Skip to main content
Premium Trial:

Request an Annual Quote

ATCC Team Uses Genomics, Proteomics to Uncover Diabetes Biomarkers


Cohava Gelber
Chief scientific and technology officer
Name: Cohava Gelber
Position: Chief scientific and technology officer, ATCC [American Type Culture Collection], 2005 to present
Background: Vice president, R&D, MannKind BioPharmaceutical, 2003 to 2005; senior vice president, immunology and cell biology, Pharmaceutical Discovery, 2001 to 2003; PhD in microbiology, Weizmann Institute of Science, Israel, 1988; postdoc work in immunology, Stanford University, 1991.

At the American Diabetes Association’s annual meeting this week in Chicago, ATCC announced that a multidisciplinary team of scientists, led by Cohava Gelber, had discovered novel biomarkers linked to diabetes.
The team used an approach touching on genomics, immunology, and proteomics for its work. For instance, by analyzing gene expression levels, Gelber and her colleagues found patterns that differed depending on how far along the disease had progressed. 
Their research will be presented at the European Association for the Study of Diabetes conference in Amsterdam in September.
Below is an edited version of a conversation ProteoMonitor had with Gelber this week about the team’s findings.
Describe the proteomics work you did for this.
We took an animal model and we identified 27 different proteins, out of which six were new. The sequences were not found in the databases. We found them in a diabetes model and we have validated and qualified one of them to be found in patients with type 2 diabetes. Most likely, we will make the connection with them, but they were not previously known to be biomarkers associated with diabetes.
Of the other 21 [proteins], we found in the literature that they were previously associated with diabetes.
We used the Surface Enhancement Laser Desorption Ionization [platform] and we used two different technologies for 2D gels. One of them is separation [and] is based on isoelectric focusing, and the other [is based] on hydrophobicity.
In each case, we ran the serum from this animal model where we had two different strains. One of them is resistant, the other became diabetic after 30 days [of being treated] with a sucrose diet. And we contrasted the two of them, and selected only the plus and the minus. We did not look, like in other microarrays, for two-fold increases or three-fold increases.
Why not?
Only because we didn’t want to end up with hundreds of biomarkers that we would have to follow. Because the model is so clean, we were looking for down-regulated or up-regulated proteins [with] a clear association with disease. Probably we’re missing other [proteins], but most likely we were eliminating a lot of noise.
Do you plan to go back and find some of these proteins you may have missed?
Yes, we’re going to do a time-course [study] with the animal model. Right now, we looked [at] before and after induction of disease, so we had four different types of groups. We’re going to look at the time of induction, before they actually had overt diabetes. And then after they have been cured, you could actually take them back. So we’re going to do that.
At the same time, we have a parallel experiment where we actually looked at very unique populations of patients that had similar modes of diabetes. There’s a group that came from Yemen. They have a very lean body, yet they have type 2 diabetes, and so we were looking at them and found several biomarkers.
That [study] will be presented probably after [the European Association for the Study of Diabetes conference in September].
Doesn’t the fact that this experiment looked only at a population from Yemen limit its utility?
Clearly, there will be some biomarkers that will specific to this population, some others that will be shared. In the United States, there are several groups descended from Africa who will share some of the homology.
It sounds like you purified and sequenced the biomarkers. Have you validated them yet?
That’s what we’re doing right now.
You said that you validated one of them.
One of them was the lead marker. … We found a low-molecular-weight protein in the serum. We then found out that this protein is made in the liver, secreted into the serum, and then it appears … as a high molecular weight doublet, so it’s probably complexing with another protein. It comes as a doublet at about 60 to 70 kilodaltons, and has a smaller fragment at about 20 kilodaltons.
And we’ve seen this pattern in normals and non-diabetics — that even the doublets are changing. It inverts in normal versus the diabetic. And the 20 kiloDalton piece that is probably coming from the complex is really not seen in the diabetic or is very, very faded. It’s very strong in the normal.
Is there any way to characterize these proteins, what types they are?
On the six, we don’t know what they are. We know that the lead biomarker comes from the serpin group. They are serine protease inhibitors. This is a very important family of biomarkers. And quite of few of them have a huge distribution in different areas of development.
