Brufsky Theorizes How Molecular Interaction Could Lead to COVID-19 Infection

Adam M. Brufsky, MD, PhD, discusses the impact of the COVID-19 pandemic on patients with cancer, as well as the science behind the proposed explanation of the immunopathology of COVID-19.

Adam M. Brufsky, MD, PhD

The rapidly evolving coronavirus disease 2019 (COVID-19) has sparked questions pertaining to the immunopathology of the virus, said Adam M. Brufsky, MD, PhD, a professor of medicine and associate chief in the Division of Hematology/Oncology at the University of Pittsburgh School of Medicine.

A recent paper published in the Journal of Medical Virology aims to begin to unify preclinical and clinical observations pertaining to the COVID-19 virus, said Brufsky, who is the lead author on the paper, coauthored with Michael T. Lotze, MD, of the University of Pittsburgh.

"[Dr Lotze] and I formulated the idea that some of the immunopathology of SARS-CoV-2 can be explained by an infection of the dendritic cell, mediated by the ACE-2 receptor and the DC-SIGN co-receptor on the dendritic cell and endothelial cells, as well as the ACE-2 receptor and the L-SIGN co-receptor on type II pneumocytes of the lung. It is very similar to mechanisms used by SARS-CoV-1 and HIV" said Brufsky.

In an interview with OncLive, Brufsky, who is also medical director of the Magee-Women’s Cancer Program, co-director of the Comprehensive Breast Cancer Center of the University of Pittsburgh Medical Center (UPMC) Hillman Cancer Center, discussed the impact of the COVID-19 pandemic on patients with cancer, as well as the science behind the proposed explanation of the immunopathology of COVID-19.

OncLive: What is known about the COVID-19 infection and the risk it poses to patients with cancer? What symptoms should people be aware of?

Brufsky: We have very limited data on COVID-19 and cancer. There is a small case series from Wuhan, China that was put out in preprint about 1 month ago. Those data suggest that people who are on active chemotherapy appear to do a little bit worse [compared with patients with cancer who are not]. Particularly, patients with lung cancer or hematologic malignancies appear to have worse outcomes with COVID-19.

We don't yet know what that means, but clearly, if patients are on active chemotherapy, it is difficult to predict outcomes with COVID-19.

I am a breast cancer doctor. The most important thing I tell my patients is that, “Individuals who have already been through chemotherapy with tamoxifen or anastrozole are probably [not at a heightened risk].” Patients who had breast cancer 5 years ago who believe that they are [at risk of developing severe complications from COVID-19] are likely not.

On the other hand, if a patient is on active chemotherapy, they should be careful. A lot of our patients are older; many of them have diabetes or hypertension. Some of our patients are men.

These are risk factors for COVID-19 that we have to keep in mind when we are treating patients.

Of course, these risk factors apply to everybody, but they have influenced how we treat patients with cancer.

How have you altered treatment for patients with breast cancer?

For example, [some of] the standard adjuvant regimens like dose-dense AC-T [doxorubicin hydrochloride and cyclophosphamide] or regular dose AC-T followed by weekly paclitaxel [have been altered].

Since COVID-19, I've started to give paclitaxel first. Though patients come in every week, by giving the paclitaxel first I know their [blood] counts are not going to drop very low. In fact, I will likely continue to give my patients paclitaxel first going forward; I like it for a variety of reasons. With adjuvant chemotherapy for breast cancer, it is not a bad idea to give the paclitaxel first because you can see how patients do with a gentler regimen.

Additionally, we are holding off on CT scans. Rather than scan a patient after 2 months, we may hold off on scanning them until after 3 or 4 months. Additionally, if a patient needs a port, we may wait to place it until after they've completed 1 or 2 cycles [of treatment]. Instead, nurses will do a peripheral [line].

How have UPMC and the Pittsburgh area been handling the pandemic?

At UPMC and in Pittsburgh, we are fortunate. We did not have a lot of COVID-19 cases. In fact, we have had few new cases in the intensive care unit (ICU) in the last few days. We have decided to open up a little bit and slowly start going back to normal. For example, I typically have about 25 patients on my clinic days. I was down to 8 to 10 at the peak of the pandemic in our area, and my clinic is slowly filling up again.

We've been doing a lot of telemedicine appointments, which have their pros and cons. I'm confident we will find our way. We are likely going to continue to use telemedicine going forward, pandemic or not.

Moving to the paper you recently published, how did you get involved in this research?

We have for years used hydroxychloroquine in clinical trials to treat breast cancer. There is a clinical trial currently at Penn using hydroxychloroquine in patients who have positive disseminated tumor cells (DTCs) in their bone marrow. The idea is that you would perform a bone marrow biopsy, use hydroxychloroquine for [an established] period of time, repeat the biopsy, and see if the DTCs are still present. Then, the results would be correlated with overall survival and disease-free survival.

