Building the Biomarker Panel in Neurology: Glial Fibrillary Acidic Protein

Since the turn of the century, the paradigm of care in neurology has changed drastically. Many patient populations with diseases and disorders—both rare and more common—that previously had few or no treatment options available, or little in terms of effective therapies, have seen first-time FDA approvals in that span. These include complex diseases such as neuromyelitis optica spectrum disorder (NMOSD), multiple sclerosis (MS), Parkinson disease (PD), Alzheimer disease (AD), spinal muscular atrophy (SMA), and amyotrophic lateral sclerosis, among others.

Although, the introduction of therapies into the clinical care landscape has not solved every problem faced by physicians and investigators involved in the process. In MS, for example, there have been more than 15 disease-modifying therapies (DMTs) approved since 1990, completely transforming the lives of patients. But the accurate measurement of disease progression in MS has remained a mountainous obstacle, ultimately proving to be a roadblock to better optimizing treatment for millions of individuals.

The in-depth measurement of various aspects of disease in research investigations can provide investigators a more granular view of often complicated pathologies, unlocking information about how well targeted treatments are performing and how disease process unfolds in the long-term. Altogether, biologic markers can be used to guide clinical diagnosis, estimate disease risk or prognosis, and evaluate disease stage. NMOSD is perhaps among the greatest examples of this utility—a key biomarker of the disease, aquaporin-4 (AQP4), was discovered in 2004, unlocking an important piece of knowledge for disease understanding. In part spring-boarded by this discovery, within 15 years, the first therapeutic for the autoimmune disease, eculizumab (Soliris; Alexion Pharmaceuticals) was approved by the US Food and Drug Administration (FDA), opening the door to treatment for thousands of patients who previously had only off-label options.

Across neurology, the pace of discovery for biomarkers has been rapid, with technologic advances in assay and instrument precision and sensitivity, as well as idea sharing from other areas of medicine, taking the possibilities to new heights. Although, despite these vast numbers of candidate biomarkers being proposed and rigorously evaluated, very few are currently integrated into routine clinical practice and the pursuit of novel markers of central nervous system (CNS) injury markers is an ongoing one.

Glial Cells

For certain CNS and neurodegenerative disorders, such as MS and NMOSD, some discussions of therapeutic targets and goals have centered around glial cells: astrocytes, oligodendrocytes, and microglial cells. These cells are responsible for a variety of tasks, including maintaining an environment for neuronal signaling (astrocytes), providing myelin wrapping to neurons in the CNS (oligodendrocytes) and periphery (Schwann cells), and the removal of cellular debris (microglial cells).

Uniquely within astrocytes, Schwann cells, and enteric glial cells, an intermediate filament III protein called glial fibrillary acidic protein (GFAP) is found. Coupled with the implications of astrocytes’ role in neuropathologies,9 GFAP has emerged in neurology as a key biomarker candidate. GFAP gene activation and protein induction have been shown in the literature to play a role in astrogliosis post-CNS injury and within neurodegenerative processes, including the rapid release of GFAP after spinal cord injury and stroke, and multiple published papers have suggested that GFAP is in some way implicated in the pathology of SMA and AD. Some, such as Gschmack et al, have also recently hypothesized that disturbances in blood-brain barrier integrity could lead to an access of GFAP autoantibodies to areas destroyed early in the pathogenesis of PD, such as in the dorsal motor nucleus, pointing to a potential pathological function in the development of PD.

GFAP in Multiple Sclerosis

In MS, particularly, much promise has been displayed in the clinical utility of GFAP as a biomarker. Neurofilament light chain (NfL) is a well-established marker of neuroaxonal damage, but data on astroglial markers in serum have been lacking for MS, and recent investigations have suggested that GFAP may offer added value in that contex. As well, data have suggested that serum GFAP may correlate with radiographic plaque burden, presenting potential to supplement measures of subclinical disease activity.

