Donor monocyte therapy for COVID-19 pneumonia: Evidence and research proposal
Could critically ill COVID-19 pneumonia patients be rescued by a single infusion of COVID-naive monocyte-like cells, collected — with a little help from a safe, well-studied adjuvant booster — from an unrelated donor; no special donor match, no exotic lab equipment, and no prior donor exposure to COVID-19 required?
Outline of the proposed treatment
- This is a proposal to use specific white blood cells from healthy donors to protect COVID-19 pneumonia patients from a common mechanism of severe illness and death (cytokine release syndrome), not from the viral infection itself.
- Nothing in this proposal depends on prior COVID-19 recovery in the donor. Nor does it require any laboratory testing of donor or recipient, other than CBC and the common biomarkers already being used in practice, plus blood type and cross-match.
- The cell type that I believe to be critical — circulating (myeloid) angiogenic cell, or “MAC” — is not a stem cell. It’s a fully differentiated hematopoietic cell. Its chemical phenotype resembles a “typical” monocyte with additional activation clusters (proangiogenic up, M2 up, M1 down). It can be routinely cultured (or, when greater specificity is called for, selected by flow cytometry) from the monocyte fraction of centrifuged whole blood. Its function is to deliver paracrine/juxtacrine signaling factors such as IL-8 (and, in hypoxia scenarios, ACE2) directly where they are needed (principally at sites of peripheral vascular damage). Its function cannot be replaced by direct infusion of its soluble products, which are ineffective and sometimes harmful in diffuse elevation.
- The proposal is to mobilize these cells directly into the bloodstream of a suitably selected and primed donor, followed by transfer of the relevant circulating cell fraction to one or more patients (depending on extraction protocol). Earlier drafts of this document placed excessive emphasis on the fact that more or less direct whole blood transfusion is one way to accomplish this. The point here is fivefold:
1) culturing takes equipment and skill and time;
2) culturing alters cells in subtle ways that may impair their function;
3) culturing grows colony-forming cells (useless because not autologous) at the expense of the actual MACs;
4) fractionating (let alone flow cytometry or donation by leukapheresis) takes equipment and skill and time;
5) fractionating may lose an unfortunate proportion of the MACs if they don’t all land in the monocyte fraction from which they’re usually cultured.
Having said that, clinicians in well-equipped developed-world hospitals will presumably opt to run the donation through apheresis and select the monocyte fraction for infusion (or, given fewer donors but more time per donor, receive it by leukapheresis in the first place). You don’t need the plasma; this has nothing to do with antibodies or globulins or any other soluble component. You don’t need the lymphocytes; this has nothing to do with T cells, memory or otherwise. You don’t need the red blood cells; this has nothing to do with oxygen transport. - Few people outside the immunology and stem cell research communities know that MACs exist. They have historically been (and sometimes still are) mistaken for endothelial or mesenchymal stem cells, which they resemble in some key attributes that are used when selecting stem cells for culture. Some of the effects of “stem cell” infusions in the clinical literature are provably attributable instead to the proangiogenic function of MACs. However, they don’t actually take up residence in tissues, and so don’t have to match the HLA type of the patient; it is normal and expected for them to survive for only a few days after infusion.
- There is indirect evidence that a typical COVID-19 pneumonia patient’s own MACs are disabled by SARS-CoV-2, turning ordinary viral pneumonia into cytokine storm and death; but this is (as yet) clinically unproven. This virus’s entry mechanism into human cells is known to interfere with the function of the ACE2 enzyme complex long before the cell is visibly altered by viral replication, and ACE2 is critical to these cells’ paracrine function in hypoxic conditions. This effect is very difficult to observe in vitro, and there may be no good way to establish this in the near term other than to attempt treatment.
- This proposal includes the use of a readily available immunostimulant, plerixafor (Mozobil), to mobilize MACs from the donor’s bone marrow into the bloodstream in about 4x–10x their usual abundance. Most clinical applications of plerixafor are thought of as “stem cell” treatment, and involve autologous donation in connection with chemotherapy, in combination with G-CSF (which has significant risks of adverse side effects). However, there exists published data indicating very high safety for plerixafor alone in healthy donors (antigen-identical siblings): https://doi.org/10.1182/blood.V126.23.389.389. Based on some further advice I have received about circadian variation in CXCR4/CXCL12 binding strength, it may even be possible to get the full 10x boost in MAC abundance by administering a quarter-dose in the early evening, 16–18 hours before starting leukapheresis. This may be expected to reduce the (already low) risk to the donor of excessive immunostimulation, depletion of bone marrow reserves, or possible ill effects due to other sites of CXCR4 blockade.
