Liquid Gold

For Jesse Berry, discovering the suitability of aqueous humor as a biopsy site for retinoblastoma, was “serendipity.” A chance decision led Berry, Associate Director of Ocular Oncology at Children’s Hospital Los Angeles, to analyze the liquid inside a patient’s eye, and find it contained crucial genetic tumor information. The hope? That the novel liquid biopsy could open the door to more accurate diagnosis, more detailed understanding of treatment response, and more reliable prognosis.

We speak to Berry to find out how she made the discovery and why she is, in her words, “forever grateful.”

How did you discover that aqueous humor made a good biopsy site?

There had previously been an agreement that no needle should enter an eye with active retinoblastoma. This is because retinoblastoma, unlike most cancers, cannot be biopsied for risk it can spread outside the eye. However, a safety-enhanced procedure for injecting chemotherapy into the vitreous was introduced in 2012.

As part of this procedure, an initial paracentesis with extraction of aqueous humor is performed to lower the overall pressure in the eye. Initially, this biofluid was discarded, but then – while actually throwing it away – there was a moment when I wondered if there was anything in this fluid. Since that moment, there has been an exciting amount of work showing that the fluid harbors DNA, RNA and miRNA which is from the tumor.

How are samples taken?

The aforementioned paracentesis involves inserting a 32 gauge needle through the cornea, and extracting some of the clear fluid in the front of the eye – fluid that will naturally replenish within one to two hours. So far, the majority of samples are taken during the injection of chemotherapy into the eye; however, given the exciting findings from this research, we can now extract or ‘biopsy’ the aqueous at the point of diagnosis and throughout therapy.

What are you hoping to find?

In general, the goal of any biopsy is to look at the tissue for diagnostic confirmation, and then look at molecular markers for prognostic information and, in some cases, treatment targets. Though there are no cells or tissues in the aqueous, we can still consider this a liquid biopsy. In the setting of a liquid biopsy for other cancers, a biofluid (often blood) is used to look for circulating tumor derived cell-free DNA that can be used for diagnosis to measure treatment response, as a biomarker of relapse and prognosis.

Mutations can also be identified for many cancers. We are looking to bring the diagnosis and prognosis of retinoblastoma to a similar molecular level by using the aqueous as our biofluid of choice for this liquid biopsy platform. Specifically, we are identifying RB1 mutations in the aqueous for diagnosis; measuring treatment response by evaluating tumor fractions in the aqueous and their response to therapy; and using cell-free DNA as a biomarker of poor prognosis by identifying gain of chromosome 6p in the aqueous.

What are the benefits and limitations of the aqueous biopsy method?

The main benefit is that – for the first time ever – we can obtain tumor DNA from the aqueous in eyes that are undergoing treatment. Previously, the only time tumor DNA was available was if the eye required enucleation because direct tissue biopsy is strictly contraindicated in retinoblastoma. Now that we can assay the DNA and other biomarkers from the aqueous in eyes undergoing therapy, we can consider using it for diagnosis and even more importantly – at least to me – prognosis.

The limitation, which was addressed in our paper, is that this is still an invasive procedure – much more so than a blood draw would be. It is arguably less invasive than a direct tumor biopsy would be, but it still involves extraction of fluid from the eye. Additionally, though we have not had any complications or issues of tumor spread, it should be stressed that using the aqueous as a liquid biopsy in retinoblastoma eyes remains experimental and should not yet be considered for clinical applications.

One of the main reasons for the blood versus aqueous study is that extraction of aqueous is more invasive, so if we could as efficiently find the DNA in the blood then we would need to reconsider our goals. In the samples taken, we did not find tumor derived DNA (with a sensitivity of 5 percent tumor fraction) in the blood. However, even if we could find tumor DNA at lower fractions in the blood, there is still a huge benefit to using aqueous: this particular biofluid is specific to each eye. About one third of retinoblastoma patients have tumors in both eyes; thus, if you found biomarkers in the blood you would not know to which eye it applied. Certainly, that question is not a big deal when it comes to diagnosis, but it’s extremely relevant when you discuss prognosis, which should be for each eye.

Can the aqueous humor predict treatment success?

We identified a novel biomarker in the aqueous which is an increased copy number of a part of chromosome 6 – “gain of 6p.”  Though this chromosomal alteration was known to occur in retinoblastoma, all previous studies on gain of 6p were from tumor tissue of eyes that had been enucleated (as there was no other way to obtain tumor DNA in vivo). However, we can now look at the DNA and the chromosome profiles from eyes actively undergoing therapy – and from eyes that have been enucleated. What we found is that if increased copies of 6p were found in the aqueous, there was a much higher risk of that eye responding poorly to therapy and ultimately requiring enucleation (2).

Could the aqueous humor be used to identify other illnesses?

Yes – and I think the list will grow. Aqueous is already “biopsied” in the setting of certain infections, and PCR can be done on the fluid to identify viruses, such as HSV. It may be possible to use the aqueous instead of direct or vitreous biopsy for other cancers or to evaluate genetic diseases in the eye in time. There is a great deal of potential here – but, aside from infections, nothing used clinically right now.

How does the aqueous humor differ from eye to eye?

Some of this detail will be in a future publication, but one fascinating discover comes from our evaluation of aqueous from children who have retinoblastoma in both eyes. These children have the same initiating retinoblastoma tumor suppressor mutation; however, the pattern of chromosomal changes in each eye is different. And that’s because the tumors (presumably) progress through tumorigenesis independently in each eye. We hypothesize that is also why children with bilateral disease sometimes respond very well to treatment in one eye and not in the other. The reason for this disparity is likely at the molecular level. Hopefully, our 6p data will show that we can prognosticate between eyes of the same child using the aqueous as well.