“GFAP is a promising biomarker for progression, [among others], and we are now working further on that by looking also into MRI metrics and thinking actively about how to validate [our] findings in a separate cohort, which is obviously a very important step to prove any usefulness or towards clinical application,” Jens Kuhle, MD, PhD, head of the MS Center and a senior physician at University Hospital Basel, told NeurologyLive® about his work assessing a custom immunoassay panel that measures the concentration of 20 proteins could classify healthy patients from MS subtypes, clinically isolated syndrome (CIS) from other MS phenotypes, as well as CIS and relapsing MS from progressive MS. Kuhle has conducted a variety of other research efforts into biomarkers for MS, including research from the EXPAND study (NCT01665144), which suggest that lowering plasma GFAP early leads to reduced risk of disability worsening in those with secondary progressive MS treated with siponimod (Mayzent; Novartis).

“The development of new clinically meaningful biomarkers in MS has been challenging,” Bruce Cree, MD, PhD, MAS, FAAN, clinical research director of the University of California, San Francisco Multiple Sclerosis Center, told NeurologyLive®. “The most useful biomarker for diagnosing MS remains oligoclonal bands and elevated IgG Index from cerebrospinal fluid. Other biomarkers, such as serum anti-JCV antibodies are useful for stratifying those at risk for progressive multifocal leukoencephalopathy if treated with natalizumab [Tysabri; Biogen]. Serum anti-AQP4 antibodies are very useful for distinguishing MS from NMOSD. Lastly, lymphocyte subset panels that include CD19 cell counts are useful for confirming B cell reconstitution in patients undergoing planned dose suspension from anti-CD20 monoclonal antibody treatments including ocrelizumab [Ocrevus; Genentech], ofatumumab [Kesimpta; Novartis], and rituximab.”

Cree explained that serum NfL remains under investigation, as although it does correlate with recent disease activity, there is a 9-week lag time in its peak concentration following MRI disease activity. “Whether serum NfL could be useful in monitoring patients, especially those with progressive forms of MS, is not well established. Although, there are certainly some strong proponents for its use. However, not all studies with longitudinally collected samples show consistent correlations with disease progression,” he said.

This challenge has, in some ways, opened the door for further research into the utility of GFAP, particularly as the field’s need for markers of disease progression has grown. Kuhle, for example, presented data at the 38th Congress of the European Committee for Treatment and Research in Multiple Sclerosis, suggesting that GFAP is a more useful biomarker for MS progression than NfL, with NfL increases disappearing after additional GFAP adjustment.

GFAP in NMOSD

In NMOSD, elevated serum GFAP appears to offer prediction of subsequent attack risk and increases during acute attacks. Multiple studies have shown that the marker can show evidence of disease severity and treatment efficacy, as well. Additionally, as AQP4 can be useful in differentiating NMOSD from MS, a recent analysis comparing the differences of double seronegative NMOSD and AQP4-positive NMOSD showed that double-seronegative patients had significantly lowered GFAP protein levels, implying an underlying pathophysiological distinction between the 2 conditions.

GFAP has shown additional utility for NMOSD in that although serum NfL levels do not differ between NMOSD, MS, or myelin oligodendrocyte glycoprotein antibody disease (MOGAD), a study from Watanabe et al suggest that the serum GFAP/NfL quotient at attack state could be a useful differentiator of NMOSD from MS with a sensitivity of 73.0% and a specificity of 75.8%, as well as a differentiator of AQP4-Ab-seropositive NMOSD from MOGAD and MS.

“The association with acute attacks is robust and has the potential to be clinically useful for diagnosing acute attacks if a point of care test were to become available,” Cree, who has also served as a clinical investigator in the development of multiple NMOSD therapies, said. Although, limited evidence is available to date for this application of GFAP in NMOSD, with only the N-MOmentum study (NCT02200770) of inebilizumab (Uplizna; Horizon Therapeutics) reporting either MRI or serum biomarker data. As a result, the development of a biomarker panel of multiple markers has been considered as a possible solution to add to the understanding of NMSOD and MS activity, severity, and progression. “However, the available data on such approaches are at best preliminary and, to some extent, disappointing diagnostically,” Cree said.