- Depending on the dosage, harvesting, and transfer protocol, one donation may result in enough cell product to treat 2 to 30 patients, or perhaps as many as 120 patients per standard dose of plerixafor (US cost estimated at $9,000). In order to meet global needs during the COVID-19 pandemic, there is an urgent need to establish screening criteria that maximize yield per donor in both high-tech (leukapheresis) and low-tech (whole blood transfusion) scenarios.
- Part of the inspiration for this proposal is a clinical study conducted (by unrelated researchers) in Beijing YouAn hospital, published in Aging and Disease. All seven ICU-admitted COVID-19 pneumonia patients in the treatment arm of that study showed marked improvement within 2–4 days of cell culture infusion, under circumstances that appear very likely to be directly attributable to the infusion itself. My interpretation of the mechanism of their results is very different from theirs, as is the protocol I propose for preparing cell product for infusion. But my estimates of cell infusion quantity and hoped-for outcomes, and the expectation of relative safety given normal precautions around infusing donor monocytes, are largely based on their work.
- I am not a medical doctor; this is not medical advice. Before this treatment protocol can be attempted in hospitals generally, it will need to be proven in clinical study, if not full-scale clinical trial. I am now working with a team that includes MDs and PhDs in relevant disciplines to launch a three-part clinical research program to evaluate whether this treatment works and how it can be scaled to meet the need.
Thesis and outline of proposal
The proposed strategy is an attempt to replicate, by relatively low-tech means, the dramatic outcome of a “mesenchymal stem cell” therapy that I have read about — which may not really depend on stem cells, mesenchymal or otherwise. This small-scale clinical study, published by stem cell researchers in China treating patients at Beijing YouAn Hospital, has demonstrated remarkable success in treating severe COVID-19 pneumonia cases. All treated patients recovered, and most left the ICU within 48 hours; controls did not fare so well (dead, dying, or still in ICU two weeks later). An unproven hypothesis about the therapeutic mechanism suggests a low-tech, low-cost path to similar results via “enhanced” blood donation.
In my interpretation, the relevant cells are MACs, formerly known as “early endothelial progenitor cells” and now as “circulating (myeloid) (pro)angiogenic cells”. They land in the monocyte fraction during leukapheresis, and can be further isolated by flow cytometry or well-known culture protocols. (Some authors divide circulating monocytes into “classical”, “intermediate”, and “non-classical” cohorts; MACs lie in the “intermediate” category.) These cells are characterized, relative to the average expression profile of the general circulating monocyte population, in the transcriptome below. You won’t find these cells in the Beijing YouAn team’s flow cytometry data; but there’s a likely explanation for that, given after my attempt at clinical interpretation.
The “enhancement” is the use of the generally available immunostimulant plerixafor (Mozobil) 4–6 hours before blood donation, which boosts the abundance of these cells in circulation by approximately 10x (full standard dose) or 4x (half dose). If this interpretation of the data is correct, there is no need for specialized manipulation of the cells between donation and infusion. The therapeutic action (though not the potential side effects!) will probably be much the same when performed in a tent with an emergency field transfusion kit as in a fully equipped hematologic oncology clinic with all the gleaming capital equipment.
COVID-19 pneumonia patients with a history of cigarette smoking are at the greatest risk of death from cytokine release syndrome. According to this interpretation, the immune systems of patients who exhibit hypoxic pneumonia together with a likelihood of SARS-CoV-2 infection can be significantly helped by infusion of CD34+ boosted donor blood. Screening blood donors for negligible history of tobacco use (and excluding diabetics), and pre-treating them with plerixafor (an approved immunostimulant), has the potential to mitigate the severest consequences of hypoxic pneumonia. This may or may not save COVID-19 patients in ICU from near-certain death in the way that the original study appears to have; that depends on whether the effective mechanism of their therapy really depends on true stem cells. But it’s more likely to be scalable in the near term.
Whatever the Beijing YouAn clinicians actually infused, it worked. The single administration in this study, to patients who already had one foot in the grave, had most of them off of the ventilator in 48 hours and walking out of the hospital (or close to it) two weeks later. The one patient that they let get into critical condition before administering treatment took a little longer, but even so he was headed for a full recovery by the end of the study. The three controls were dead, in acute respiratory system distress, or stable but still in the hospital.
