Expert Insights: From Bench to Program Lead

Mitzi Rettinger: Welcome to Expert Insights, the DCN Diagnostics podcast, where we explore the trends, challenges and opportunities shaping the future of rapid diagnostics. I'm your host, Mitzi Rettinger. Today, I'm excited to introduce a voice that many of our clients will soon get to know well, Helen Hsieh. Helen is a Senior Scientist here at DCN and was recently promoted to Program Manager with a PhD in Chemistry and a strong foundation in biophysics. Helen has worked on everything from COVID-19 and malaria assays to maternal health diagnostics, really combining her scientific depth with a sharp eye for assay performance. Helen, welcome to the show.

Helen Hsieh: Thank you, Mitzi, I'm glad to be here and I'm looking forward to this.

Mitzi: Let's get started with your background. You've had a really interesting career academia, big industry, biotech, and now DCN. Can you walk us through how you got into diagnostics and what drew you specifically to lateral flow?

Helen: Well, I started in academia, where I was quantitating the binding between biological molecules, and when I left academia and moved into industry when I graduated, I was still quantitating the binding between biological molecules. This becomes very important in medical diagnostics, because how tight that binding is affects how good your the sensitivity of your test is going to be. And as I grew in the medical diagnostic field, one of the things I learned is that what people would really like is a point of care test. That's where you get that information right there on the patient bedside. And while there have been many, many technical advances in the field of point of care, actually the LFA, which has been around for a while now, is still the standard in a lot of cases in that it is so easy to run point of care patient bedside. In home testing, you can hand it to anyone. In fact, that's what we did recently. During the recent pandemic, we handed it to everyone. They can run this test. It is so easy to run, but that simplicity exists because the test developers have put all the work in under the hood to make it easy for everyone to run.

Mitzi: Thanks so much for sharing that story. And I agree, I think when we think back at how people thought about pregnancy tests and now people are trying to develop cancer assays all on LFA, it really does show how we've grown in this field so much. Something that I've always admired about your approach is that you're extremely grounded in material science, but also super practical. So from your experience, what are some of the most common early-stage problems that you've seen when developers first try to build a lateral flow test?

Helen: The lateral flow test infrastructure, it consists of multiple pieces of paper like materials. There's usually something called a sample pad, that's where you put the sample. A conjugate pad that's where your label is going to go in the final test. There's the membrane, which is where all the action is on the test, and the wick that helps pull all the fluids through. So when you put all of these together, this is the fluidic system of the LFA. And you want to combine these in a way that they will all work together if you're running other tests. For those of you familiar with running an ELISA test, you buy a plastic plate. Somebody has already figured out how to make the plate work. It's a plastic plate. They work. Buy the right kind. In this case, you get to tweak a lot of things to make all the fluidics work. And you want to combine them correctly. Try a couple of different variations that will work for whatever your system is, because each system is going to be a little different. The other challenge that people will have is how early do they move from a buffer system into a sample matrix? The sooner you can move into your sample matrix, the better it is for testing to make sure that your reagents work on this lateral flow system, that everything flows properly. Flow is an important component of this test and that everything is behaving. And then you can go in and optimize things later.

Mitzi: Okay. I really love that. I think, you know, it really reinforces what we try to get across to new clients, especially new clients that don't have a technical background. And then our basic lateral flow training participants that come in and they don't realize, you know, you used a word tweak, which is always something. In science, people typically say you can't tweak an assay. But I think, you know, what's interesting is the whole part of developing a lateral flow is that's what you're doing all the time until you lock it down, you know, until you get it past that feasibility and you move into optimization and further development. Because those pieces all work together. I think you use the word, you know, it's a fluidic system. And I think that's why the tweaking is so important. And I recently was sitting in asking a lot of our training participants, what's that one thing that you learned that surprised you? And someone said, there are way more variables than I ever thought coming into this. I thought it was going to be simple to put these pieces together and it was going to work if I had, you know, the right, you know, combination of biomarkers, etc.. I think that's you're definitely hitting on a lot of things that we try to, to talk through as far as the complexities around it. I once asked an attendee of the training class what the most important thing was that they had learned, or what the most important variable to adjust was. And she said everything. It all matters that and you have to and you have to prioritize.