Serum and Blood GFAP

Evidence has emerged indicating that blood GFAP levels hold the serum’s potential to reflect, and possibly predict, a worsening of disability in individuals with progressive MS, and literature has been continually published in recent years suggesting that blood GFAP levels can be even used to detect subtle injury to the CNS. “Although criteria for determining a MS diagnosis have improved over the years, a clinically impactful blood test that can cleanly distinguish MS from other neurological diseases has yet to be developed. I think this is rather surprising given the extensive research thus far devoted to understanding the genetics and immunology of MS,” Cree said.

This lack of development may be in part because of technologic limitations. In those with demyelinating diseases such as from NMOSD and MS, the median cerebrospinal fluid GFAP level is 8601 pg/mL and the median serum GFAP level is 167 pg/mL, but conventional enzyme-linked immunosorbent assay (ELISA) measurements typically assesses proteins at concentrations above 10⁻¹² M. Although such clinical and technical challenges still need to be resolved for its full potential to be realized, released GFAP does reach the blood through an impaired blood-brain barrier and/or glymphatic efflux, and the advance of ultrasensitive single-molecule array (SIMOA) technology has expedited the realization of such potential.

The dynamics of GFAP after release will still need to be explored to determine accurate blood GFAP half-life, as does more information need to be elucidated on the mechanisms and pathways of GFAP that is released from brain into the blood. As well, the intervals for testing blood GFAP levels and guidelines for biomarker-based decision must be established for such tests to become regular in practice. That said, blood testing has been discussed as a preferred method for long-term biomarker measurement tools, particularly because of the ease of collection compared with lumbar puncture, which is required for the current measurement of many biomarkers. If long-term and consistent measurement is to be undertaken for more precise prognostication and assessment of therapeutic efficacy, blood may be among the favored routes to take, at the least in combination with other methods.

“As far as multiple testing is concerned, having a cerebrospinal fluid biomarker is not something that will be very easy to translate into clinical practice because we don't do serial lumbar punctures in clinical practice,” Claire Bridel, MD, PhD, a neurologist in the Department of Clinical Neurosciences at Geneva University Hospital, told NeurologyLive®. “A biomarker that that is identified in cerebrospinal fluid can be a candidate that can then be worked on to develop tests in blood. Blood is much easier to draw and to analyze. The disadvantage of blood is that it's further away from the CNS, and there are a lot of other things that modulate it and can potentially modulate biomarker levels in the blood compared to cerebrospinal fluid. So, these are really 2 complementary approaches. But we cannot reasonably ask a patient to have to ask to have multiple lumbar punctures throughout his or her disease and time because it's just too invasive.”

Although, Cree explained, serum GFAP also has the potential to be assessed at bedside, and “the development of similar bedside kits for rapid testing could be highly advantageous when there is diagnostic uncertainty about ongoing disease activity.” Multiple studies in spinal cord injury and traumatic brain injury have suggested that GFAP has utility and that a rapid test, done with blood or serum, would be clinically useful.

GFAP in Alzheimer Disease

In addition to MS and NMOSD, recently published data have suggested that geometric GFAP concentration is significantly increased in carriers of familial AD relative to noncarriers, with levels beginning to differ at presymptomatic stages of disease, roughly 16 years prior to the estimated symptom onset. As a result, senior investigator Nick Fox, MD, the director of the Dementia Research Center at University College London, and colleagues, concluded that the data "support plasma GFAP being a biomarker of early AD pathology."

Fox et al also noted that these findings are consistent with other recent studies that showed GFAP increases being associated with subsequent decline in global cognition, amyloid accumulation, and conversion to dementia. In 2021, findings published in JAMA Neurology highlighted the value of plasma GFAP as a sensitive biomarker throughout the AD continuum, from preclinical AD to AD dementia. Using a SIMOA assay, plasma GFAP levels were found to be significantly higher in individuals with preclinical AD in comparison with cognitive unimpaired amyloid-ß-positive individuals and were higher among individuals in symptomatic stages of the AD continuum. Not only were plasma GFAP magnitude changes consistently higher than those of cerebrospinal fluid GFAP, but plasma GFAP also more accurately discriminated Aß-negative patients compared with cerebrospinal fluid GFAP.