The original, scientifically published protocol (with which I had no involvement) at http://www.aginganddisease.org/article/0000/2152-5250/ad-0-0-216.shtml rescued every COVID-19 pneumonia patient who received it in the ICU of Beijing YouAn Hospital.
It appears to me that the data gathering in this paper is in good faith. (There are clearly problems with control selection and lack of randomization; but unless someone had a crystal ball that picked out seven ICU patients on the verge of a turnaround that is far from characteristic of this illness, outcomes like this are either real or outright fraudulent.) Though also in good faith, the interpretation section reads to me as… fanciful. However, reading between the lines of their procedural descriptions, I see a quite different interpretation whose mechanism is a great deal more straightforward.
If my interpretation is even approximately correct, similarly dramatic results may be reachable using fully differentiated cells (variously known as circulating (myeloid) angiogenic cells, CACs, MACs, or “early” endothelial progenitor cells) that are normally mobilized from bone marrow into the bloodstream in response to vascular damage. These cells are a million times more abundant in the bloodstream than true endothelial stem (colony-forming) cells — let alone mesenchymal stem cells, which are rarely if ever reported to be found in circulation in adults. The catch is that these cells need to come from an outside donor, because the patient’s own MACs are being destroyed, or at least severely impaired in function, by SARS-CoV-2 infection as rapidly as they are mobilized. The donor cells need not be COVID-19-resistant in an immunological sense, and it is irrelevant whether the donor has previously fought off COVID-19.
(A liter of adult human blood normally contains about 7,000 million white blood cells, of which about 40 million, or about half a percent, are MACs. About 50 are true ECFCs, according to the estimate in the paper that introduced plerixafor for this purpose. Yes, that’s only about 250 circulating ECFCs in an entire adult bloodstream. A dose of plerixafor increases the abundances of both MACs and ECFCs in circulation about tenfold within 4–6 hours.)
This therapy can reasonably be applied to any patient that is likely COVID-19-positive based on clinical criteria, without waiting for a positive result on an RT-PCR test. SARS-CoV-2 tests are great for epidemiology, but their absence is no obstacle to immediate diagnosis of COVID-19 pneumonia given the clinical presentation. As reported (secondhand) by a New Orleans ER doctor (personal communication):
CXR — bilateral interstitial pneumonia (anecdotally starts most often in the RLL so bilateral on CXR is not required). The hypoxia does not correlate with the CXR findings. Their lungs do not sound bad. Keep your stethoscope in your pocket and evaluate with your eyes and pulse ox.
…
Basically, if you have a bilateral pneumonia with normal to low WBC, lymphopenia, normal procalcitonin, elevated CRP and ferritin — you have covid-19 and do not need a nasal swab to tell you that.
What the clinical study appears to report
Let’s look at the clinical sequence for the critically severe case (Case 1), which is narrated in detail in Leng’s Supplementary 3 (Supplementary Materials are available at http://www.aginganddisease.org/fileup/2152-5250/SUPPL/20200229073402.pdf). This patient was admitted to the ICU and put on supplementary oxygen on January 30th, treated on January 31st, and rolled out of ICU on February 5th with no remaining fever or shortness of breath.
I’m not a clinician, but I don’t think the mechanism of treatment here could have had much to do with fighting the virus directly. First, read the narrative about the critically ill patient in Supplementary 3 and look at the CT scans in Figure 2. Think about the timeline. Peak pneumonia symptoms were on or about February 2, two days after MSC infusion. CT on February 9th looks pretty similar to CT on January 30th; that’s about 4 days continuing downhill from the point at which they put him on supplemental oxygen, 6 days improving till they took him back off. That’s a more or less normal pace of recovery after turnaround for viral pneumonia.
Table 4 of Leng provides partial blood testing results throughout this interval. There are several things that jump out at me. The most remarkable is that C-reactive protein, which was at insanely high levels (105.5 ng/mL on the 30th, 191.0 ng/mL on the 1st), dropped to 13.6 ng/mL by the 4th and then plateaued, still at 10.1 ng/mL on the 13th. So I think that tells us that the treatment resolved the extreme systemic inflammatory response within about 3 days, while the viral infection didn’t clear for another week or longer. (RT-PCR still showed viral load as of the 6th.) Neutrophil counts were elevated throughout the decline, turnaround, recovery period. The cytokine storm mechanism didn’t operate through them, but the response to viral infection did.
