Friday, 16 June 2017

Attention Alemtuzumab users: When is an Inhibitory Antibody not an Inhibitory Antibody? When its Neutralizing!!

Free Grant Idea of the Month
#ClinicSpeak #NewsSpeak

If you are using, taking or thinking of taking alemtuzumab you need to read all of this post. At the very least the bottom bit.Tweet, Retweet and Facebook or whatever to your circle of friends

  To continue explaining our recent paper and the last bit for Detective Week, so MD is back in his Civvies next week.

Interpreting Lymphocyte Reconstitution Data From the Pivotal Phase 3 Trials of Alemtuzumab David Baker, Samuel S. Herrod; Cesar Alvarez-Gonzalez,Gavin Giovannoni; Klaus Schmierer
JAMA Neurol. Published online June 12, 2017. doi:10.1001/jamaneurol.2017.0676

Meaning Controlling this B-cell hyperrepopulation after alemtuzumab administration may limit the risk for secondary autoimmunity if administration can be performed safely.

  • Alemtuzumab induces binding (could cause allergic infusion reactions and could stop drug having optimum effect) and neutralizing (could stop drug working) antibodies in most people. 
  • These issues were known, but not reported clearly in pivotal trial results
  • They are a problem for some people and the drug does not work
  • The neutralizing antibodies increase with time and the third/forth injection cycle may be more problematical
  • Should there be a test be repeat infusions?
  • Some people do not respond to alemtuzumab.

We have shown that alemtuzumab blocks tolerance induction and that it does not deplete B cells from the bone marrow or lymphoid tissues well. 

It only takes about 6 days to make an antibody response, with 3-4 days to sensitize a T cell so within a day of finishing the 5 day infusion you could have an anti-drug response because all the immune system is seeing in those first few days is loads of alemtuzumab

We knew that there is an anti-globulin response as it had been reported, as had ways to reduce them.

Somerfield J, Hill-Cawthorne GA, Lin A, Zandi MS, McCarthy C, Jones JL, Willcox M, Shaw D, Thompson SA, Compston AS, Hale G, Waldmann H, Coles AJ.A novel strategy to reduce the immunogenicity of biological therapies.J Immunol. 2010185(1):763-8.

Those studies talked about the high incidence of antibodies that bound to alemtuzumab. This did not specifically mention neutralizing antibodies that would block alemtuzumab function,

Normally this type of information is published in pivotal phase III antibody trials.

However, anti-alemtuzumab antibodies only received cursory mention in the pivotal trial reports, save to say they were of no importance. Neutralizing antibodies other than those to beta interferon (control group) where not mentioned

So it made us do more digging

The Prof Coles had done an open access paper that you can all read.(CLICK ON ME) Alasdair J. ColesAlemtuzumab Therapy for Multiple Sclerosis Neurotherapeutics. 2013; 10(1): 29–33.
There is a section:


"Despite being humanized, alemtuzumab induces anti-alemtuzumab binding and neutralizing antibodies in as much as 30 to 70 % of patients 1 month after the first and second cycles, respectively. However, because the interval between treatment cycles is at least 1 year, such antibodies usually become undetectable before the next cycle [24]". 

These were not mentioned in the pivotal trial report. Why not?

So 30% after 1 month.  Is this right? 

Well let's look at the data.
Someone shouldn't have prompted us to hunt, by saying they don't exist. They do! 

You don't have to do a freedom of Information to find the information it can be found here.

CLICK HERE) Turn to page 34/35

It says on page 34

"The majority of patients (691/811, 85.2%) treated with 12 mg/day alemtuzumab in the pooled Phase 3 studies tested positive for anti-alemtuzumab antibodies. Of the 85.2% of patients who tested positive for anti-alemtuzumab antibodies at any time-point during the course of the study, 92.2% (637/691) tested positive for inhibitory antibodies".

