Introducing… Fundamentals

There’s no debate that the last few years have seen some recurring themes in pathology and laboratory medicine. Infectious disease goes without saying – but we’ve also heard more than ever about digital and computational pathology (and the telepathology that goes hand-in-hand with it!) and about the rise of molecular pathology and the omics revolution.

What does that mean for us – and for you? We’ve set aside some space in our print and digital worlds to showcase these disciplines in more detail. We hope this will give us the opportunity to share new developments as they happen, and to make it easier for busy professionals to stay abreast of these quickly changing topics. But don’t worry – we’ll still cover the rest of the pathology world in the same detail as before, from practical tips in the lab to big-picture looks at training and career topics!

To give you a taste of what’s to come, we invite you to take a look at some of our new “fundamentals” content. Scroll on to read more – and keep an eye out online and in print to see more articles like these!

Digital Pathology

(PANDA) Challenge Accepted

Challenging the world to develop artificial intelligence for prostate cancer grading

Over 10,600 digitized prostate biopsies; 1,290 developers; 65 countries; and just one goal: to design an algorithm that could automatically assign a Gleason grade group to a biopsy sample. A fun challenge for developers and digital pathology enthusiasts alike, this is the aim of the PANDA (Prostate cANcer graDe Assessment using the Gleason grading system) challenge, set by Wouter Bulten and colleagues from various institutions (1).

Why crowdsource artificial intelligence (AI) this way? The team wanted to take the next step toward automating Gleason grading for prostate cancer – and the challenge was the perfect setup. To help developers in their quest, the researchers released all training data from two large studies on automated Gleason grading, along with slide-level labels and label masks. Though no method was off limits, teams had to follow some instructions to make sure biopsies could be loaded correctly during the evaluation stage. What made the PANDA challenge unique was that teams were asked to submit their algorithms; test data was kept hidden from participants so that cheating wasn’t an option and results could be translated to other datasets.

Final entries into this largest known histopathology competition were validated on a private set and were given a time limit of six hours to analyze 1,000 biopsies. In a blog post on his personal website, lead author Wouter Bulten gave insight into the selection process from there. “After the competition ended, we selected 15 teams to join for extensive independent validation of their algorithms on new data. The selection was based on the score on the final leaderboard, method description, and scientific contribution. The latter criteria were used to ensure we had a good representation of algorithms for the analysis (2).”

Bulten and his team then reproduced these algorithms and tested them on external validation sets from Europe and the US – yielding agreements of 0.862 and 0.868 with expert uropathologists. In his post, Bulten highlights how global challenges like this can benefit the development of AI tools. “Challenges are often a powerful way of crowdsourcing new AI innovations. The PANDA challenge was no exception; due to the scale of the competition, after 10 days, one of the algorithms was already at the level of the average pathologist. In the remainder of the challenge, many teams caught up and improved further. This speed in development was also driven by extensive sharing of tips and tricks through the challenge forums (2).”

Will we see more of these crowdsourcing development challenges? Hopefully – not only do they sound like a great way to engage a talented pool of developers in pathology, but they also promote interdisciplinary working between laboratorians and technologists to achieve a common goal.

References

1. W Bulten et al., Nat Med, 28, 154 (2022). PMID: 35027755.
2. W Bulten (2022). Available at: https://bit.ly/3BagU6m.

Molecular Pathology

Keeping the Balance

Examining the role of d-serine in kidney disease and physiological function

Previously, d-serine has been associated with kidney function and disease activity; however, the biomarker’s function remains unclear. In two new studies, researchers have investigated the physiologic functions of d-serine and the potential of d-serine clearance as a measure of glomerular filtration rate (GFR). We spoke to Tomonori Kimura, senior author on both studies, about the research – and how it will support future evaluation of kidney dysfunction.

Measurement of glomerular filtration rate using endogenous d-serine clearance in living kidney transplant donors and recipients (1)

How do we currently evaluate kidney function?

