable to readily differentiate MS from these MS mimics or differentiate MS subtypes. This is particularly true during the early stages of disease progression. Thus, patients with different diseases or disease subtypes are often diagnosed as having the same disease and treated the same way when, in fact, they have very different diseases and require different treatments.
All these factors seriously complicate the diagnosis and proper treatment of these diseases since patients with the various subtypes of MS and the confounding diseases respond to different treatment approaches – Proper treatment depends on accurate diagnosis.
A number of different drugs are available for the treatment of MS patients. Although there are fairly good drugs available for patients with RRMS, the drugs available for treating PPMS and SPMS are not as effective and often involve the use of chemotherapy drugs, which were developed to treat cancer patients.
In general, response rates for many commonly prescribed drugs range from as low as 25% to 80% (Spear et al., 2001). This means that, in some cases, up to 75% of patients given a specific drug will not benefit from the treatment. Many MS patients respond well to current therapies. Unfortunately, it is estimated that up to 50% of MS patients on Interferon-beta therapy will continue to experience significant relapses and disease progression leading to severe disability (Byun et al., 2008). This highlights the added need for new MDx tests to predict which drug is most likely to work for each patient and to check whether the drug is working after therapy starts.
Thus, as suggested above, there is a critical unmet need for better methods to diagnose, prognose, and monitor patients with these complex diseases.
DioGenix is an early-stage diagnostics company that has developed a novel strategy for the development of next-generation MDx assays using panels of biomarkers to improve the diagnosis of difficult-to-diagnose diseases. DioGenix is currently focused on developing MDx assays for neurologically-based autoimmune and inflammatory diseases. We are currently focused on developing MDx assays for MS.
DioGenix uses state-of-the-art genomics technologies and sophisticated biostatistics to quickly identify and validate novel panels of genomic biomarkers that represent “gene signatures”. A gene signature is a panel of genomic biomarkers whose pattern of gene expression correlates with disease status. They can be used to 1) provide disease diagnosis and prognosis; 2) predict disease progression and regression; and 3) predict and monitor a patient’s response to therapy. Optimally, the biomarkers included in a gene signature will be associated with multiple biological networks that are affected in the disease being studied. One can think of a gene signature as a unique molecular fingerprint. Since genomic biomarkers measure changes in biological processes at the molecular level, they can more accurately identify and differentiate similar diseases and disease subtypes, even when patients display very similar clinical symptoms. Gene signatures form the groundwork for the development of critically needed MDx assays and improve patient management for difficult-to-manage diseases like MS and TM. They also may be used as key components in future theranostic applications to facilitate the delivery of the right drug to individual patients, based on the individual patient’s own genomic profile.
DioGenix evolved out of Gene Logic, a leading Genomics Service company with more than 10 years experience generating high-quality genomics data and building comprehensive genomics databases including BioExpress®. BioExpress® is a comprehensive database of human genomic and clinical information. It contains “whole-genome” expression profiles for more than 12,000 clinical samples covering more than 400 different disease types. DioGenix maintains a close working relationship with Gene Logic with preferred access to BioExpress®, in addition to their extensive clinical network, biorepository and genomics data production lab.
DioGenix is leveraging this relationship and has established new relationships with clinical experts and prominent organizations in the MS and TM research communities, including Dr. Benjamin Greenberg, Director of the Johns Hopkins Encephalitis Center and Co-Director of the Johns Hopkins Transverse Myelitis Center; The Accelerated Cure Project for Multiple Sclerosis (ACP) and The Transverse Myelitis Association (TMA).
How is DioGenix developing Next-generation MDx assays for MS?
Gene Signature Discovery is typically performed using DNA microarrays. DNA microarrays can measure the expression of thousands of genes in a single assay and are used for large-scale gene expression studies capable of determining the gene expression profiles of virtually all known human genes simultaneously.
Gene Signature Validation involves the following steps:
Clinical gene expression analysis for MDx testing requires the analysis of a small number of genes (10’s to 100’s) compared to Gene Signature Discovery. This is partially due to the high cost of running high-density microarrays in a clinical setting and economic pressure to keep healthcare costs low. Therefore, a critical factor in developing a genomics-based MDx assay is to identify the smallest number of biomarkers possible that provide the requisite clinical utility.