The blood chemistry shows severe elevation of creatine kinases, dropping abruptly between February 2nd and 6th (why didn’t they measure it on the 4th?). There is no comparable rise in troponin. So that’s likely to reflect inflammatory muscle damage, but not cardiac or skeletal muscle; specifically smooth muscle. Mild AST elevation, but nowhere near in the same league. Again, I’m no clinician, but that sounds a lot like vascular damage to me. And like the C-reactive protein levels (in which an acute elevation usually indicates hepatic response to inflammation), we see a much more abrupt pattern of recovery in CK than the CT/neutrophil/bilirubin data shows.
That kind of acute C-reactive protein spike appears to be associated with IL-6 stimulation of hepatic CRP production. The smooth muscle layer of blood vessels dumps IL-6 into the bloodstream in response to vascular damage, and specifically in response to hypoxia; this response is especially potent in the smooth muscle cells of pulmonary arteries: https://respiratory-research.biomedcentral.com/articles/10.1186/1465-9921-10-6 . IL-6 secretion patterns in alveolar cells themselves appear to be rather more complicated: https://www.jci.org/articles/view/117392. (I don’t have the training to know whether other axes of CRP provocation might be more significant here; and exact identification of the mechanisms after the fact, from data this coarse, is neither likely nor necessary. It wasn’t the clinical signs of vascular damage that led me to the hypothesis below, but as far as I can tell, this data at least doesn’t contradict it.)
Why to consider disregarding the clinical study’s interpretation
None of this response to treatment is a plausible outcome of infusing mesenchymal stem cells. It doesn’t even appear to follow from repair of alveolar or vascular damage; the patient was still on supplemental oxygen to address patent hypoxia until February 9th. But abrupt arrest of early cytokine release syndrome is a plausible outcome of infusing MACs — cells that mediate paracrine signaling which quells the inflammatory call to action at its source. And historically, these cells have constituted the great majority of freshly harvested “stem cell” fractions, whether derived directly from bone marrow or from circulating cells via leukapheresis. These misidentified “early EPCs” are fully differentiated and don’t last in culture beyond 4 weeks. Quoting from the abstract of the 2003 paper that identified them as a separate population (https://www.ahajournals.org/doi/10.1161/01.ATV.0000114236.77009.06):
We cultured total mononuclear cells from human peripheral blood to get two types of EPC sequentially from the same donors. We called them early EPC and late EPC. Early EPC with spindle shape showed peak growth at 2 to 3 weeks and died at 4 weeks, whereas late EPC with cobblestone shape appeared late at 2 to 3 weeks, showed exponential growth at 4 to 8 weeks, and lived up to 12 weeks. Late EPC was different from early EPC in the expression of VE-cadherin, Flt-1, KDR, and CD45. Late EPC produced more nitric oxide, incorporated more readily into human umbilical vein endothelial cells monolayer, and formed capillary tube better than early EPC. Early EPC secreted angiogenic cytokines (vascular endothelial growth factor, interleukin 8) more so than late EPC during culture in vitro. Both types of EPC showed comparable in vivo vasculogenic capacity.
This is important because there is otherwise a pretty serious problem with my interpretation. The clinicians in the Beijing YouAn team had no part in the preparation of the cell culture labeled “mesenchymal stem cells”, which could have had practically anything in it given only what we know from this paper. The extent of their published “cell preparation” method is rather minimal:
The clinical grade MSCs were supplied, for free, by Shanghai University, Qingdao Co-orient Watson Biotechnology group co. LTD and the Institute of Basic Medical Sciences, Chinese Academy of Medical Sciences. The cell product has been certified by the National Institutes for Food and Drug Control of China (authorization number: 2004L04792, 2006L01037, CXSB1900004).
I don’t know whether there’s any means of verifying what’s in a “cell product” based on its NIFDCC “authorization number”, though I can guess that “mesenchymal stem cell” harvesting/culture protocols from 2004/2006 are likely to be less precise than current practice in many immunocytology labs. But I can observe that the prevailing cultural practices in parts of the stem cell research community, especially those that procure their cell cultures from outside their lab, still seem to use terms like “MSC” or “BM-MSC” or “EPC” to refer to what are fairly obviously heterogeneous cultures. Clinically focused researchers also often omit to run the cell product through flow cytometry or mRNA assay before starting their investigative protocol. (This is not just a China issue; see, for instance, the otherwise quite careful-seeming paper at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4736322/.)