Glomerular filtration rate (GFR) using exogenous inulin has previously been considered the gold standard of kidney function measurement, though it is time-consuming, labor-intensive, and expensive – and, therefore, rarely performed.

Because of this, measuring or estimating GFR using endogenous molecules has been widely adopted. Creatinine clearance (C-cre) correlates strongly with GFR; however, it suffers from major proportional bias and overestimates GFR due to kidney tubular secretion of creatinine into urine. Estimated GFR (eGFR) – calculated using a combination of age, sex, race, and serum levels of creatinine or cystatin C – is convenient and useful for screening chronic kidney disease (CKD). Most equations for eGFR show relatively small bias; however, they are not precise enough to provide a correct estimation for individual patients. Furthermore, creatinine has another problem – because it is waste from the muscles, blood creatinine levels are higher in younger populations and lower in elderly people. Therefore, kidney function in older populations is often overestimated.

Tell me about your new method for evaluating kidney function…

We use d-serine clearance to measure glomerular filtration rate (GFR), which describes the flow rate of filtered fluid through the kidney. GFR is calculated as clearance, which is the quantity of substances in urine that originated from a calculable volume of blood. For the substance used, the product of urine concentration and urine flow equals the mass of substance excreted by the kidney during the urine collection period.

We measured levels of d-serine in blood and urine and calculated the d-serine clearance (C-dser) level. C-dser is less biased than C-cre and, because both C-dser and C-cre are precise enough to measure GFR, we combined the two methods. Unsurprisingly, we found that this method greatly reduces bias and can measure GFR with a high performance.

Did you encounter any surprising findings?

The first time we saw the strong correlation between clearance of d-serine and inulin was astonishing (even for those of us who knew the marker’s potential). In 2016, I was interviewed by The Pathologist regarding our first report that d-serine is a kidney biomarker. We also reported in 2019 that d-serine reflects GFR and, since then, have observed that d-serine reflects GFR. Even so, we did not anticipate such high agreement between d-serine clearance and GFR.

What’s next for the research?

The study was conducted in Japan with a limited number of participants. Now, we need to validate our findings by expanding the cohort to multicenter studies in an international setting so people around the world can be offered the test. Once international research has been conducted, we can immediately use the test in clinics.

d-Serine Mediates Cellular Proliferation for Kidney Remodeling (2)

What prompted you to study d-serine in relation to kidney function?

In our previous study, we found that d-serine is a vital biomarker for kidney function – but its physiologic function in the kidney is relatively unclear. d-serine is a neurotransmitter that works via the N-methyl-d-aspartate receptors in neurons, though it is also supposed to function in the kidney. Treatment with extremely high doses of d-serine causes acute kidney injury in rodents, suggesting a direct effect of d-serine on the kidney at least at the supra-physiologic level; however, its role in human disease remains unknown. 

Additionally, a lack of key information (such as tissue distribution and mechanisms of action) has obscured the physiologic functions of d-serine. In the previous study, we noticed that not all d-serine is excreted into urine – a fraction is taken up by the kidney tubules after glomerular filtration. Therefore, we considered that there should be some physiological function of d-serine in the kidney at the physiological level.

What were your findings?

d-serine has a cellular proliferative effect on the kidneys. For example, when kidney size is reduced, blood levels of d-serine increase due to the reduction in urinary excretion. In this condition, increased levels of d-serine mediate kidney remodeling, which increases the urinary excretion of d-serine and reduces its blood levels. As mentioned in the previous study, d-serine reflects GFR. The classic studies of d-serine were conducted at very high levels, whereas the physiological level of d-serine in the human body is very low. In the blood, l-serine is present at about 100 μM, whereas d-serine is around 2 μM. We initially thought that d-serine might function at the same level as l-serine but, to our surprise, when we reduced d-serine concentration to the physiological level, its proliferative effects became clearer.

d-serine is an activator of mTor signaling – an essential mediator of nutrients. The effects of l-amino acids in activating mTor signals are often examined, whereas those of d-amino acids have not been. Sometimes, d-amino acids are used as negative control for mTor signals by testing at supra-physiological doses; however, I would say that the true relationship has been overlooked because the function has not yet been tested at the physiological dose.

