Why Rituxan takes so long to kill Waldenstrom's macroglobulinemia

Posted to IWMF-Talk on 1/20/11

The question of why Rituxan can in some people take many months before seeing results is a very interesting question and I don’t think anyone has a good explanation. The most obvious guess is that the IL-6 production stimulates the IgM production and so it "masks" the true impact of the treatment for months.

We do know a lot of bits and pieces that might explain what is going on.

Here are some things we know:

1) In imaging studies in mice, sufficiently high doses of Rituxan kills virtually all the WM cells in the marrow in a matter of days. This was done by imaging studies in rats.

2) In the same mice studies, using a lower dose had virtually no effect on the WM.

3) Rituxan triggers elevation of IL-6 which causes increased IgM production (the "IgM flare") that can last for months. Therefore, the “true results” of the killing on infusion day can be “masked” from us for quite some time

4) I don't know what the half life of Rituxan is on its own in the body. The Wikipedia page lists 30 to 400 hours but that is much more likely the result of the body clearing the Rituxan, rather than how long the antibody itself lives for. Does anyone know this number?

5) Rituxan is cleared from the body at a pretty fast rate (exponentially as you'd expect). In 21 days, 75% of the stuff they put in you has been filtered out by your body. You can see the curves here if you look at Figure 1: http://www.skirsch.com/wm/wmLowDoseRituxan.htm  What that means is that if you are trying to use the least possible drug and achieve a given minimum blood concentration, you should use frequent small doses rather than infrequent large doses.

6) Ofatumumab (similar to rituxan) has no incremental effect after a certain body concentration is reached. There is a range of concentrations where it is linearly effective. Above 10ug/ml, there is no value in increasing the drug. I have not seen similar graphs for Rituxan. Has anyone?

7) It is not well understood which killing mechanisms Rituxan primarily uses (ADCC, CDC, etc). We know it is a combination of mechanisms, but I think there is still some disagreement as to the primary mechanism of action.

8) It is likely the case that one of the main mechanisms through which Rituxan works is in the recruitment of Natural Killer (NK) cells to aid in cell death. But our body only has a finite amount of these and if you deplete them, then the efficacy of your Rituxan infusion is compromised by some unknown amount. Once your NK cells are depleted, it takes about a month to fully regenerate them. This means that it is may not be a good use of drug to start with Rituxan doses every week if you have all of your B-cells intact since it's possible that you'll use up all your NK cells in destroying your good B-cells before you even get to your WM cells in the marrow. It sure would be nice if someone measured CD-20 positive cell population over 4 weeks with a single 10mg dose of Rituxan and then tried a 50mg dose to compare it (adjusting for body mass of the recipients). You'd have to do this in different people since you'd want people who have not been previously treated with Rituxan (or allow a year between treatments if in the same person to allow the B-cells to regenerate). As far as I know, the data (amount of Rituxan needed to wipe out your good cells) isn't available which is surprising to me, but if someone knows this, please let me know! Once we have that info, we'd know how much drug to give people to wipe out all the good CD-20 positive cells so that we can then give the body a month to recharge the NK cells before we attack the remaining WM cells in the marrow.

9) The dosing of Rituxan was derived from 500/4*3 which = 375. It's a long story and the numbers were determined by the FDA telling the docs how to do their clinical trial and the numbers had nothing to do with maximizing effectiveness. Basically, drugs are given to you based on Maximum Tolerated Dose (MTD) under the assumption that more must be better. That is not always the case (although it is often the case). And even if it was the case, in a lot of patients who have infusion reactions (or post infusion reactions), using the standard dose may not be the best way to treat the patient (this is, of course, really hard to determine).

10) There has been little to no work on optimizing the dosing of Rituxan and virtually no interest on doing a clinical trial to understand efficacy vs. dose. With a drug this popular, you'd think things would be done to maximize effectiveness and minimize side effects so that more people could benefit with less risk. Virtually everyone gets the standard dose.

11) It could be more effective to wipe out the good CD-20 cells and then allow time to recharge your NK cell population to repopulate before you start a chemo round. In general, you want to minimize the number of chemo rounds you have to go through. So why not start it after you've cleared away the "decoys" (i.e., the good CD-20 positive cells in the peripheral blood)?