Therefore, that makes about 78% of people having neutralizing antibodies to a humanised antibody that depletes T and B cells. This is staggering. Chimeric mouse human anti CD20 only makes 15-25% of people develop binding antibodies in the same time frame.  Humanized anti CD20 induced less than 1% neutralizing antibody response according to data presented at ECTRIMS

Now turn to page 35
So at 1 month after first injection there is a whopping 62.4% of people making binding antibody and of them 86.8% make inhibitory antibody so 86.8% of 62.4% = about 54%, so 30% is a bit low.
At month 13 after the second injection it was 94.3% of 83.2% = about 78%
However, without access to the actual individual data points this is just an estimate.

These were not mentioned in the pivotal trial report in a way that explains this. Why not?

What happened in the phase II?

Coles A. J., D. A. Compston, K. W. Selmaj, S. L. Lake, S. Moran,
D. H. Margolin, K. Norris,P. K. Tandon,CAMMS223 Trial Investigators
. 2008. Alemtuzumab vs. interferon beta-1a in early multiple sclerosis. N. Engl. J. Med. 359: 1786–1801

In the phase 2 study in patients with early relapsing–remitting multiple sclerosis, alemtuzumab was given as annual cycles. "Most patients’ anti-globulin responses had resolved prior to the second and third exposures. They reported reported that only 1 of 208 (0.5%) and 51 of 194 patients (26.3%) had significant alemtuzumab-binding Abs (above a pre-specified threshold of 2000 U/ml) at months 12 and 24". 

So sounds like not much of a problem but reports in rheumatoid arthritis indicated there was over 60% of people making antibodies within twelve months, so is this a fudge by the phase II investigators to downplay the antibody response?

Does it tell us that thresholds above 2000U/ml are important and if so important for what? Adverse events? Blockade of Function?

It was not mentioned, how many people had a low titre below 2000U/ml, maybe we should have guessed it must have been 50% of the sample as the figures for the phase II are 29.3% verses 75.4% in the phase III. Furthermore no mention of neutralising antibodies.

Now, you may argue that the data is there for all to read...but it is clearly not well known otherwise I would have not been told that neutralizing antibodies don't exist. 

A humanized antibody creating neutralizing antibodies in about 70-80% of people, tells you there is something weird going on. Alemtuzumab was the first humanised antibody and this was designed to reduce immunogenicity., but it must be a relative effect as it is clearly immunogenic.

Anti-Drug antibodies may be consistent with alemtuzumab blocking immune tolerance, as we have reported (as a cause of secondary autoimmunities) this week.

We thought it was surprising that this information has been omitted from the phase III write up. 

But the information could be found.
Those EMA chaps, called it neutralizing (pg 89) so Google search picked up a search for "alemtuzumab" and "neutralizing antibodies" 
Missed the stuff of (pg35) about "inhibitory antibodies" completely. 

Why the name change of a very well-defined term?   

If you look at the table above. What does a titre of 6.5 million mean?

In the table we can see levels of antibody at up to 6.5 million
Titre = the concentration of an antibody, as determined by finding the highest dilution at which it is still able to cause agglutination of the antigen
So it means someone made a shed load of antibody against alemtuzumab!
It would take about 6 days to make an neutralizing antibody from scratch, about 3 days to make the T cells and then they help the B cells. So first set of doses taking 5 days it would be no problem. It would have killed the T and B cells before it could be stopped. 
The second cycle would be OK too because only 0.6% (5/789) of people would have neutralizing antibodies present at the time of infusion but if there is a need for a third infusion which occurs in about 50% of people then about 31% (239/764) have pre-existing neutralizing responses and this could cause problems. However the level (titre) many be low so it is not devistatingly important. Also 75% of people have pre-existing binding antibodies and this could lead to allergic reactions. In the reports I only found 1 case of anaphylaxis although 239 people had pre-existing antibody. So the risk is low. However once generated antibody responses will persist
for life and may be boosted each time you get a new course.
It is published that alemtuzumab does not block vaccine responses and so plasma cells and the T cells that help them (T-dependence) make antibody, persist.