However, the Beijing YouAn study authors (or, more likely, some of their co-authors, probably recruited to help after the clinical success was evident) did perform a careful analysis of their cell culture, seemingly after the fact (they don’t say so explicitly, but it’s implicit in the sequence of their narrative). They ran it through flow cytometry and single-cell RNA sequencing, resulting in a credible report that it was dominated by cells that are CD45–/CD14– (and therefore not differentiated hematapoietic cells) and also CD34– (and therefore mesenchymal rather than endothelial). The latter is attested only through an assertion about scRNA-seq results, but the culture protocol typical of attempts to produce “mesenchymal stem cell” cultures routinely selects for CD34– cells (presumably through the “adheres to plastic” criterion dating back to their identification by Friedenstein et al. in 1976).
Is this fatal to my interpretation? Frankly, I ignored their cytometric data at first, because I was focused instead on finding a plausible mechanism for the clinical results. And then I misread it, resulting in a howling error of misinterpretation. And then I looked harder at how stem cell culture protocols work, and found that there are working scientists who are seriously questioning some of these CD markers anyway. CD34 expression, generally considered a marker for hematopoietic progenitor cells and their progeny in vivo (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3578962/), is known to inhibit cell adhesion and induce cell rounding; see https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3903688/. It also appears to be at least partially reversible, and perhaps to be expressed differently in open circulation and in culture than in the hypoxic environment of the bone marrow.
I can’t say whether the skepticism in https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3846603/ towards CD34 as a reliable indicator of endothelial vs. mesenchymal stem cells is warranted, or whether to heed the call in https://www.the-scientist.com/news-opinion/call-to-stop-using-the-term-mesenchymal-stem-cell-64862 to abandon the entire term “mesenchymal stem cell” as a category error. (For a survey of literature that relates MSCs to different types of EPCs, see section 5 of https://www.hindawi.com/journals/sci/2018/9847015/.) But I can say that if their culture, like most historical “endothelial progenitor cell” cultures, started out dominated by fully differentiated cells, they would probably not have survived the extra 2–3 weeks and 3 culture passages that appear to have ensued between clinical infusion and flow cytometry. (A culture passage involves extracting a fraction of the remaining living cells from a parent culture and reinoculating them into fresh culture media.)
The mesenchymal stem cells were cultured in DMEM/F12 medium supplemented with 2% FBS, 2% GlutaMAXTM-I, 1% antibiotics and 2 mM GlutaMAXTM-I at 37°C with 5% CO2. After three passages, MSCs were immune-phenotyped by flow cytometry for the following surface markers: CD105, CD90, CD73, CD29, HLA-DR, CD44, CD14 and CD45 (all antibodies from BD Pharmingen, San Jose, USA). And MSCs were tested for adipogenic, chondrogenic and osteogenic differentiation to identify their characters.
It’s even imaginable that they did infuse CD34–CD45–CD14– cells, but that they had the potential to respond to the extreme cytokine signaling environment by differentiating to CD34+CD45+CD14+ MACs or their functional equivalent. In any case, I’m going to focus on the thing that did jump out at me from the beginning in their cytochemical data, which is the claim that their cells were ACE2 negative, and return to building the case for a treatment proposal around “ACE2-negative” donor MACs. (Note that ACE2-negative, in this context, does not imply an impaired ability to produce ACE2; it means that the donor’s cells are not epigenetically primed, by chronic hypoxia, nicotine exposure, diabetes, and/or other factors, to express histochemically visible ACE2 in the absence of acute hypoxia.)
The treatment proposal
The novel proposal is to attempt to obtain a useful volume of CD34+ cells — consisting mostly not of true stem cells, but of circulating (myeloid) angiogenic cells, which are another cell type that may be exceptionally vulnerable to SARS-CoV-2 infection in a hypoxic patient — directly from blood donors with the help of plerixafor. (This is an “orphan” but not hard-to-manufacture drug, routinely used worldwide to stimulate CD34+ cell release in chemotherapy adjuvant protocols.) The donor should be non-tobacco-smoking (preferably with negligible past history of smoking tobacco or anything else); non-hypoxic (not currently suffering from pneumonia or COPD or sleep apnea); non-diabetic; and ideally not taking ACE inhibitors to control hypertension (though this factor appears much less significant than hypoxia, nicotine, and diabetes).