Helen: Yes. Yeah, I think that's where the expertise comes, the prioritization of what are you going to change and not change first, and in what order do you do that.

Mitzi: So that's certainly something that you've mastered. Let's talk about the client work for a minute. You know you're stepping into this new, you know, new role fairly new role to you as a program manager. What's something you're looking forward to as you get more time with clients and more clients in a shorter amount of time, as you'll be managing multiple programs

Helen: In that first meeting with the client, we get to have an in-depth discussion on what is it they really want. And that's always very interesting in that you learned there the clients who their clients who know exactly what they want. You have clients who already have expertise in LFA or in product development. And in that case, maybe we're collaborating on troubleshooting or we're guiding them through iterative testing. And then sometimes you have clients who have a little less commercial experience. And it's it makes me feel really good to help them figure out what those next steps are. They might be a startup coming from academia where they don't quite understand what's involved. When you're thinking about manufacturing or they understand in theory, but in practice it means, well, you actually should get five lots of materials and make sure that three of them are working. If you've got something a little more tricky from a reproducibility from the vendor viewpoint. It makes me feel really good to help them understand. What those trade-offs are for manufacturing and what does it mean to produce this in real life versus having a graduate student make this at 3 a.m.?

Mitzi: Yes. Yeah. Solving problems, finding solutions, feeling purpose and helping our clients, you know, with the outcome that they hope to have. That tends to be a real common theme amongst the DCN team. I do hear that quite a bit. Is there a particular project or a moment that stands out from your time at DCN, where something just finally clicked, or where you were helping a client to avoid some costly misstep?

Helen: We had a client who was, I will highlight they were running a non-standard assay, not in an LFA system, and they were converting to an LFA system. And as part of that, they didn't have a control line. Normally, a control line is not a big deal to add, but because they had a non-standard membrane, it was actually a little more complicated than one would have expected. So I feel that because of the expertise we had here and because I could consult our team members, that having that team around us is great, I was able to get to a much faster production of that control line so they could drop it in and everything is still working. We didn't break anything in the meanwhile and just get to that point faster for the client. In my role, I get to hear a lot of challenging assay development projects. But I'll tell you, I love hearing these stories. It never gets old to hear successful outcomes and to learn something new that we did to help one of our clients.

Mitzi: So thanks for sharing that. You've worked on tests for a huge range of diseases. So COVID, TB, STIs, malaria, maternal health. What's something you've learned across all of these different projects that you think applies to almost every lateral flow program, no matter the analyte?

Helen: Broadly speaking, it would be to get in your real clinical samples as soon as possible. You've got biological variability from person to person. You even have biological variability within one person over the course of a day. So that biological variability testing for interferences is very important. The other very different component from the biology would be designed for manufacturing. You are going to eventually want to produce a test on a larger scale. That means you are going to need to get enough material in. So if you need capture reagents, you need milligram or more amounts of that. And if your vendor can't produce that, you're not going to be able to make your test. You need reproducible materials. So, test multiple lots of your LFA materials. If it only works on one lot of your membranes, something is over optimized and you're going to have to either buy a large amount of the one lot, which is kind of a not a great way to go, or you're going to need to work out how to make your test work on multiple lots from the same membrane from the same manufacturer. So those would be the, I guess, two points that apply in very different directions across most LFA tests.