McCarthy CL, Tuohy O, Compston DA, Kumararatne DS, Coles AJ, Jones JL.Immune competence after alemtuzumab treatment of multiple sclerosis. Neurology. 2013;81(10):872-6

So as you can't have it both ways once the antibody forming cells are created some must survive to become active the next time it sees its target and for people with alemtuzumab making antibodies after the first round of injections, the next time is day 1 of the second set of infusions by day 2 those plasma cells will have woken up. Hopefully not enough by day 3 to stop the antibody depleting.

In the CARE-MS  documents we received, some effort is made to show that the presence of binding or neutralizing antibodies do not influence depletion, efficacy or infusion responses, we didn't get to see information presented on the individual responses. However not surprising as only 5/789 had pre-existing neutralizing antibodies and so not enough to make a conclusion.

What could this all mean? 
You could get infusion reactions (anaphylaxis as an extreme allergic response. This occurs as a small fraction), but this would be treated with steroids and anti-histamine, which is standard treatment against the infusion reactions that occur in 90% of people anyway. 

So the influence of binding antibodies and anaphylaxis potential is masked.  

However 75 % of people will have circulating antibodies at the end of second infusion cycle and 30% have neutralizing antibodies, so the risks of not depleting on the third cycle is increased. 

Based on what we know this is not a reported problem for the treated population but at the individual level it would be important to see.
A neutralizing antibody response could mean that the drug stops working. Based on that reported, it is said that they were not a problem in the trial. Maybe that is why it wasn't mentioned in the pivotal trial papers..which every one would read if interested in using alemtuzumab.
So as a group it is not important, but it may be important to an individual. 

This is the important point
So if people are in your care or you are taking alemtuzumab the question is are they still depleting and is the depletion maintained? 
What if they are treatment failures and need another dose, about 50% people do. With three treatment cycles the data says that many people do very well so again on mass the drug is working
But what about the people needing 4 or 5 cycles? 

Is the antibody depleting? 
If you know about neutralizing antibodies, you could get them checked or switch treatments if it is not working.

So are neutralising antibodies important?
It is said that in the EMA report that the presence of antibodies are not important in the two year trial on a number of outcomes, in terms of efficacy and safety at the population levels, meaning no eyebrows were raised.

Let's go back to the paper at the top and see what was said in the rest of the immunogenicity section.
"If persistent, neutralizing antibodies become problematic in patients who have received multiple alemtuzumab cycles. We have shown that pre-treatment with an altered version of alemtuzumab, which no longer binds to its target, can effectively induce tolerance to alemtuzumab itself [24]".
So if persistent there is a problem so this issue appears known but again not published properly. Why was this not properly exposed, to show data? A couple of lines in a review is not exposing the problem. 

Am I being a DramaQueen?  If you know the about information above I am a DramaQueen, if not an Investigative Journalist.

Now you can think about this issue if you are treating people or if you are taking the medication.

Now you can say, look at the label

Go to the EMA and you get

It says I'm being a drama queen because such a high level of neutralizing antibodies is misleading. 

Best get your info from the FDA

See. It was there all the time so I'm not revealing anything new.

You just have to look

As it is Friday and you get supplements on Friday, I'll give some extract bits of detective work done since submission, not in the current paper. 

Why do this? I could and perhaps should wait and publish it first.

I do this because I think, you need to know now and shouldn't wait until I publish it (or someone else publishes it). As it may be relevant to you. I know there are people in the World, where this is indeed relevant and the neuros will have a dilema of what to do next.

Are the Anti-Drug responses Important at the Individual level?  Simple answer is Yes.

At the population level, I agree these antibodies have no effect within the trial, however back to the data, I can see two things.

(a) There are people who do not deplete well , independent of anti-drug antibodies

(b) There are people who  do not deplete well because they make anti-drug antibodies.