Typical plerixafor administration protocols in oncology settings involve a 4–6 hour delay from subcutaneous injection to the start of leukapheresis. There are reasons to think that this is not optimal given healthy donors. The published data for healthy donors is promising as regards safety, but doesn’t tell us much about optimizing for yield, given that the donors were chosen because they were HLA-identical twins of cancer patients. So given that there’s a need for a larger-scale study of donor safety, donation protocol optimization (starting with first-hand guidance from clinicians, which suggests use of known circadian rhythms in CXCR4/CXCL12 binding strength) can be conducted in that context.
In developing-country settings, leukapheresis may not be available or economical for mass treatment. There’s no reason to insert the leukapheresis obstacle here if enough donors are available, and I imagine that there’s some risk to the donor of depleting bone-marrow-derived cell reserves in the process. A back-of-the-envelope calculation, based on the cell volumes and abundances in the clinical study and in the original plerixafor (AMD3100) trial at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1895468/, suggests that a full dose of plerixafor would mobilize close to three times as many MACs per pint of donor blood as the number of “MSCs” infused per patient in the clinical study. A middle ground might be to administer a half-dose of plerixafor (reported to boost MAC abundance by 4x rather than the 10x of a full dose: https://www.bbmt.org/action/showPdf?pii=S1083-8791%2817%2930979-5) and draw one pint of blood.
The subsequent transfusion requires no intermediate processing and no donor matching beyond ordinary blood typing; any minimally equipped hospital anywhere in the world can perform this procedure, given supplies of plerixafor and non-smoking blood donors. (While whole-blood transfusions are now a rare phenomenon in the developed world, I wanted to emphasize that there are no complicated cell fractionation and preparation and culture steps involved; if you don’t even have a centrifuge or a lab tech who can isolate the buffy coat, you can still do this. I don’t know anything about preparing blood fractions, but I assume that there are non-trivial logistics involved in tracking a plerixafor-enhanced donation through the existing process. And while I think it’s great to separate out the rest of the blood for other uses, it’s not obligatory in an urgency-driven attempt of an unproven treatment.)
The hypothesized mechanism of therapeutic action
The goal is to supplement the recipient’s own immune system with a critical type of “white blood cell” involved in repairing damaged blood vessels (the MACs in the diagram below). If this analysis is correct, these cells are being killed by viral infection as fast as the patient’s own bone marrow can produce them. But donor cells, which are not lit up in the same way with the virus’s chemical attack targets, can do their job well enough to halt the “cytokine storm” that would otherwise send the patient into an autoimmune death spiral.
These circulating (myeloid) angiogenic cells, known as MACs, CACs, or “early” EPCs, are not true stem (pluripotent progenitor) cells; but they do comprise the majority of CD34+ “stem cell” cultures in older studies. (They are similar in chemical phenotype, but not in appearance, to monocytes and M2 macrophages. Note that IL-6 also polarizes macrophages towards M2 upregulation.) For a cogent summary of what they do and how they initially got mixed up with “endothelial progenitor cells”, see https://www.researchgate.net/publication/319836034_Tissue_regeneration_using_endothelial_colony-forming_cells_Promising_cells_for_vascular_repair.
The SARS-CoV-2 virus enters cells by grabbing onto a “doorknob” called ACE2, which is a transmembrane metalloenzyme important in blood pressure regulation. (This is a gross oversimplification, in light of recent developments in understanding tissue-local (pro)renin-angiotensin systems; but it’s sufficient for virological purposes.) Type II pneumocytes in the alveoli express surface ACE2 under hypoxic stress in general, becoming prime targets for viral replication. Surveys of ACE2 expression using antibody stain protocols show some accessible ACE2 elsewhere, particularly in the walls of blood vessels, though non-smokers exhibit far less immunohistochemically visible ACE2 than smokers. This contributes significantly to the demographic pattern of serious COVID-19 illness.