Mitzi: That makes sense. There are a number of projects that have kind of ended up at DCN, you know, and I and I get I have a privilege of getting to touch almost all of those in the very beginning that where they were struggling and where their assay was failing was exactly what you said they had been using contrived or made up samples in the early phases and didn't start with samples early, clinical samples early. So certainly I've seen people fail and have to do rework because of that. So that's certainly very relevant. And when we talk about design for manufacturing, I mean somebody can have a great idea and this great diagnostics, but if it's not developed in a way that can make it commercially viable and at a cost that the market can bear, depending on what is your market, then a lot of that is just waste of time. So it's nice to hear you say how important that is to DCN from the very, very beginning that we are thinking about that as soon as we start a feasibility program. So that's really nice. So let's go a little bit deeper into some technical ground you've really mastered. Labels can seem like a plug-and-play choice, but they come with a ton of hidden tradeoffs. So what are some of the things that developers might not expect when they're selecting between gold latex fluorescence or cellulose nanobeads?

Helen: When you're selecting a label there's the trade off basically between sensitivity and cost. That's usually where your trade offs end up being made. You also want the label to have good stability when you dry it down and be compatible with the reader if you are choosing to use a reader. So I'm going to start in order of sensitivity because this is also the most expensive. Generally speaking, if you want the most sensitive system, you'd start with fluorescence. But now you do have to have a reader. It's not optional. You will have a reader and you do have the cost of the reader as well as the fluorescent system. Cellulose nanobeads are next in sensitivity, but they tend to cost more than the latex or gold does. And the gold and latex have similar sensitivities, but how well they're going to work can be very assay dependent. The chemistries are usually different when you put your detector reagent onto those beads. Latex usually costs more. It gives you more colors though. So if you're doing a multiplex and you want different colors, you can get different colors. Gold is usually going to be red for you. One other thing to consider is that while they a lot of these chemistries to put your detector reagent onto, these materials have been standardized. They do kind of drop into two generally different categories. One is passive conjugation. So a lot of gold is done with passive conjugation. And a lot of latex conjugation is done with a covalent conjugation. You might have a protein. It's usually an antibody. But you might have a protein that cares. Which chemistry are you applying. And it might work on one or not the other. So you want to be in the final layer label you're going to be using. Don't start in the cheap fast one and then move over because if it doesn't work, then you just have to re-screen things, which will just be aggravating.

Mitzi: That makes a lot of sense. You know, something we sometimes hear from our clients during some consults is they'll say, well, we're already working with this label, or we've partnered with this company for this label, or we heard this is the label to use. And then whether they're physically or mentally kind of locked in to one specific target. Keeping that in mind. On the flip side, DCN is agnostic to label choice. I mean, just as you as you were just describing, you know, you come into a feasibility. You're putting everything out on the table to, to think about what's going to be the best for the assay and not specifically or what you said in the very first, what are you call you have with the client? What is their outcome? What do they really want? And you're putting all that thought into kind of what's that next step. So when you're reviewing a new LFA concept or an early prototype, what do you look for? What kind of gives you that confidence that. I think this is going to work or makes you pause? Like, I don't know if we should use that label.

Helen: What gives me confidence is when the client has well-defined goals so they know what they want. Are they more interested in sensitivity or specificity? Yes, everybody wants both, but usually there's a trade off from a developing the assay. Are they looking at how long it's going to take to develop the assay or how much money? For example, if they would like me to screen antibodies, I would like to get a lot of antibodies in at once and screen everything. If I screen four antibodies and they don't work very well. So then we bring in four more antibodies. You are literally waiting for the vendor to ship you more antibodies. And then you're going back and screening some of the first ones you use the first time. So getting ten antibodies in at once will make the whole process move much faster. But now you paid for ten antibodies up front and you're not using at least half of them, and probably more. You're trying to narrow things down as quickly as possible. From that viewpoint, if they can hand me materials where the capture and detector reagents are already known to work together. That makes me feel very good for a confidence viewpoint. They've worked in one format. Can we convert them? They should convert if they worked in an ELISA. Still not a guarantee. There's always the devil in the details. That does give me more confidence that an LFA can be built with those particular materials. I like to hear that the reagents are stable at room temperature. We are eventually would like to be striping some of these materials on membranes, which can be. If you're making a lot of them a process that takes an hour or longer, especially if you're looking at reel-to-reel. So those materials need to be stable at room temperature. I'd like to have at least a milligram. I can work with less. I want a milligram. Please give me a milligram. If getting a milligram is hard. Manufacturing is going to be a problem later. Later. Downstream. Those are some of the things I like to see when someone hands me a feasibility system.