Looking through the EMA data, since we submitted the paper above, I can find an example of each of these (written and ready to go:-). Now I can't say they didn't deplete because I didn't get the before and after cell counts (I have asked), but one can say they didn't deplete very well. Therefore, there appears to be non-depletors following treatment with alemtuzumab. 

So £28,000 (in the UK) for something that does nothing 

Below is a graph found on the internet that told me to look at the data and told be where to look in the thousands of pages within an EMA document. 

Only the first thirty days are the important period when the cells are being deleted. We did not see it in the EMA stuff we looked at, but this is from the EMA website. (Check it in the document below)

Lemtrada, INN-alemtuzumab - European Medicines Agency 

When you are looking for a needle in a haystack. Graphs are what you need. But raw data is what we need. I have drawn some boxes to help you focus on the points I'm making.

You only look at the first 30-40 days because repopulating B cells skew the data after that.

Is the blue above the red? If it is, does it mean being having anti-drug antibody is important to individuals? They are not depleting as well as those without antibody.

Are there non-depletors?, If there are, is it one in a blue moon?
The graph above suggests there are non-depleters. You can see it is not common. 

I'm not much of a human genome mapper, but there are people out there who are and one could easily look to see how common polymorphisms are to CD52, and other gene products that affect lymphocyte depletion. 

CD52 is a peptide of 12 amino acids, anchored to glycosylphosphatidylinositol (GPI). Since it is highly negatively charged and present on sperm cells and lymphocytes, it has been conjectured that its function is anti-adhesion, allowing cells to freely move around

Turns out there are polymorphisms (differences) in the  CD52 molecule

Oko A, Wyrwicz LS, Glyda M, Idasiak-Piechocka I, BiƄczak-Kuleta A, Kaczmarczyk M, Drozd A, Ciechanowicz A, Czekalski S.CD52 gene polymorphism and its potential effect on the response to alemtuzumab in renal transplant recipients.Ann Acad Med Stetin. 2009;55(2):22-6.

CD52 is a small glycopeptide leukocyte antigen present on selected subpopulations of human cells. From the clinical point of view this protein is an important target for therapeutic interventions aimed at leukocyte depletion in hematological malignancies and post-transplant immunosuppression. Recently, two variants of CD52--rs1071849 (A119G; Asn40Ser) and rs17645 (A123G; I1e41Met)--were discovered. We now report on the distribution of these variants in kidney graft recipients and controls. Our bioinformatics findings suggest that CD52 polymorphism may affect the efficiency of GPI anchor formation and thus may indirectly alter the response to anti-CD52 agents.

There are no obvious loss of function mutations (i.e. knockouts) when I looked through over 60,000 genomes from the Broad Institute data base. The polymorphisms above are reasonably common meaning about 30-40% of people would be homozygous for these variants. If it truly affects response to alemtuzumab it will be affecting a lot of people. Is this why the blue dots above are higher than the red dots? 

However, there is only one other relatively common polymorphism besides the two above, that occur as homozgotes and this occurs in Latinos. Do they deplete different from the Europeans. A 5 minute bit of data analysis if you have the raw data 

Is it something else I suspect. We know some gene variants affect, how well you deplete. Do they influence responses to Rituximab, ocrelizumab, etc? 

Based on the total population, we can say that in the first two years of treatment, the antibody neutralisation is not an issue, and published data (ECTRIMS2016) shows you can deplete after three or more cycles of alemtuzumab. This was shown in