That study got me thinking. I knew there was literature in lupus pertaining to the hypoglycemic action of hydroxychloroquine. With all the discussions about the use of hydroxychloroquine in COVID-19 pneumonia, I thought it would be interesting to investigate these potential hypoglycemic effects.

I was also having conversations with some infectious disease specialists taking care of patients in the ICUs near the peak of the epidemic in the New York City metro area. Those physicians told me that about 50% of patients [with COVID-19 in the ICU], had diabetes and another 20% had prediabetes. Another proportion of patients had unexplained hyperglycemia.

Therefore, putting this together, I thought that perhaps the hyperglycemia could be a key to understanding the pathogenesis of SARS-CoV-2.

What preclinical research did you conduct to flush this idea out?

We started looking at reports of animal models. There is a model called the non-obese diabetic (NOD) mouse that has a lot of ACE-2 in the lung. When the NOD mouse gets insulin, its ACE-2 in the lung apparently goes down. However, the ACE-2 activity stays the same. When you look carefully at the mice, it appeared to be the amount of glycosylated ACE-2 that goes down rather than ACE-2 itself, as the Western blot antibody test that is being used to measure the protein in the mice is actually measuring glycosylated ACE-2 rather than all ACE-2.

It is important to understand that glycosylation and glycation [are different]. Glycation is when sugars chemically bind to proteins, like hemoglobin A1c or the crystallin in the lens of the eye. That is directly and chemically adding glucose molecules. On the other hand, glycosylation is a cellular enzymatic process that, in response to metabolic changes such as diabetes, could add or subtract sugar [molecules] to proteins at very specific sites in the endosomal compartment of the cell.

We also discovered in the literature that a number of patients with SAR had unexplained diabetes with their infection, which resolved when their infection resolved.

According to this research, how can we understand the immunopathology of these viruses?

The theory is that the disease can be explained by the amount of glycosylated virus , the amount of glycosylated ACE-2 receptor, and the amount of various proteins of the immune system that are glycosylated around day 7 or 8 when the inflammatory reaction peaks. If there is a lot of each, the response can be cytokine storm.

SARS-CoV-2 does mutate, and it allows us to track its spread. Some of these mutations result in non-synonymous amino acid changes within the spike protein, and other proteins, of the virus. We don’t know for sure whether these changes alter the characteristics of the virus, but some believe that they may.

GISAID [The Global Initiative on Sharing All Influenza Data] is a German group that catalogs SARS-CoV-2 viral sequences in a database. Nextstrain is a group that [runs] an open-source website that can allow us to visualize the changes in the viral sequence through time. In looking at the database, we can see where the various mutations are, and when, to crude approximation.

With all the caveats to this, specifically uncertainties and biases in the analysis, there appeared to be a mutation in the viral spike protein (S protein) where an aspartic acid at residue 614 is changed to glycine. This mutation predicts to increase glycosylation at a site at amino acid 616.

How can we extrapolate these results to COVID-19?

There is a lot of work that has been done with SARS. [Scientists] made pseudotypes of the SARS virus—a lentivirus—that they can work with. They put various SARS proteins in the lentivirus to see what happens. The reasoning behind this is that they cannot work with SARS unless they are in a biosafety laboratory where it could be investigated safely.

These investigators made a viral construct that added a SARS-CoV S protein to a lentivirus and infected various cells. The construct bound to the ACE-2 receptor to as expected, but also used a second receptor called the dendritic cell-specific intercellular adhesion molecule-3-grabbing non-integrin (DC-SIGN).

DC-SIGN is a mannose binding protein that is responsible for looking for carbohydrate antigens on pathogens. Bacteria, parasites, and viruses have carbohydrates on them. They are non-specific receptors, but they work with ACE-2. Then DC-SIGN helps to phagocytose [the virus] into the dendritic cell.

I started having conversations with Dr Lotze, who is one of the world experts on the dendritic cell that we have at the University of Pittsburgh. We came up with the idea that the some of COVID-19 can be explained by an infection of the dendritic cell, mediated by the ACE-2 receptor and the DC-SIGN co-receptor on this cell.

How did you and Dr Lotze formulate this theory? What scientific evidence is there to suggest that this interaction could add to our understanding of COVID-19?

[This theory could] explain a lot because these lentiviral pseudotypes of SARS show that the more glycosylation you have on the spike protein, the higher probability that DC-SIGN is going to help [phagocytose the virus] into the cell. The more glycosylated sites there are in the spike protein coat, the higher chance virus will bind to these receptors.

That possibly argues that if you have more glycosylation, the disease should be more virulent. We went searching for possible glycosylation mutants, and we may have found them. Also, we may have found some mutants that predicted for less glycosylation.

The virus has a glycoprotein coat. The virus recognizes ACE-2 and DC-SIGN and binds. It is then phagocytosed into the cell and sits in the lysosomal compartment. This causes dysfunction of the dendritic cell and confuses the immune system, resulting in a temporary loss of other immune cells such as T cells and B cells. The virus appears to hide in the dendritic cell, waits for the immune cells numbers to go down, and is expelled. This is called the immune synapse.