We also found that reagents sold commercially as l-serine contained 1 percent d-serine. This is not the fault of manufacturers because, when we synthesize serine chemically, l- and d-serine are usually produced at a 1:1 ratio. Manufacturers then remove the d-serine “contamination” – but removal is not easy and there is no way to monitor chirality in the final product. Therefore, most l-amino acids sold contain d-amino acids to some extent. The results of our study showed that d-serine can exert its function at 1 percent of the level of l-serine, meaning that if the concentration of cellular culture medium is 400 μM for serine, there is likely enough d-serine (around 4 μM) present to affect the results of a cellular proliferative assay.

Why are the findings important – and what could they mean for patients with kidney disease?

To our knowledge, there is no way to increase kidney function or enlarge the kidney but, if you could do so by even a small fraction, this could help halt the kidney’s decline. Now, we know that the kidney can remodel to increase its function by using d-serine-related signals – and further studies of d-serine and its signals may lead us to a method of maintaining kidney function or even curing kidney disease.

References

1. M Kawamura et al., E Clinical Medicine, 43, 101223 (2021). PMID: 34934934.
2. A Hesaka et al, Kidney360, 2, 1161 (2021).

Infectious Disease

Parasite Predictors

Defining a new way to predict clinical outcomes in malaria patient

The World Health Organization estimated 241 million malaria cases and 627,000 deaths from malaria worldwide in 2020 alone – respective increases of 14 million and 69,000 from 2019. Levels of Plasmodium parasites in the blood are often used as a measure of disease severity – but patients can suffer from poor clinical outcomes even in the presence of low peripheral parasitemia.

To find out how, researchers from the University of Campinas investigated the relationship between peripheral parasitemia and total parasite biomass and host response in Plasmodium vivax malaria (1). Their analysis revealed two patient groups (Vivax-low and Vivax-high) based on differences in total parasite biomass, but not peripheral parasitemia. Vivax-high patients showed significant differences in hematological parameters, glycocalyx breakdown, endothelial cell activation, and cytokine levels regulating different hematopoiesis pathways, and displayed more severe thrombocytopenia and lymphopenia than Vivax-low patients, who were more closely aligned with healthy donors.

“When we separated some of the patient plasma samples to stimulate endothelial cells, we observed strong modulation of the cellular monolayer without direct or indirect interaction with the parasite for endothelial dysfunction,” said co-principal investigator Fabio Trindade Maranhão Costa when talking to Agência FAPESP about the research (2).

A strong association between total parasite biomass and markers of endothelial cell activation, thrombocytopenia, and lymphopenia severity was found when patients’ signatures were combined. The findings demonstrate that combining host parameters and total parasite biomass could better predict infection outside the bloodstream than peripheral parasitemia.

Based on their findings, the authors suggest that “changes in clinical parameters and biomarkers detected in the plasma of P. vivax patients are the result of both systemic host responses and local infection in tissue reservoirs such as bone marrow and spleen (1).” Costa adds, “Biomass, rather than parasitemia, is associated with several problems. This is extremely novel and hasn’t been investigated in depth until now. Our findings highlight the importance of parasite biomass in the bone marrow and spleen. It’s clear that these two organs play a major role in vivax malaria complications (2).”

With P. falciparum and P. vivax the two most common causes of malaria, the research has significant implications for unraveling the two. “Every step we take in our research confirms that P. vivax is different from P. falciparum,” Costa told Agência FAPESP (2). “From afar, they may seem the same, but when you look more closely you detect the differences. This is important for the development of more effective treatment, specific controls, and even a vaccine.”

References

1.  JL Silva-Filho et al., Elife, 10, e71351 (2021). PMID: 34585667.
2. Luciana Constantino (2021). Available at: https://bit.ly/3IJnnYz.