Mitzi: Thank you. I love hearing how you think through all the different scenarios and what's I guess, provides you more confidence when you just get started. And on the other hand. What's the moment you realized it was time to pivot or maybe rethink an assay strategy entirely? Is there a red flag you've learned to trust?

Helen: There is a point at which sometimes the clients have given us reagents that they tell us should work in an LFA, and then later, as we progress and they're not working, it turns out nobody has ever tested these reagents, or they tested a different version of them from a different vendor, and they're custom made. So having materials that have never been qualified is a concern.

Mitzi: You know, it's kind of dovetails into what you're just saying. You know, someone says they know this is going to work and it doesn't. You've worked with a lot of interdisciplinary and commercial teams. So what's your approach to communicating complex assay trade-offs with business stakeholders or teams that are not in R&D?

Helen: Generally speaking, when you're talking to interdisciplinary groups, it helps to one, understand who your audience is since it's interdisciplinary but also avoid technical jargon. It is very easy as a scientist to use a bunch of acronyms. So stop doing that. That said, the first, I'm going to give you a technical jargon in that a lot of times in my field, the client will have a target product profile or TPP. The business people also understand TPP, generally speaking. So if we talk from there, everybody usually understands what we're talking about. And any assay considerations we can translate into the TPP will generally make sense across the board to everyone.

Mitzi: You know, I think it's important to involve the client, no matter how technical they are or aren't, and help them understand options and trade-offs. And I think being transparent, guiding the client based on my experience with you and the team, this is just really something DCN does very well, which is nice. So before we wrap up, I want to ask, what keeps you excited about this work? Like, you've been doing this for a long time and your energy is contagious. Like, what motivates you?

Helen: One thing I have realized about myself is I enjoy solving puzzles, and each new client with each new project is a chance to do problem solving. I get to problem solve in an area I understand, which is the LFA part, but I also get to usually learn something new in that the client's assay will usually have a target or sample matrix that I'm I personally am not familiar with. So I get to learn something new about that particular disease or protein or sample matrix. That's what I enjoy about coming to work.

Mitzi: I love that. Any final advice for someone who's just getting started in the LFA world, whether they're a bench scientists or they're managing their first diagnostic development program?

Helen: I would say don't rush your material selection. Going back to the LFA materials, which are your microfluidic system. Test 2 or 3 more if you can manage it, but you can start with 2 or 3 different materials with each iteration, finding some space where your assay is working. That said, do go to real samples as soon as you can understand what materials you have. Are your sample pads pretreated with chemicals already? Sometimes people will want to treat their sample pads as well. Did your capture reagent come with a carrier protein from the manufacturer? Read the fine print in your in your certificate of analysis. When you screen your reagents and are optimizing. Do this in as close to the final form as possible. For example, if your final test is you know you're going to use a fluorescent reader, do your screen with fluorescent particles. If you screen with gold, because it would be very fast to screen with gold particles. It could be that you select an antibody that doesn't like your chemistry to put things onto the fluorescent particles, so start with the end in mind. My last suggestion would be to talk to your coworkers. They might have experience with your sample matrix that you don't have for example. And they can give you some suggestions. Then always consult your coworkers. You're all on the same team. I get to see that here at DCN. Everybody's always willing to give me a tip on what I could be doing to make the assay better.

Mitzi: Those are all really important points to remember and very insightful. Helen, it has been a pleasure having you on. And for our listeners who want to dig deep into the kind of bench level insight Helen is sharing, check out our new Lateral Flow 101 blog series on DCNDx.com. Thanks for listening and we'll see you next time.