A critically important aspect of this entire process, particularly during the early phases of Signature Discovery and Validation is gaining access to large numbers of well-characterized patient and control samples. Although Gene Logic’s extensive biorepository contains more than 45,000 human and animal tissue samples, each with detailed clinical and experimental study information, this is insufficient for the development of commercial MDx assays requiring FDA approval. This repository was created to include a broad range of tissue types and diseases. As such, it does not have a comprehensive collection of samples for each disease. Even though it contains more than 3,500 samples from patients with immunological diseases, there are a relatively small number of MS samples, most of which are represented in BioExpress®. To supplement this repository, we have established a strategic collaboration with The Accelerated Cure Project. This gives us access to their extensive repository of biological samples from patients with MS and other demyelinating diseases.
The Accelerated Cure Project is a nonprofit organization with the stated mission of “curing MS by determining its causes”. They have created a comprehensive biorepository of blood samples and clinical data from patients with MS; other demyelination diseases including TM, NMO, ON and ADEM; and matched controls. They make these samples available to researchers investigating the causes of MS. Although they are focused on aiding research to understand the causes of MS, they recognize the importance of these other diseases to the study of MS. For more information on the Accelerated Cure Project go to their website – www.acceleratedcure.com – and see the article by Jana Goins in the Spring 2007 issue of The Transverse Myelitis Association Newsletter (Volume 7 Issue 2, page 21). Jana is the Study Coordinator for ACP at Johns Hopkins School of Medicine.
As of February 14, 2008, the ACP biorepository contained blood samples from a total of 970 subjects including:
Molecular Diagnostic Assay for Multiple Sclerosis
Figure 2 indicates where new MDx assays are most likely to make a significant difference in the current diagnostic process to improve healthcare for MS patients.
Figure 2: Points of intervention in Current MS Diagnostic Process where MDx assays are most likely to change how patients are treated.
A simple, highly-accurate, blood-based MS Risk Assessment Assay (assay in figure 2) would significantly reduce the number of spinal taps for CSF tests and MRIs for patients presenting with symptoms of MS. Likewise, a highly-sensitive blood test capable of confirming the diagnosis of MS and differentially diagnosing patients with MS versus patients with related demyelinating diseases, such as TM, NMO, ON, and ADEM (assay in figure 2) would reduce the costs and uncertainty associated with this complicated disease.
MDx assays also are needed to both predict and monitor therapeutic efficacy (assay in figure 2). As mentioned above, up to 50% of MS patients on Interferon-beta therapy will continue to experience significant relapses and worsening disability (Byun et al., 2008). The availability of MDx assays capable of predicting which patient will respond to a given drug will help ensure that each patient gets the most appropriate therapy available. Up to 40% of patients on Interferon-beta therapy will develop interferon resistance and no longer respond to this drug. This type of assay, along with an assay that can assess how well a patient is responding to therapy, would provide tremendous benefit to MS patients and forever change how MS patients are treated.
Initial DioGenix MS Study Results
We hypothesized that some number of genes should be differentially expressed in the blood of patients with MS compared to patients with inflammatory demyelination diseases that are difficult to differentiate from MS and healthy individuals. Furthermore, these dysregulated genes would represent the initial set of biomarkers for the development of a blood-based MDx assay that would provide a definitive diagnosis of MS for patients in the early stages of disease progression.
Statistical analysis of data contained in the BioExpress® database identified an initial gene signature comprised of more than 250 genes that are differentially expressed between MS patients, non-MS patients and controls. This preliminary study involved analysis of whole-genome expression data from 19 blood samples including 11 confirmed MS patients and 8 healthy control subjects accrued from a single clinical site (Figure 3).
Figure 3: Initial gene signature differentiates between blood samples collected from MS patients, lupus patients and healthy controls
Figure 3 presents the data from this experiment in graphical form using Principal Components Analysis (PCA). PCA is a mathematical method used to visualize the potential usefulness of a particular gene signature. Tight, well-separated clusters of related samples indicate that the initial gene signature may have clinical utility. Figure 3 demonstrates that one of our initial gene signatures can identify all 11 MS samples within a group of 23 samples that also includes 8 healthy controls and 4 patients with lupus.