Boyko AN et al. Lymphocyte depletion and repopulation is consistent across alemtuzumab treatment courses in patients with relapsing-remitting multiple sclerosis: 6-year analysis of patients from the CARE-MS studies . Mult Scler J 2016
ECTRIMS Online Library. Boyko A. Sep 15, 2016; 146494
In patients who received only the original 2 courses of alemtuzumab, mean CD4+, CD8+ and CD19+ counts at the start of Course 2 were 0.28, 0.24 and 0.34 x 10*9/L, respectively. Both T and B lymphocytes were depleted after the second course of alemtuzumab (N=369), reaching their lowest levels at 1 month after Course 2 (mean CD4+: 0.05; mean CD8+: 0.06; mean CD19+: 0.03 x 10*9/L). CD19+ counts repopulated quickly and reached the lower limit of normal (LLN; 0.1 x 10*9/L) by 3 months. T cells repopulated more slowly; CD8+ counts reached LLN (0.2 x 10*9/L) by about 6 months, and CD4+ counts reached LLN (0.5 x 10*9/L) at Month 24. Once reaching plateau (CD19+, 15 months; CD8+, 30 months; CD4+, 24 months), cell counts were stable through 48 months and remained above LLN. Depletion and repopulation of T and B cells were similar after 1 or 2 alemtuzumab retreatments (Course 3: n=79; Course 4: n=10).

(see it takes 2 years for CD4 levels to return to the lower normal limit so this may be why it takes 2 years for the secondary autoimmunities to occur and you can see the CD8 cells (regulatory cells) are back quicker so the pathogenic cells return in a regulated environment.) 

At the population level there is deletion and activity as can be seen above, but in the responders the CD4 level at 6 months post alemtuzumab is about 0.2 x 10*9, but in the failers requiring another cycle the CD4 level appears around 0.4 x 10*9 in the 6 months before retreatment. Is this relevant?

What is the deletion after cycle 1, cycle 2 and cycle 3 and cycle 4 and how does this relate to neutralizing antibodies? What is the data at the individual level?.  What is the titre of antibody that is clinically relevant for lack of efficacy. It looks like the data is there to address this pretty quickly.

There was no mention of depletion failures in the MS CARE extension abstracts (Boyko et al 2016) or phase II data 

Kousin-Ezewu O, Azzopardi L, Parker RA, Tuohy O, Compston A, Coles A, Jones J.Accelerated lymphocyte recovery after alemtuzumab does not predict multiple sclerosis activity.
Neurology. 2014 Jun 17;82(24):2158-64

An omission or does it simply not happen, that there are no people who do not respond well to treatment? I suspect the occurance of this is low.

However, some people do stop responding to alemtuzumab and the logical reason is due to antibody neutralisation, it happens with all protein therapeutics. Again we have to go no further than the internet and do some digging and some data interpretation. 

We can go back to published data that was used to say that it is safe to go onto fingolimod after alemtuzumab. The lymphocyte number was shown for 13 individuals who failed alemtuzumab and went onto fingolimod. Six people had three cycles and it is clear that the two people in the bottom panels I selected stopped depleting. Probably because of neutralizing antibodies.

These people need to be changed, together they got about £60,000 worth of antibody to do nothing. 

Importantly, it gets even more complicated because after alemtuzumab, it can cause white blood cells to become CD52 negative and so unresponsive to drug.

Brett SJ, Baxter G, Cooper H, Rowan W, Regan T, Tite J, Rapson N.Emergence of CD52-, glycosylphosphatidylinositol-anchor-deficient lymphocytes in rheumatoid arthritis patients following Campath-1H treatment. Int Immunol. 1996;8(3):325-34.

CD52 is a glycosylphosphatidyl-inositol (GPI)-linked glycoprotein expressed at high levels on normal T and B lymphocytes and at lower levels on monocytes, while being absent on granulocytes and bone marrow stem cell precursors. The emergence of CD52- lymphocytes of both T and B cell lineages was observed in three out of 25 rheumatoid arthritis patients treated with the humanized antibody Campath-1H in phase II clinical trial. 

Whereas the majority of CD52- B cells had disappeared from the peripheral blood by 3 months post-treatment, both CD52- CD4+ and CD8+ T cells persisted in the circulation for at least 20 months. In the two patients that were tested, the GPI-anchored surface molecules CD55 and CD59 were also absent on the CD52- cells, although expression of other cell surface transmembrane, proteins (CD3, CD4 and CD2) was unaffected. The CD52- cells maintained a stable phenotype in vitro despite repeated re-stimulation in culture. Both CD52- and C52+ clones, established from one of the patients, responded to a similar extent to several T cell mitogens, as assessed by proliferation, suggesting that a general defect in expression of GPI-linked molecules does not impair T cell activation. These data show that an immune attack against a GPI-anchored surface molecule can result in the selection of a GPI-anchor-deficient cell population. Despite the persistence of CD52- T cells in the peripheral blood, no adverse reactions associated with the presence of these cells were noted in any of the patients; in fact they responded with longer remission times after Campath-1H treatment than the other patients in the trial.