Note that histological studies are not likely to capture the role of circulating cells. The circulating angiogenic cells are ACE2-positive under hypoxic conditions (and in smokers generally), but they probably get washed right out of tissues being prepared for antibody-based immunohistological staining protocols. (And besides, they aren’t gathered there in uninjured tissues in the first place; they only congregate where there’s vascular damage awaiting repair.) So they haven’t shown up in histological studies of which tissues are likely to be attacked by SARS-like coronaviruses, and they’re not widely recognized to be a critical factor in the mechanism of COVID-19 infection and disease progression.
The hypothesized vulnerability of these cells may be also be a factor in the initial SARS-CoV-2 infection of oral / nasopharyngeal mucosa, and may help explain why the authors of an important 2004 post-SARS histology paper (https://www.ncbi.nlm.nih.gov/pubmed/15141377) said they couldn’t explain the upper respiratory symptoms in SARS. (The role of cigarette smoking in ACE2 expression was also not known at that time, and I would consider any histological or epidemiological analysis that does not control for that factor nearly valueless. In addition to the papers I have cited, there’s a wealth of new preprints trying to evaluate whether former smokers still have more ACE2 in places that are relevant to SARS-CoV-2 infection, how much of the effect is due to hypoxia vs. nicotine, how much ACE inhibitors and diabetes matter once smoking history is corrected for, and so forth.)
Independent of epigenetic changes due to chronic hypoxia and nicotine exposure, type II pneumocytes are known to express surface ACE2 under acute hypoxic conditions, such as those resulting from severe pneumonia. In contrast, vascular tissues under hypoxic stress signal for help from MACs, whose role includes the secretion of multiple paracrine signals including ACE2 (or at least a sufficient fraction of the ACE2 complex to bind to the ACE2 antibodies used in immunohistochemical research). The long-range component of the vascular hypoxic stress signal acts on the same CXCR4-based mechanism that anchors MACs and ECFCs to their bone marrow origins that plerixafor does. (This appears to involve microRNA exosomes in a cardiac-specific context, per https://www.sciencedaily.com/releases/2019/03/190313132244.htm; I have so far been unable to determine exactly how this works for general vascular damage.)
Under systemic hypoxic conditions, I hypothesize that a patient’s own MACs express so much surface ACE2 that they are extremely vulnerable to SARS-CoV-2 infection, especially if the patient is a current or former cigarette smoker. In a non-hypoxic non-smoker, MACs do not necessarily express surface ACE2 until after they arrive at the site of vascular injury and respond to the local signals of vascular damage. And while the expression, and even shedding into plasma, of ACE2 by MACs and/or ECFCs is an essential part of the local repair process, it appears to be the MACs’ release of localized, concentrated IL-8 that is critical in the context of acute lung injury and microvascular damage. (See https://www.archivesofpathology.org/doi/pdf/10.5858/arpa.2013-0182-RA for the role of IL-8 in recruiting neutrophils and the harm associated with elevated diffuse IL-8 in acute respiratory distress syndrome. For specific details on MACs and IL-8 paracrine signaling, see https://iovs.arvojournals.org/article.aspx?articleid=2360958.)
The short-range signal of vascular damage, which recruits MACs and (to some degree) true EPCs to the site, appears to be hypoxia-inducible factor 1-alpha (or perhaps something whose production it triggers). See the “acute lung injury” section of https://www.nature.com/articles/s12276-019-0235-1 and the discussion in https://www.ncbi.nlm.nih.gov/pubmed/26880989. Note also the role of CXCR4 in type II pneumocyte spreading after acute lung injury. Details are in https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5530913/ (ATII = type II pneumocyte). These are reasons not to want to gum up CXCR4 receptors by administering plerixafor to the actual pneumonia patient, along with the likelihood that it is wholly redundant to endogenous CXCR4 blockade effects (and may even exacerbate incipient cytokine release syndrome).
Clinicians in US hospitals (New Orleans, personal communication) report that it is not the direct tissue damage done by SARS-CoV-2 infection that is killing COVID-19 patients; it is the systemic inflammatory response. Under this hypothesis, since the MACs aren’t surviving long enough to do their job (which is to act as the “traffic cop” at the site of vascular damage), the patient’s immune system goes into overdrive (cytokine release syndrome) and kills the patient by mistake. MACs recently transfused from an ACE2-negative donor (or freshly differentiated from ECFCs transfused from an ACE2-negative donor?) are not particularly vulnerable, because they don’t start expressing the virus’s target signal until they arrive where they are needed.
