Researchers identify autoantibody targets in COVID-19 patients

By Sam HancockReviewed by Aimee MolineuxFeb 14 2022

Coronavirus disease 2019 (COVID-19) has shown a very wide variety of symptoms, from short-term flu-like effects with relatively little danger, to organ failure and severe lung damage, to long-term fatigue. Some researchers have also suggested that there might be some risk to fertility, with angiotensin-converting enzyme 2 (ACE2) highly expressed in male reproductive cells.

Researchers from Proteomics Core have been investigating auto-immune responses to natural human antigens in COVID-19 patients.

A preprint version of the group’s study can be found on the bioRxiv* preprint server while the paper undergoes peer review.

*Important notice: bioRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.

The study

The researchers identified the total IgG response of two sets of COVID-19 patients from different ethnic groups and their responses against 1,318 naturally folded human antigens. One cohort was recruited at an ICU in Doha, Qatar, while the second was recruited from the ICU of a hospital in the United States (US). They managed to gather a total of 97 cases (49 in Doha, 48 in the US), and 76 controls. They began by using KREX high throughput autoantibody assay technology to discover IgG autoantibodies that could affect the outcome of COVID-19 infections. They initially found 1,600 proteins, but reduced this to 1,318 in the replication cohort for increased stringency. The antibodies found varied in location, with most in the cytoplasm, nucleus or cell membrane. The KREX assay reported RFU values that correlate directly with antibody titers, with higher RFU values indicating higher titers or repeated exposure.

General intensity distribution was initially calculated on the mean antibody-antigen titers across all samples. KEGG-Brite-based Voronai treemaps could be used to further examine the titers, and around 1,150 proteins were assigned to the annotation. Almost all showed strong IgG signals, with the highest autoantibodies found against mostly structure-related proteins, including RBPJ, TPM1, TACC1, KRT19, PTPN20, TBCB, KRT15, AFF4, HSPD1, and CBFA2T3. The lowest titers tended to be against cytoplasmic proteins involved in phosphorylation.

In the Doha cohort, the researchers performed a differential expression analysis between cases and controls using a T-test. They found that 57 proteins had significantly altered autoantibody responses, forty of which showed increased responses, and 17 decreased. They found the most elevated responses against ATF4 and the sperm protein associated with the X chromosome N4 nucleus.

A binarized autoimmune response analysis, assuming that all samples with a response exceeding standard deviation by one as positive and all others as negative, alongside a Fisher's exact test, showed 25 patients had higher RFU values for SPANXN4, compared to only five in the controls. They also found significantly higher autoantibodies against ATF4, recombining signal binding protein J (RBPJ) and programmed cell death 5 (PCD5), but only SPANXN4 showed the required significance. When examining samples collected from patients six weeks after recovery, the researchers also observed strong correlations between autoantibody responses for certain proteins that remained highly elevated in recovery, including SPANXN4, STK25, TRAF3IP1, AMOTL2, PSMD4, and PPP1R2P9. A repeat of these comparisons in the US patients showed much of the same results.

Following this, the researchers performed principal components analysis (PCA) of RFU data from the two cohorts, which revealed a strong overlap between the two cohorts, which did not separate into discrete clusters. Pearson correlation analysis revealed a high correlation between the autoantibody responses of the two cohorts. When the two cohorts were combined and compared to the controls, the scientists discovered that 56 proteins showed a significantly altered response, with 35 increased antibody responses and 21 decreased. The proteins that showed the highest effect size were SPANXN4, ATF4, STK25, and PRKD2. Sequence homology and antigen specificity analysis were used to check for cross-reactivities. While a few proteins did show significant sequence homology, many of these were included in the initial KREX analysis and showed no significant changes.

Conclusion

The authors have revealed several autoantibody responses that are significantly elevated during COVID-19 infection, with some, such as SPANXN4, remaining elevated for extended periods post-infection. They suggest that this could have some association with the long-term chronic health issues some COVID-19 patients suffer, known as 'long-COVID'. They identify many of the physiological functions of the proteins, with structural proteins and reproductive proteins both present.

As well as this, they show there is little to no difference in the elevated proteins between the Doha and US cohorts. These results and the authors' analysis highlight the importance of immune responses and could be used to help clarify the functions behind long-COVID.

*Important notice: bioRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information.