So if MS is a problem of B cells then they will still be dealt with..phew? If it is a T cell issue then 12% of people may not deplete as well as they could. Does this happen in MS?.

This is due to cross linking of CD52 this can be done with alemtuzumab but would be augmented by binding antibodies cross-linking the alemtuzumab binding to CD52. Thus crosslinking CD52

I could have kept this last bit of detective work quiet, so I could publish it first.

But if you are using/recieving alemtuzumab I think you should know and monitor your depletion. If you are not depleting, are you neutralizing the response? and do you need to think about a switch?

Now that the existence of neutralizing antibodies is in the open in the academic literature in a searchable way, there will be a race to publish the first case that neutralizing antibodies actually impact MS at the individual level.

I guess you realise there are probably two places with a head start if they trawl their data, but as you can see it is all there already if you know where to look. (However in the cases above you don't know if they had neutralizing antibodies), but I know it happens.

If you are a non-depletors (on first infusion) maybe get you neuro to give ProfG an email, so we can try work out why, if this not known already, depletion may not be that good. I will volunteer to do this, we have the technology, but we don't have the cash:-(.

It may determine whether you can switch to another IgG1, as we know certain genes determine whether you deplete with rituximab too, or whether you need an non-antibody treatment.

Is this an issue for ocrelizumab? 

This bit of pharmacogenomics may help inform treatment choices.

However, if you are not depleting maybe it is time to change. 

Maybe it is time to Publish the data, so that we don't have to dig to find the information. The answer is there.

CoI: None.  DrK, ProfG multiple


  1. Say an RRMS patient is diagnosed with leukemia: would they hit 2 birds with 1 stone by switching to alemtuzumab (Campath) ?

    Is it the same formulation and protocol that treats both diseases?

    Tony Fonda

    1. The cancer dose is higher and I believed free

    2. Free?

      So a higher dose would hit 2 birds with 1 stone?


    3. Yes 2 birs one stone.
      To stop people buying the CAMPATH-1 cancer vial and diluting it for the lemtrada dose, they removed CAMPATH from the market and repackaged the alemtuzumab to the lemtrada formulation and increased the price twenty times. Bit I think they offer the cancer dose on a compassionate basis

  2. But theres a silver lining in that these non-depleters will have reduced risk of secondary autoimmunity, right?

  3. MD your work on these matters is very impressive and I hope it'll benefit so many more into the future as those dedicated few such as yourself continue to gain improved understanding of this disease and the ways in which DMTs and HSCT work/impact.
    Right now for the embarrassing, I'm so not a scientist, question: If your lymphocytes fall through the floor post 1st infusion you're depleting and don't need to raise any concerns with your neuro - am I correct? Sorry to reduce your incredible post to such a level - but just so as I understand.

    1. Yes correct, first infusion and lymphocytes fall through the floor then everything is going to plan. You should expect to see the same each time you take it

    2. Thanks for all your work on this. Antibodies to Alemtuzumab have been a niggle in the back of my mind ever since I heard of them. This explains it all really well and I find your posting style with different colours/fonts much easier to read than big, uniform blocks. A timely reminder to ask my neurologist what my depletion response to round 2 was.

    3. I think you should look one month after each cycle and if you need third or more cycles then find out your neutralizing status. I am not sure it will predict if you will neutralize, until Genzyme publishes the data it will be difficult to know. I can't force them to publish the data.


Please note that all comments are moderated and any personal or marketing-related submissions will not be shown.