There is compelling evidence (S. Joshi et al., J Cell Physiol. 2019, https://onlinelibrary.wiley.com/doi/epdf/10.1002/jcp.28643) that true ECFCs shed ACE2 under hypoxic conditions. Zikuan Leng et al. report that their MSC culture was ACE2-negative at the time that it was histochemically analyzed. Evidence from a murine model (Yaqian Duan et al., Circulation Research 2019, https://www.ahajournals.org/doi/10.1161/CIRCRESAHA.119.315743) suggests that ACE2-capable MACs are essential to maintenance of vascular walls, and evidence from studies of type II pneumocytes in a murine model (B. Uhal et al., European Respiratory Journal 2013, https://erj.ersjournals.com/content/42/1/198) shows that ACE2 expression is downregulated in these cells during cell proliferation. So while the evidence that infusion of ACE2-negative MACs can have a significant effect on clinical outcomes is indirect at best, there are multiple mechanisms by which donor-MAC-induced stimulation of replication in the patient’s own ECFCs might be expected to mitigate vascular damage and the associated cytokine-storm-provoking signals.
MACs can’t differentiate into vascular endothelial cells and take up residence in blood vessel walls. But it’s not at all clear to me how donor ECFCs or MSCs could have done so either. MACs can stimulate recruitment, migration, and proliferation of tissue-resident ECFCs. Unfortunately, the Beijing YouAn study isn’t going to be a direct guide to whether this effect is enough to help patients or how large a quantity of MACs would be needed to accomplish a rescue. So dose estimation for any therapy that may arise from this hypothesis is less crisply determined than I originally hoped. (A pint of donor whole blood 4–6 hours after a full dose of plerixafor may be expected to contain approximately 2 × 10⁸ MACs and 250 ECFCs, versus the 6 × 10⁷ MSCs reported to have been infused into the critically ill patient who recovered in the Beijing YouAn study.)
If donor MACs can help a COVID-19 patient, it won’t be by actually fighting the virus. They would not take up residence in the patient’s tissues and multiply and differentiate, as true colony-forming cells would (which is why true “stem cell” treatments must generally use self-donated cells or be accompanied by dangerous immunosuppressants or complex immune-marker-stripping techniques). MACs could simply mediate the repair of localized vascular damage, and more importantly, shut off the systemic inflammatory signals that are driving the cytokine storm. The patient’s own immune system—which generally responds effectively to the viral infection as long as the patient doesn’t die first—would have to do the rest.
Plerixafor, aka Mozobil, is currently expensive (about $9,000 for the typical dose used in the oncologists’ leukapheresis protocol) and produced in low volume in the Western world; but production could probably be ramped up very quickly. I did some patent searching with the help of Google Translate, and I think there’s at least one mid-sized Chinese pharmaceutical manufacturer that could scale up plerixafor production cheaply. See https://patents.google.com/patent/CN102653536A/en; I’m working on reaching the inventor to discuss feasibility. This is important because, while it’s currently an expensive drug, I think that’s entirely about recouping R&D costs. Not that there’s anything wrong with recouping R&D costs! But unit manufacturing costs and R&D costs have different implications for the potential scalability of a treatment.
This treatment, if it works, isn’t just for people who can access experimental treatments in stem cell research hubs. With an ordinary degree of coordination among governments — or, frankly, even in the face of active government opposition and disinformation, as long as NGOs and hospitals are determined and non-smoking blood donors can be recruited — this is a treatment that could be administered successfully, if it works anywhere, in any minimally equipped hospital anywhere in the world.
As I learn of flaws in this analysis, I’ll make every effort to correct the resulting misinformation, and to ensure that an up-to-date version is accessible through https://bit.ly/Plerixafor-COVID-19 and (just in case a problem develops with that link) https://bit.ly/CD34-Transfusion. (I’ve received some much appreciated guidance from Dwight McKee MD, CNS, ABIHM, Integrative Cancer Consulting, Aptos, CA. Many thanks also to Y. Jarajapu for gently directing me to observe errors in a previous version, and to D. Motyka for very helpful advice as to what busy clinicians are likely to want to know first, which I have taken only in part. None of the above have endorsed anything appearing in this essay, and all errors remain mine.) In the meantime, perhaps readers could pass this to people who are in a position to refute it, or even maybe give it a try.
I can be reached at this email address (m.k.edwards@gmail.com) or by phone/text at 831–239–3341.
Michael K. Edwards
Santa Cruz, CA