Can USP, NIST, & NIIMBL Conquer The Biggest AAV Quality Challenge?

By Anna Rose Welch, Editorial & Community Director, Advancing RNA

It has been almost a year since I first caught word of a significant new AAV measurement/quality collaboration between U.S. Pharmacopeia, National Institute of Standards and Technology (NIST), and National Institute for Innovation in Manufacturing Biopharmaceuticals (NIIMBL).

For those of you working in the AAV gene therapy space, you’ll know that a year goes by in the blink of an eye. A lot can and has certainly happened in the gene therapy space in the last 365 days — including a two-day FDA adcomm devoted to AAV safety and toxicity. But regardless of the number and breadth of discussions lavished on AAV quality today, the industry has not yet been able to cross “Answer the most daunting AAV quality-related questions” off their to-do lists (to the chagrin of analytical teams the world over).

As the CGT industry struggles to sufficiently characterize its AAV and the FDA grows increasingly inquisitive about what each vector is bringing to the party, I reached out to my good friend Fouad Atouf, VP, global biologics, science, and standards at USP. Together, Atouf and I, along with a panel of a dozen gene therapy manufacturing experts, unpacked the ins-and-outs of this ongoing USP, NIST, and NIIMBL initiative — as well as the CGT industry’s greatest needs from such a collaboration.

Of course, our conversation covered the important basics — the “why,” “what,” and “how” of this collaboration. However, what also emerged from this discussion with Atouf were the challenges and opportunities facing USP and the gene therapy industry in achieving a common quality “language,” if you will, for AAV in the years ahead. 

Vector Half Empty Or Half Full? There’s A Bigger Issue To Address First

Following the FDA’s adcomm on AAV safety and toxicity last September, there were two central takeaways for those on the CMC side of the industry. First and foremost, limits to certain quality attributes, particularly empty-full capsid ratios, were not specified for a handful of reasons — a lack of reference standards being a big one. The panel also rejected the notion that dosages should be capped as a safety measure.

Not surprisingly, these two takeaways had the industry breathing a sigh of relief. Limits in a fledgling industry like CGT can spell additional complexity in terms of cost, equipment needs, and of course, time — that under-appreciated COGS. That said, however, there were also arguments voiced that we cannot and should not be afraid of certain reasonable limits — particularly for the empty-full capsid ratio — to achieve a more knowledgeable and higher-quality, standardized future. The FDA’s Peter Marks may be urging the industry to “just pick a CQA!”; now, some in the industry are taking a page from Marks’ book and urging the FDA itself to “Just pick a ‘viral vector viability minimum’!” (In fact, Dark Horse Consulting just recently submitted a proposed draft guidance on AAV empty capsid testing for the FDA’s consideration.) 

Regardless of which stance you take — “let there be reasonable limits” or [figuratively speaking] “stay away from my analytics”— such arguments at this moment in the industry may be slightly premature. As I see it, there are two important steps the industry must ascend before it can decisively know exactly what — and how much — is and should be contained inside of a viral vector and within a dose of a gene therapy. First of all, the industry needs to know which analytical methods will be the best suited to measure each of the viral vector’s quality attributes, including empty-full capsids. And secondly, the industry must also understand how to accurately and reproducibly use these methods in conjunction with publicly available reference standards.

This is where my conversation with Atouf began. As we’ve become increasingly aware, advanced therapies demand a lot of their makers from a quality perspective because they comprise so many different critical materials.

“On the surface level, a gene therapy is a single construct comprising a critical component — that critical component being the viral vector equipped with the necessary genome sequence,” Atouf explained. “If you don’t address the quality of that viral vector, you’re not addressing the quality of the product. But for AAV in particular, the viral vector manufacturing process and the materials involved in the triple transfection step — namely the three different plasmids — are also an integral piece of the gene therapy puzzle. Quality has to go all the way back to those different components, as well.”

Though reference standards will be essential for achieving high-quality AAV gene therapies, Atouf emphasized that reference standards will not necessarily be a cure-all for the industry’s current woes. Even in cases where standards exist for evaluating the critical components of an advanced therapy, what isn’t yet known are the best tools with which to measure the quality of that vector.

The ABC’s Of The USP/NIST/NIIMBL Collaboration 

Enter the newest collaboration between USP, NIST, and NIIMBL.