In a similar effort to better distinguish dementia with Lewy bodies (DLB) from AD, researchers at Cleveland Clinic assessed the differences in cerebrospinal fluid levels of both GFAP and soluble triggering receptor expressed on myeloid cells 2 (sTREM2), a myeloid cell activity marker that has been genetically linked to AD risk. Led by Lynn Bekris, PhD, a molecular biologist at the Cleveland Clinic Lerner Research Institute, the study included individuals clinically diagnosed with AD, DLB, mild cognitive impairment (MCI), Parkinson disease (PD), and cognitively normal controls. The P-value for α-synuclein vs GFAP in entirety was also significant, driven by the significant P-value for the DLB patient group.

“Discerning dementia with Lewy bodies from Alzheimer's disease can be a challenge for the clinician in some cases and represents a critical unmet medical need,” Bekris told NeurologyLive®. “For some of our previous findings, we suspect that biomarkers of inflammation differ by different pathologies, so by amyloid and tau, and then others are also looking at alpha synuclein. Since both GFAP and sTREM2 are elevated in AD, and ADRD, but little is known about the relationship between them. We thought that this might tell us whether a particular disease group might have more astrocyte versus microglia pathology.”

Bekris explained that their data showed that the sTREM2/GFAP ratio was observed as lower in the MCI stage, and GFAP was significantly higher in the MCI stage compared with the other groups assessed, likely reflecting the lower sTREM2 relative to GFAP. She added that elevated GFAP levels have been observed by other researchers in multiple neurodegenerative diseases, but her group was surprised by the significance of it in their data. Its importance, though, still lies in its relationship with the role of microglia in AD pathologies.

“We have observed in autopsy studies that there's astrogliosis in these disorders. We also know now that microglia are very important,” Bekris said. “There are kind of 2 camps where people are studying astrocytes and people are studying microglia, where is it we have markers that are reflecting those different activities—for example, GFAP and sTREM2. We could perhaps use a ratio to understand, in a more precise way, which individuals are exhibiting a more astrocytic response versus a myeloid response. That's why we wanted to look at the ratio, to see if there was a significant difference between these groups. And it looks like there is.”

A Full Panel

Ultimately, although GFAP has emerged as a topic of consistent interest with regard to the information it may provide to those treated chronic neurologic and neurodegenerative disease, like all markers of biologic activity, it remains limited alone. At the end of the day, diseases such as MS, NMOSD, and AD are complex, and an accurate and detailed biomarker panel to measure their downstream and long-term effects will likely require more than just a few markers. GFAP measurements alone—in cerebrospinal fluid or blood—will not uncover all the information required.

“We will probably need multiple biomarkers for every single patient to capture all the different dimensions of MS because in remitting MS, there's already progression happening, it's just very subclinical. It’s very difficult to identify, but it's already there,” Bridel said. “If we have a very sensitive biomarker, a biomarker of progression may be useful in relapsing remitting patients. And the opposite is also correct, in my view, in progressive patients—there can be residual acute inflammation. I don't think we will have biomarkers for subtypes of MS, but rather, biomarkers of biological processes will be which will be more or less important in one stage of the disease versus the other.”

In addition to the limitations posed by the complicities of neurologic diseases, the utility of any biomarker utility lies in how it can be used in clinical settings alongside other measures, biomarker or otherwise. Although ratios of GFAP and other markers appear to have some potential to inform clinicians, even in a complex panel that includes myriad biomarkers are still just one piece of a multitool approach to disease measurement. Imaging techniques and clinical exam scales remain other large pieces of the puzzle.

“It remains unclear whether serum GFAP will add value to what can be determined radiographically,” Cree explained. “This is especially true for patients on highly efficacious medications, such as anti-CD20 monoclonal antibodies, the reason being that these treatments are so effective that new radiographic disease activity is almost completely silenced with treatment. Given that nearly every MS patient undergoes MRI at the time of diagnosis, and often during treatment, how serum GFAP will add to the neuroanatomical data remains to be determined. Markers with high specificity and sensitivity for insidiously progressive disease and disability worsening remain elusive.”