This viral synapse mechanism has been identified in SARS and in HIV, so it is plausible to be operative in SARS-CoV-2 infections.

There was an interesting protein-interactome study published recently in Nature that took 26 SARS proteins and evaluated what cellular proteins bound to them. They found 332 proteins that interacted. They found that of those 332 proteins, 40% were in the endosome. Clearly, there is something important going on in the endosome of cells in COVID-19.

The third most common mutation found right now in SARS-CoV-2 is in the ORF3a protein and is called Q57H. ORF3 is an ion channel that appears to be responsible for expulsion of virus from the lysosome. If you have this mutation in ORF 3a, Q57H, that sits in the middle of an ion channel, it could possibly to make the virus weaker because the virus cannot be expelled.

That is the idea behind what is happening in this D614G mutation. Maybe you are getting a second [mutation] in this strain of virus that attenuates it by interfering with the viral synapse.

The vast majority of time, when you introduce a virus to a novel host, it attenuates over time. This mutation in the ORF3a protein looks like it could be attenuating. Of course, this needs to be tested.

Groups around the world are starting to find multiple mutations in SARS-CoV-2 that can possibly change the structure of viral proteins. You can take these mutations and apply them to the protein interactome study that was published in Nature to try to possibly predict what they would do.

One question that has been asked is, “Has this been done in cell culture?” It turns out that it has. There is a group in China that published a paper looking at 11 different patients in Wuhan with 11 different viruses and compared them with the viruses seen on the West Coast of the United States and the ones seen in Europe. There was clearly a difference in virulence. The one similar to European strains were more virulent in cell culture; they killed more vero-E6 cells and had a higher viral load versus the ones that were more similar to viruses from the West Coast of the United States.

How could the immunopathology of COVID-19 explain the disparity between transmission and mortality in the West Coast versus the East Coast?

In Nextstrain, it looked like the D mutation was on the West Coast of the United States and the G mutation, which appeared to travel to Europe from China, was on the East Coast of the United States.

When you look at the mortality rate per million, it is much higher in New York than in California and the whole West Coast. That was strange to me. How is it that in California where nearly 40 million people live, there are roughly 2500 deaths, nearly 7 times less than New York City, with a population of 10 million? It could be that people are spaced further apart in California, but what about San Francisco Bay Area, which is quite crowded?

Virology experts saw this D614G mutation as well. Wherever a virus with that mutation went, it appeared to take over from other viruses. In Sheffield, England, the number of people in the hospital with the virus with the G mutation versus the D mutation did not appear to be statistically significantly different. There was a non-significant increase of patients with the G mutation who needed the ICU.

What are the next steps with this research?

The dendritic cell explains a lot, but this is again, just a theory that needs to be proven. I sent this theory to a lot of people asking for their opinions. A lot of people wrote back saying that this is an interesting idea, but that is not likely to get into a major journal because there are no data to back it up. That is perfectly understandable.

Based on these findings, what is your assessment of the current investigational treatment strategies for COVID-19?

We have some observational studies that state that hydroxychloroquine has benefit, and others that say it does not. We all await randomized clinical trials.

Hydroxychloroquine appears to alkalize the lysosomes. The virus needs an acidic environment to function, so this could modulate the virus. Hydroxychloroquine is also an immunomodulator that therefore could possibly blunt the cytokine storm.

There are other interesting things out there that are being explored. If you think about COVID-19 as an infection of the dendritic cell as well as other cells, it becomes clear, but I am going to wait until the observational and randomized clinical trials come out to mention those.

How could the strategies proposed in the paper impact the trajectory of COVID-19 and potentially other related viruses?

An important thing to understand is that after the 1940s, the way we did science and medicine is that we found an idea in the bedside and went back to the bench to figure it out scientifically by coming up with theories and then performing large clinical trials to test them.

COVID-19 is a once-in-100-years medical crisis. Given the speed with which the crisis hit us, some of us went back to the way we used to do things; make empiric observations at the bedside and then appeal to the literature to try to explain them. Then take the theory back to the bedside and see if it explains more of what you see. If it doesn’t, refine the theory or change it. Eventually the science will get done and support it or not.

What is your take-home message regarding the COVID-19 pandemic?

We were stunned by this [pandemic]. People were dying and we had no idea why. Our ICUs were filling up.

What is happening now is that, after getting over the initial stun, physicians of all specialties—oncologists, infection disease experts, endocrinologists, virologists, molecular biologists, and immunologists, among many others—are all excited to figure out this tough problem. This is why we all went to medical school.

The [healthcare community] is starting to get a handle on this. We still need to be cautious since much is not known and we only have theories. However, coherent testable frameworks are starting to appear to try to assist us in figuring out where we go from here.

Brufsky A, Lotze MT. DC/L-SIGNs of hope in the COVID-19 pandemic [published online ahead of print May 6, 2020]. J Med Virol. doi:10.1002/jmv.25980