12) Once you deplete your good CD-20 cells with Rituxan, it takes about a year for them to regenerate back to normal (assuming you stop all Rituxan). This means that once you wipe out the peripheral cells, things are pretty clear for the Rituxan to go after the harder to reach (and better protected) WM cells in the bone marrow. The Wikipedia page on Rituxan specifically notes that Rituxan works great on peripheral B cells but hardly works at all on cells in the marrow which is where we want to do the real killing. So there are really two phases in your Rituxan treatment: phase 1) knock out all the peripheral B cells and Phase 2) go after the cells in the marrow. You have to do phase 1 before phase 2 since you'll never make it to the marrow if there are good B-cells around that will attract the Rituxan.

13) Rituxan maintenance infusions have proven to be effective, although there is no consensus as to the best dosing schedule for this. Several schedules seem to work about equally well.

14) The size of the flare seems to be dose dependent. The higher your Rituxan dose, the larger the flare. I tried this on myself with a much smaller Rituxan dose and got hardly any flare. This wasn't surprising to me that the flare would be proportional to the dose.

15) Consider all of the above my personal opinion and observation. Others may disagree. The talklist welcomes sharing different points of view.

So I'd love to know the answers to some of the questions posed above, e.g., I'd love to know what the concentration vs. efficacy graphs for Rituxan look like (where efficacy is measured in killing in the marrow after the peripheral B cells are eliminated). I'd love to know how many mg/m2 you need to deplete your good CD-20 cells in the peripheral blood and what period of time is required to do that. Can you do it with a single standard dose? Multiple smaller doses? Can it be done in 30 days? Etc. It would seem that this wouldn't be hard to measure and should be a starting point before initiating any chemo/Rituxan therapy.

 The CD20 positive cells that Rituxan attacks do not generate most of the monoclonal IgM we see in our blood.  The IgM is, for the most part, generated by WM plasma cells which are CD20 negative.  The CD20 positive cells differentiate into CD20 negative plasma cells and the WM plasma cells are relatively long-lived.  Thus, even if the Rituxan wipes out most of the CD20 positive cells, the IgM will continue being generated at virtually the same rate until the WM plasma cells die off. 

I'm not an MD, but this has been the general theme I think I have seen expressed in some of the literature on WM and also was discussed openly at some of the presentations in the last Forum.  I have also spoken with my Mayo oncologists and the general idea was confirmed. 

 Regarding the flare, my own experience is that I endured six separate IGM flares associated with exposure to Rituxan but ultimately had virtually no response over the long run.  My IgM and BMI stayed effectively unchanged before and after the many Rituxan treatments. My interpretation has been that IL6 increased to cause the flare, in line with what the Harvard studies suggest, and that this increase in IL6 lasted for at least six weeks but no more than twelve weeks.  

Treon once told me nobody knows how long the WM plasma cells live.

The Rituxan response data suggests the answer is probably different for different people, but clearly it is on the order of a few months (per logic below)

The reason chemo gives you a faster response than Rituxan is because they kill both types of cells (child and adult) whereas Rituxan only kills the children.

So the answer to the "delayed" reaction is at least two-fold:

  1. masking by increased IL-6 stimulating IgM production in plasma cells (plasma cells are not CD20 positive)

  2. masking because Rituxan isn't killing any of the IgM secreting WM cells (only their cells that would give to their replacements)

The bottom line is that the killing happens virtually immediately. It's just that it takes it a while for the dust to clear so you can see it if you are looking at IgM.

The key is that IgM is a lagging indicator of progress.