Using more rigorous statistical analyses we were able to derive a number of smaller gene signatures with varying assay performance characteristics. These results give confirmation of our original hypothesis and provide us with a large set of candidate diagnostic markers to work with.
To further assess the utility of these biomarkers in a clinical assay, we challenged our gene signatures using additional data residing in the BioExpress® database. We added in data from additional blood samples that were collected from 24 independent MS patients and 2 asthma patients (Figure 4).
We were still able to clearly identify all 35 MS samples. Interesting, the 2 asthma samples clustered with the control samples.
Realizing that a commercial diagnostic assay would require use of a smaller set of genes, an attempt was made to develop a “minimal” MS gene signature. We compared the ability of multiple sets of genes (two or more) from the original MS gene signature to properly identify MS samples in a mix of MS, lupus, asthma and control samples. We found that as few as two genes from the original gene signature were able to accurately identify virtually all 35 MS samples (Figure 5).
Again, the two asthma samples clustered with the controls. The 4 lupus samples were clearly distinct from the asthma and control samples but clustered closer to the MS samples. In fact, one lupus sample overlapped with one of the MS samples at the edge of the MS cluster. This is not altogether surprising since MS and SLE are both autoimmune-based diseases.
Although lupus is not really a confounding disease for MS in the clinic, they are related since both are autoimmune diseases. Furthermore, patients with SLE and other autoimmune diseases can display early symptoms similar to MS – and TM. A good discussion of the diagnostic challenges these diseases pose can be found in the article by Julius Birnbaum in the Spring, 2007 issue of The Transverse Myelitis Association Newsletter (Volume 7 Issue 2, page 18). The data presented here demonstrates the potential of this approach.
Asthma is an inflammatory disease of the respiratory system and does not affect the central nervous system. Taken together this data suggests that our signature does not detect patients with miscellaneous autoimmune and inflammatory diseases. This signature appears to be specific for MS, although we do not have access to enough samples from patients with related diseases such as TM, NMO, ON, and ADEM to test whether this signature or one of the smaller signatures derived from this signature is capable of differentiating MS from these other diseases. This is one of the challenges we now face. We need access to blood samples from large numbers of patients with each of these diseases to test this hypothesis. This is critically important for us to determine the sensitivity and specificity of our signatures and complete the clinical validation phase of product development. We hope that we will be able to get these samples from the ACP biorepository in the near future.
In summary, MS is a complex disease that results from the interplay of both environmental and genetic factors. It is clear that there is a critical need for improved methods to diagnose, prognose, and monitor patients with MS and other demyelination diseases. DioGenix is developing a series of next-generation molecular diagnostic assays that will facilitate 1) the early diagnosis of MS; 2) differentiate MS from other demyelination diseases; 3) predict which patients are most likely to respond to specific drugs; and 4) monitor drug response and resistance in patients receiving therapy.
As discussed above, MS shares many common features with other demyelinating diseases, including TM, NMO, ON and ADEM. For example, inflammation plays a critical role in causing nerve damage, a hallmark of all these diseases. As such, the information we generate studying MS should help elucidate the underlying causes of TM and other demyelinating diseases. Furthermore, this information will provide the framework for the development of diagnostic tests for all these diseases.
However, to accomplish these goals we need access to hundreds of patients diagnosed with MS and patients with TM, NMO, ON and ADEM, as well as healthy controls. We encourage patients with these diseases and family members to get involved and donate blood to The Accelerated Cure Project for MS. This will benefit efforts by DioGenix and others to develop both molecular diagnostic assays and improved therapies for patients with these devastating diseases. Hopefully, some day in the near future, patients presenting with TM, ON or ADEM will be able to undergo a simple blood test and find out whether or not they have experienced a first attack of MS or a one time only demyelinating event. This information with be critically important for deciding on the best short and long term treatment strategies. This can only happen with your help.
Brody, J.E., From Multiple Sclerosis, a Multiplicity of Challenges, The New York Times, March 4, 2008.