How proteomics helped diabetic kidney disease research advance

Dr. Krolewski and his team at the Harvard Medical School found 56 proteins to be significant in diabetic kidney disease patients. Potentially, these could serve as prognostic biomarkers for disease progression and treatment response. This is how adding proteomics to the methodologies elevated their research.

Dr. Andrzej Krolewski, the head of the Genetics and Epidemiology section of the Harvard Medical School, has been working on the genetics of type 1 and type 2 diabetes. For the last 30 years, he has been actively investigating diabetic kidney disease in patient cohorts from the Joslin diabetes center, where he is also a researcher.

Andrzej Krolewski

Read more about Dr. Krolewski, Professor of Medicine at Harvard Medical School, and Joslin Diabetes Center on their website.

Before proteomics, genomics was the answer

Having started one of the first biobanks of its kind in the late 1990s, Dr. Krolewski collected data on 5000 patients over the past 10 to 20 years to understand the disease progression of diabetic kidney disease. This was the time where human genetics as a field was making huge advances, with the human genome project being just one of them. The promise of genetics as the solution to all our biological questions and conundrums captivated even Dr. Krolewski, who began genetic studies in earnest to make progress in his research.

However, despite his efforts, the results of this research were underwhelming, and he needed to utilize a different approach, using a different form of -omics.

Moving to proteomics

Having had no luck in exploring the human genome for answers to his research questions, Dr. Krolewski turned to the human proteome instead. His initial work was backbreaking; he and his team had to perform countless ELISAs and used up invaluable sample specimens to make any scientific progress. Their application of proteomics analysis to a high-quality longitudinal study cohort led to many discoveries, although these were hard-won.


He then became aware of Olink and ran a pilot study on over 1000 proteins for a subset of his samples.

Proteomics methodologies improved as time passed and Dr. Krolewski became aware of SomaScan. This technology uses SOMAmers, a modified form of nucleotide aptamer, to simultaneously analyze thousands of proteins at once, compared to one at a time with ELISA. He then became aware of Olink and ran a pilot study on over 1000 proteins for a subset of his samples.

Olink technology delivers on accuracy and sensitivity

Comparing this data to that of SomaScan, Dr. Krolewski found that Olink was hard to beat on accuracy and specificity. Olink also offered him the flexibility and scalability that SomaScan could not, enabling the progression from protein biomarker discovery to more targeted studies focusing solely on the proteins found to be most interesting after screening. In this way, he was able to scale down analyses from 11 Olink panels, down to just 5, saving on costs and precious sample.

Dr. Krolewski and his team used these 5 panels to run a study on 500 individuals from the Joslin longitudinal study cohort, for which their first paper has now been published (DOI: 10.1016/j.kint.2020.07.007). Overall, 56 proteins were found to be significant in diabetic kidney disease patients, and these could potentially serve as prognostic biomarkers for disease progression and treatment response.


Watch an interview with Dr. Krolewski to learn more about him and his work with Olink.

Dr Krolewski and his team are now working with Olink to develop 21 of these proteins into a customized panel to determine which of them may be used to predict kidney disease development in diabetic patients. He also plans to use the panel to conduct further studies on patient response to treatment, as well as make further advances in treatment for this disease, which are currently very limited.


Ihara K et al.,2021, A profile of multiple circulating tumor necrosis factor receptors associated with early progressive kidney decline in Type 1 Diabetes is similar to profiles in autoimmune disorders. Kidney Int. DOI: 10.1016/j.kint.2020.07.00