This repeats part of a message posted on Tuesday. Steve and Bob,

This repeats part of a message posted on Tuesday. The best data happens to be older than most. Why best? Simply because there is sufficient information for one to do a reasonably thorough analysis - but the data spans only 35 days post infusion. Sufficiently thorough analysis has not been done in any of the papers that I have seen. You will note that a scientific outcome emerges that argues against both small dose and high dose levels. Again I have not seen any articles mentioning this result...maybe the analysis indeed has not been done? ----------------- When I take a close look at the data published by Maloney and co-workers back when Rituxan was undergoing trials (1994), I find classical evidence that Rituxan reduces concentration in serum of individuals in at least two conventional stages or compartments of mechanism. The first is probably a simple clearance mechanism where material is deposited in various parts of the blood vessels and/or excreted. This has a short half-life of a couple of days. The second mechanism allows the infused drug to persist and decline steadily at much slower rates, the half life being a few weeks. People who showed no response in clinical trials exhibited only the first efficient clearance mode. They did not transition into the slow mode of elimination, but cleared the infused mAb rapidly with half life of a day or so. By extension, it would appear that the durable second pharmacological stage is that in which the desired immunological interactions take place…B-cells and killed and so-on.   The interesting result to emerge from my little analysis is that the half life of each process depends on the administered dose of Rituxan. The kinetics occurring in an individual case depend on the peak concentrations reached after infusion. For 100mg/m2 infusion, the first stage has half life 3.0 days (+/- 0.9 days). It is 4.9 days for 250mg/m2 infusion and 5.2 days at 500mg/m2 infusion. Thus, the half life for initial clearance of Rituxan lengthens (process slows down) as the dose level is increased and settles down to a constant time for the high end of tolerable doses. The predominant therapeutic mechanism is evidently the second. Its half life is 7.2 days (+/-2.3 days) at 100mg/m2 infusion. This increases to 23 days at 250mg/m2 dose. But at very high dose, 500mg/m2, the half life is always short…about 7 days.   Along with all this, the initial serum concentration and the concentration of Rituxan consumed through the durable second stage process vary also with dose level. They are not simple linear functions. Neither increases substantially until the dose exceeds 100mg/m2.  Indeed, the math suggests that the serum concentration available to the therapeutic mechanism doesn’t rise markedly until the infusion level exceeds about 200mg/m2. Thereafter, it rises briskly. Infusion of 100mg/m2 provides 18ugm/ml serum Rituxan to the durable mechanism, 250mg/m2 provides 41ugm/ml and 500mg/m2 delivers 217ugm/ml.

  Now, it is most likely that the benefit of available mAb depends both on its concentration within the serum and the time for which that is present. As an indicator, integrate the decay curve over time from zero to forever (the easiest option). This gives a guesswork number that might tell if an optimum dose level emerges from those old studies. Normalize the result to unity for 100mg/m2 infusion. It has jumped to 7.2 at 250mg/m2 dose, but achieved only 11.6 at infusion levels of 500mg/m2. This rate of gain does not continue at elevated doses. Taking into account the trends, I conclude the most efficient dosing level is in the range from ~250mg/m2 up to ~350mg/m2. As the limit of drug dosage tolerability approaches, the drug is cleared just as rapidly as the body is able – hence it does not get time to work well. Go higher and you do real internal damage to the patient.   I can produce some detailed curves if you want - but they cannot be posted here, of course.   Best, Colin


On 1/20/2011 6:15 PM, William Block wrote: > At a San Francisco event last Fall, I talked with Steve Treon about the rather haphazard way in which 375 m2 came about and asked why there had never been any research on whether a smaller dose would be as effective. His answer was straightforward: The drug companies have absolutely no interest in investingating whether a smaller dose would be as effective, and as their funding is essential to the research, the research does not get done.

Hi Steve, I have read that the rituxan half life is of the order of 15 to 20 days rather than hours. Calculating maximum concentrations If 20 days is used yields a maximum concentration from 375mg/M^2 after 4 weekly infusions at slightly over 2000mg/M^2. That is significantly more than 1 infusion every 3 months would give. I do not have the reference but it would seem possible to monitor this analytically over a course of 4 infusions.

From Colin Perrott
First: my apologies. I hate to do this to you all and I did not set out to be smart here. The subject is evolving as we breathe. BUT the measurements reported by Maloney et al for "non-responders" were for those people in whom there was no observable reduction in the numbers of circulating B-cells as a result of the Rtx. For them decline of serum rituxan level proceeded at a rate that was not measurably different from responders. Yet by contrast, "responders" exhibited almost total elimination of the peripheral B-cell population during this first stage. Where did the rituxan go?

If we think about the behavior of human IgG for comparison, we find long-standing data that shows infused IgG (IVIG) declines in a "biphasic" mode. In the first "alpha" phase, the observable half life is a couple of days. In the second "beta" phase it declines at a half life ranging from 20 to almost 50 days, depending on the base serum IgG level. The serum concentration drops to about 40% of the initial infusion within the alpha phase (about three days total). Where did the IVIG go?