We excluded some potential serpins that had some homology to our proteins, so we did not find [other candidates for serpins]. We found obviously insulin signaling biomarkers, and we found quite a few biomarkers that are important in lipid metabolism. And we found quite a few biomarkers that are associated with inflammation.
Do you know if all 27 proteins need to be present for a person to develop diabetes?
Actually, we don’t think we need all of them. We performed an analysis on 62 individuals — 30 with type 2 diabetes, and 32 normals — with a cluster of five [proteins] and we found that they provide a [disease] predictive value of 93 percent.
What’s the plan for the 21 proteins that aren’t novel?
We’re going to validate them, we’re going to put together two different types of assays. One of them will be a Luminex assay. We will take antibodies to the different proteins, the different biomarkers, and we’re actually making [antibodies] for the new ones as well.
Then we will test them with the Luminex. We have the ability to go up to 100 tests in a tube. And we will screen serum samples from collections of patients and normals.
We’re trying to find an early test before the patient is exhibiting glucose disruption. We’re looking for early detection before there’s damage due to the disregulation of glucose. We’re also looking for biomarkers that are associated with complications.
At this point, we’ve got some confirmation that we indeed saw at least one of the new [proteins] in patients. We’re going to do the same for the constellation of biomarkers … using human samples. So that will be the qualification.
After qualification, we’ll perform screening.
Can you put a timetable on this?
Probably 18 months.
Then you’re going to apply for approval from the US Food and Drug Administration?
What’s the need for a prognostic for diabetes?
There’s definitely a need for early detection of diabetes. According to the statistics, there are several million [people in the US] walking around with diabetes. Even before it is overt, there is an ongoing destruction of internal organs that is initiated. And with the proper treatment we can minimize the damage.
Are there prognostic tests for diabetes already on the market?
Not that I know of. In past American Diabetes Association [meetings], they have talked about the need for biomarkers for early detection, and also as a theranostic to find out the effect of new therapeutics. With the new therapeutics, they’re talking about preserving beta cell mass, and actually making sure that the patients will not have their complete beta cells destroyed.
In animals, you can actually sacrifice the animal and measure beta cell mass or visually, by immunohistochemistry and other methods, you can look at the state of the pancreas and the beta cells and the endocrine part. But in humans, there is no marker that can show you that. All the markers are tied with the insulin production.
We hope that our test — and it will probably not be a single marker; it will be a constellation of markers — will enable early detection of disease, and hopefully select patients who may suffer complications, and then they can be treated ahead of time.
Why did you choose this path, using proteomics to develop a test?
We chose a multi-prong approach because we thought our competitive advantage will be the fact [that] in our research department we practically have all the different disciples working in concert.
We decided on disease-specific biomarker discovery as a global strategy … and diabetes was the first one. We also have cancer as a parallel program. We chose diabetes because we have quite a few scientists with interest and experience.
We have a very strong collaborator, we have a very large group in Israel from Hebrew University [led by] Itamar Raz.
They provided us with tissue samples. Of course Dr. Raz provided us with guidance in the collaboration, but the entire research, per se the proteomics and the personnel were here in ATCC.
Once we validate [the markers] and put together the assays, we’re going to do the clinical trials with the assays, and then it will all be with the investigational new drug application.
We’re also thinking about spinning off the program into probably a private company out of ATCC.

File Attachments
The Scan

Pig Organ Transplants Considered

The Wall Street Journal reports that the US Food and Drug Administration may soon allow clinical trials that involve transplanting pig organs into humans.

'Poo-Bank' Proposal

Harvard Medical School researchers suggest people should bank stool samples when they are young to transplant when they later develop age-related diseases.

Spurred to Develop Again

New Scientist reports that researchers may have uncovered why about 60 percent of in vitro fertilization embryos stop developing.

Science Papers Examine Breast Milk Cell Populations, Cerebral Cortex Cellular Diversity, Micronesia Population History

In Science this week: unique cell populations found within breast milk, 100 transcriptionally distinct cell populations uncovered in the cerebral cortex, and more.