“Standards-setting organizations — and USP in particular — are no stranger to collaborative studies,” Atouf shared. “USP runs studies, which are multi-level approaches assessing a single factor. In this case, we’ve partnered with NIST, which has a long history and great experience in measurement science, and with NIIMBL, which has a diverse membership across industry, academia, and government. Together, we are carrying out an interlaboratory study to improve the measurement of AAV quality attributes, with a particular emphasis on empty-full capsid ratios.”

There are two critical facts about this collaboration to note upfront. The first is simple: this collaboration is not currently working on the drafting and publication of a documentary standard. As Atouf pointed out, this collaboration’s end-goal is still upstream of the documentary and reference standard. Before we can press “print” on any standards, the end-user needs to understand the suitability of and variability between the many different analytical methods available for use today. 

Secondly, the term interlaboratory here is also important to note. It’s a tale as old as time that reproducibility of analytical testing procedures is a common pain point in the development of any therapy. Ensuring that the selected method is performed accurately and that those results can be reproduced regardless of who performs the testing, and where, is only one (big) part of the challenge. The other is figuring out just which of the many analytical technologies out there will provide the most information about the quality attributes of a novel therapeutic product.

In the advanced therapies space, the pains of method selection/application are even more acute, seeing as legacy analytical methods and the procedures for running them are less tried and true in the context of a gene therapy. The sheer number of assays (~30+!) needed to evaluate an AAV gene therapy alone is another complicating factor, thanks to AAV’s biological complexity. If you need any proof that AAV is an overachiever, look no farther than the fact that there are eleven identified AAV serotypes — and, unfortunately, the phrase “seen one, seen them all” does not apply to viral vectors.

However, Atouf emphasized that the analytical technologies — as opposed to the product and/or materials — will be the primary stars of this USP/NIST/NIIMBL collaboration. The ultimate goal is to learn how the different analytical technologies compare to each other in measuring AAV empty-full capsid ratios.

“We started by picking a few of the most commonly used analytical methods and are having researchers in different labs use those methods to analyze the same reference material [i.e., a commonly used serotype of AAV],” Atouf outlined.

The methods selected for this study include AUC, cryo-EM, SEC_MALS, UV, PCR (including dPCR, ddPCR, and qPCR), and ELISA.

“With this parallel assessment, we can benchmark how each of these methods perform, what information they each provide, and the inter-lab variabilities that arise in the process of analyzing the same reference material,” Atouf continued. “We will then compare and rank these methods and determine their most appropriate use cases.”

It goes without saying that each analytical tool has its own benefits and limitations, which can be dependent on the therapy itself, the phase of development, and/or the maturity of the manufacturing paradigm. For example, seeing as this space is currently suffering from scale-up constraints, some methods (e.g., analytical ultracentrifugation/AUC) that require large sample sizes will be ill-suited as release assays for certain development programs and/or stages of development.

Following the conclusion of the currently ongoing pilot study in Q3 of this year, the partners will launch a full-scale inter-lab study. Of course, we can look forward to future publications showcasing the results for both the pilot and full-scale studies.

When all is said and done, the hope is that this collaboration’s head-to-head comparisons will aid companies in selecting the most suitable method(s) for measuring the empty-full capsid ratio at their particular phase of development. But we shouldn’t necessarily expect measuring empty-full (or AAV analytical characterization in general) to become cookie cutter-like in the years ahead, either. As Atouf himself acknowledged, any method and reference standard will still need to be verified to suit each company’s individual process and their product’s biological nuances.

Once this collaboration reaches its conclusion at the end of 2023, Atouf expects reference standards will be a natural next step. But what these future AAV reference standards will ultimately look like is hard to predict.

“If you were to ask me if the desired outcome of this collaboration is to arrive at a single reference standard, I’d say the answer — in part — it is a possible yes,” Atouf surmised. “But I also think it could be more complex than that. We need to let the scientific work drive the outcome and we may potentially deliver a panel of reference standards for developers to use to sufficiently address all the critical quality attributes of AAV.” 

Though the quantum leap-style development of COVID-19 vaccines in 2020 was a marvel for the entire world, there is also something to be said of the underrated “baby steps” that every nascent industry has to take to lay the groundwork for standardization. Throughout our conversation Atouf kept reiterating that, to fill the most prominent CGT industry gaps, we simply have to start somewhere. As he concluded:  

“Will this collaboration be 100 percent perfect right out of the gate? No, but what is important is that we have the framework in place and the right minds tackling this necessary work.”

Stay tuned for part 2, in which Atouf and I continue to discuss the challenges of ascertaining the quality of AAV viral vectors and how these complexities are shaping USP’s thought process around and approach to establishing reference standards for AAV gene therapies.