I chose the words "deposited in various parts of the blood vessels" for a reason. In the parlance of cancer therapy, we always talk about killing and death. Man you just gotta smack those cells. Shift your mind to living and birth, to what happens in the placenta, to what happens at epithelium throughout our bodies. Binding of the human IgG via those odd things called FcgRIIB receptors on epithelial cells transfers about half of the total IgG from the blood, where it is generated, into extravascular spaces - in the tissues. There it is protected, persists, and is re-expressed to return to the circulation when needed.

Does rituxan do just the same thing? A few articles say it inevitably does, although no radiolabled studies have been done to prove so (to my knowledge anyway). Some established therapeutic uses of rituxan require the mAb to penetrate epithelial layers. I have never seen mention of such behavior in the context of hematologic cancers. But I really do believe that an extravascular store of rituxan must be created. And only this evening I read a note from John Paasch which indicates that he may well be able to produce hard evidence for just that occurrence. Thanks John, neat work. Then the "alpha phase" for rituxan could amount to potentially retrievable storage of as much as 45 to 80% of the infused dose. But if on the other hand it binds to FcgRIIA, we won't ever get it back. It will indeed have been lost.

a few points.

- in everything i have read, it's the cancerous lymhocytes that proliferate, while the plasma cells (further down the line) do not. when i put this to one of my docs recently he contradicted that and said no, the plasma cells also divide. i didn't feel able to challenge him on that. he could be wrong. is he categorically wrong??

- as a patient who only has insignificantly elevated IgM, but congregating cancerous lymphocytes which form dangerous masses in the body, especially around the spine, i'm personally more interested in treatment that kills off the proliferative lymphcytes than the whole area of tackling IgM and IgM producing entities, be they lymphocytes or plasma cells. to date the treatments i've had have done so, in varying degrees - altho of course without clarity as to which components of the combo therapies i've had have actually been most effective.

- in either case - whether chasing the IgM or the faulty underlying cells - another important factor must be the rate of proliferation. the more indolent the proliferation the better and vice versa - and presumably this rate of proliferation will be another factor that varies from person to person. i believe we get a rough reading of degree of proliferation with our biopsies. mine have been between 1 - 5%.

raphael altman oxford, england

He is categorically wrong.

Well, two things on that. First, the question was - do plasma cells proliferate? The answer is no, they do not. Secondly, there is no such thing as a WM plasma cell, they are malignant lymphoplasmacytoid cells. Plasma cells are normal cells. No cancerous cell of any type in the body is still called by its normal name as they have mutated. So, do the lymphoplasmacytoid cells that are more plasmoid than lymphoid in WM proliferate? I doubt it and the paper you quoted does not seem to answer that. There is no clinical evidence that they do. If they did, rituxan would not work on them as they have lost there CD-20 markers. Some people have to wait a long time for rituxan to work on them. The IgM continues to go down for months after the rituxan has left their blood stream. In all likelihood, that represents the dying off of the plasmoid cells. If those plasmoid cells were replicating, I doubt that we would see that die off.

Let me remind all those who are discussing the lymphocyte vs the plasma cell as they apply to Waldenstroms that if you look at your bone marrow biopsy, you will find the term lymphoplasmacytoid cells. Additionally, WM is a subset of LPL or lymphoplasmacytic lymphoma. This term denotes that the cells are arrested between the lymphocyte and the plasma cell in maturation. The exact point of maturation arrest, however is variable amongst us. Some have cells closer to being a lymphocyte and don't make a lot of IgM per cell. Those arrested closer to a plasma cell make a lot of IgM per cell. What those cells have in common is that they don't know how to die. Not only that, the plasmoid cells have lost their CD-20 marker. This probably affects the time response to rituxan as the lymphoid type cells are easily killed and the plasmoid cells are not. Tom

> > On Sat, 22 Jan 2011 14:02:35 -0000, Raphael Altman >  one of my docs recently he contradicted that and said no, the > plasma cells also divide. i didn't feel able to challenge him on that. > he could be wrong. is he categorically wrong?? > > > Raphael,

> Raphael,

No, I don't think he is wrong. It is never as cut and dried as we sometimes have to make it sound when we describe a more complicate process. However, the B-cell/plasma cell line for WM does appear to end in senescence - and the plasma cells are the last cells in this line. Thus, the number of times that plasma-like cells can replicate is more limited than it is for their B-cell precursors. Some models suggest that the number of maturation compartments needed to sustain our blood cell numbers is around 30 which is far greater than the number of names we give to the different cell stages in the B-cell/plasma cell line. And there is certainly some additional self renewal that occurs besides this at each stage. Let's say that a cell derived from the eternal stem cell can divide only 30 times before the line become senescent. For a perfect clonal expansion (all daughters, no self renewal) this would mean that the daughter cells outnumber the stem cells by about a billion to one (2^30 to 1). Those cells in compartment 30 would outnumber those in compartment 10 by about a million to one... etc...

Analagously, I think that for cancer we have to be careful about just attributing all proliferation to one very specific cell type (e.g., lymph or plasma). In some cancers at least, the only eternally proliferating cell is referred to as a cancer stem cell - which may or may not look like the cells that are observed in highest abundance. The cells that are observed in the highest abundance may still proliferate and differentiate (like they do in WM) but the number of times they can do this is limited and so the cancer cell line dies out if no mechanism (e.g., a stem cell) is present to replace them from below. The so-called cancer stem cell model - is certainly applicable to many cancer types but I have not seen it applied specifically to WM except in vague terms. I have yet to see somebody describe what a WM stem cell might look like - or to really address how or where it might exist in our body.

Still, I think it helps me to consider WM in terms of a cell dynamical model. The cells that are observed to be present in the highest (or at least the most abnormal) abundances are probably not eternally dividing, but are instead the ones that have longer life spans and/or are slower to differentiate than their more normal counterparts. Something in our marrow supplies them to the body faster than our body's other natural control mechanisms can remove them. For some of us, it seems that a temporary (and sometimes almost permanent) balance is struck whereby the disease does not progress because the production rate of new cells becomes balanced - or at least nearly balanced - by the rate at which they are removed. CAM methods may be appropriate in such cases, where the aim is to change the environmental variables that affect the internal cell dynamical balance in positive directions.

For WMers who happen to be experiencing disease progression, however, and for those whose cumulative progression has resulted in symptoms that are intolerable, the time comes to change the balance by promoting removal of the cells that are causing the problem (e.g., most types of treatment). If that balance is simply uncontrollable then the auto- and allo-HSCTs are used to reset everything.


Below is a study involving Fludarabine, but with well written abstract (below) and nice figures and charts and probably a good model for Rituxan, too. Perhaps all should read carefully and make their own minds up what it means. I'm not a WM expert, but these guys are.

Clin Lymphoma Myeloma. 2009 Mar;9(1):53-5. Assessment of bone marrow response in Waldenström's macroglobulinemia. Varghese AM, Rawstron AC, Ashcroft AJ, Moreton P, Owen RG.

HMDS Laboratory, St James's Institute of Oncology, Leeds, UK. Abstract In this study we used bone marrow flow cytometry and immunohistochemistry to evaluate response to fludarabine therapy in patients with Waldenström's macroglobulinemia (WM)/lymphoplasmacytic lymphoma. Responses in serum M protein were typically delayed with a median time to maximum response of 6 months following the completion of therapy (range, 0-18 months). In contrast, bone marrow responses occurred promptly in responding patients such that there were no detectable clonal B cells at the end of therapy in 55% of patients assessed. Persistent monoclonal plasma cells were, however, readily identified by CD138 immunohistochemistry, explaining the persistence of serum M protein in these patients. This simple observation has significant implications for the assessment of responses in WM as well as the design of future therapeutic strategies.

The term lymphoplasmacytoid or lymphoplasmacytic is a pathological term used to describe a lymphoid tumour in which plasma cells can be seen. Many people consider WM to be a disease with features intermediate between those of plasma cells and B cells - I think this assumption is incorrect - when you look at the bone marrow of pts with WM you can clearly identify both B cells (CD20+ CD138-) and plasma cells (CD20- CD138+). The number of plasma cells will vary from patient to patient and I think it is the extent to which this differentiation process occurs is a major determinant of the clinical features seen in the patient - patients with a lot of plasma cell differentiation will likely have high levels of IgM and vice versa.


Steve Kirsch's Waldenstrom's Macroglobulinemia links