Tuesday, July 31, 2012

What is BCL-2 and why should we inhibit it?

Another class of investigational treatments working their way through the system are the BCL-2 inhibitors (named as the "B-Cell Lymphoma gene number #2).  These investigational therapies will likely be relevant to patients with CLL or NHL.   They may actually be relevant to a number of cancers but CLL/NHL will be be where they get evaluated first. 

BCL-2 inhibitors have been a long sought goal for the pharmaceutical industry but they have had to try targeting the protein a number of ways before they found what looks to be the promising set of compounds that are moving forward now.

I thought a little biology might be fun - so let's talk about the mechanism by which these work.  Within a cell, there are a variety of "organelles."  In the same way you have a liver and kidneys, cells have their own organs - known as organelles.  These have names such as the golgi apparatus, endoplasmic reticulum, mitochondria, lysozymes, etc.  The mitochondria is the equivalent of the "power plant" for the cell.  It is where the vast bulk of energy is synthesized for the normal function of the cell.

In normal physiology, it is important to have a way to get rid of cells that you don't need anymore.  The main mechanism by which this "programmed cell death" occurs is called "apoptosis."  It turns out that the mitochondria plays an important role in apoptosis.  It contains a variety of enzymes known as caspases tucked away between two membranes.  When the caspases are released from the mitochondria into the cell, they function as the "cellular executioners" to bring about an orderly demise of the cell.

At the outside surface of the mitochondria, there is a constant battle between two sets of proteins.  These conflicting proteins create a balance between the pro-survival and pro-death signals.  Depending on environmental cues or external signals, the cells fate is played out.  The main "pro-survival" protein is known as BCL-2.

You might ask why your body doesn't just turn to mush when you take one of these drugs as it should release the death signal throughout mitochondria everywhere.  It turns out though that cancers are "addicted" to BCL-2 in ways that most other cells are not.  The inside of a cancer cell is not as friendly as other cells and they need some mechanism to protect themselves from the death signals they constantly receive.  In follicular lymphoma, pieces from two chromosomes (14:18) break apart to ensure that too much BCL-2 is produced.  In CLL the mechanism is less clear, but may involve genomic alterations in the 13q14 region (see my prior post on micro-RNA's).  In either case, there is too much BCL-2 protecting the mitochondria, so no matter what "death signal" (i.e.. chemotherapy, immunotherapy, etc) you send to the cell, it survives.

For this reason, pharma has long wished to create a good BCL-2 inhibitor.  The first attempt was a molecule known as Genasense.  This used a technology known as "anti-sense technology" which has not proven itself to be a good way of making a drug.  The next significant foray into BCL-2 inhibitors was a series of compounds generated by the folks at Abbott Labs.  Abt-737 was an IV formulation of what became the oral Abt-263.

Abt-263 has reported activity in CLL - but as is often the case with drugs, we discover things along the way we didn't expect.  In this case, there is a protein very similar to BCL-2 known as BCL-xL.  It is the main protein keeping your platelets going and when you inhibit BCL-xL platelets drop really fast.  This makes it unappealing to think about combining it with other types of CLL/NHL treatment and may prevent the drug from finding a home in our treatment choices.

The next research treatment coming down the pipeline is Abt-199 which also has the designation of GDC-0199.  This drug was designed by the same folks at Abbott that made 737/263.   Abstract 546 at the 2012 European Hematology Association details the clinical experience to date.  This link downloads a long PDF.  Best to use the search function after getting it open.

Sunday, July 29, 2012

How I approach follicular lymphoma (part 1) - patients with low risk disease

A lot has changed in follicular lymphoma in the past few years. Quite a few items considered standard of care just a few years ago have been replaced by new standards. I thought it might be worth while putting forward my take on all of it. For the purposes of this post, I want to talk about patients with follicular lymphoma and "low risk" disease.

In order for us all to be on the same page about what constitutes "low risk" disease, I need to immediately introduce a scoring system called the FLIPI scale. It stands for "follicular lymphoma international prognostic index"

Follicular Lymphoma International Prognostic Index

This scale uses age > 60, stage III-IV, hemoglobin < 12, number of nodal areas > 4, elevated LDH. Patients with 0-1, 2, 3, 4-5 are grouped together in risk strata. Lower scores are better and higher scores are worse.

A good starting place for patterns of care in follicular lymphoma is the "lymphocare" study published in 2009. What is startling is how much has changed in the last three years.

Follicular Lymphoma in the United States: First Report of the National LymphoCare Study

The punch line from the paper is that in a mix of community practice sites (80%) and academic sites (20%), about 2800 patients were enrolled at diagnosis and followed in order to characterize treatment choices and outcomes. 18% were placed on "watch and wait," 14% were given rituximab alone, 52% were given rituxan with chemotherapy. When R-Chemo was given R-CHOP was given to about half, R-CVP to about a quarter, and R-Fludarabine to about 15%.

I find it unusual that so much has changed so quickly. For a disease like follicular lymphoma, this treatment mix has been turned on it's head by several high profile publications.

Let's start with "Watch and wait." For many patients, the idea of "watchful waiting" is a tough recommendation. I completely understand the sentiment that if you have cancer you ought to do something about it. The problem from a historical perspective is that we didn't necessarily have anything that really changed the long term prognosis. That historical precident was established for quite a few years by the BNLI study which started accruing patients a LONG TIME ago. In this study patients with advanced stage FL were randomized to watch and wait vs immediate chlorambucil.

Long-term effect of a watch and wait policy versus immediate systemic treatment for asymptomatic advanced-stage non-Hodgkin lymphoma: a randomised controlled trial.

This study found that with long term follow up, the folks who went on "watch and wait" actually did a little better than patients who got up front chlorambucil (keep in mind that chlorambucil is an "alkalating" agent which causes DNA damage). A generation of oncologists followed this paradigm for nearly 20 years. Unfortunately, chlorambucil is such a lousy drug (I only use it in the "very unfit patient population"). One point to highlight for later is that the average survival of either arm on this study was only between 5-6 years. Furthermore, this wasn't terribly different than the overall natural history of the disease.

The natural history of initially untreated low-grade non-Hodgkin's lymphomas.

Over time however, more effective regimens such as CVP, or CHOP came along. With more effective regimens, watch and wait remained a good option, but the pendulum moved back toward combination therapy. By the time of the lymphocare study in the mid 2000's watch and wait was a distinctly small minority. Having done my fellowship at Stanford with Ron Levy and even Saul Rosenburg, I have certainly watched and waited on quite a few patients.

The biggest change to this equilibrium has been the introduction of rituximab. Rituximab is not chemotherapy - it is "immunotherapy." It is an "engineered antibody." You make antibodies to flu, e. coli, etc. This is just an antibody against the lymphoma. I tell patients it coats the outside of the cancer cell and "focuses" the immune system. There may even be some degree to which it helps "train" the immune system to fight the lymphoma.

There are several important studies that have made this option more welcome in the minds of physicians. The first was a single arm phase 2 study of four doses of rituxan which showed an overall response rate of 70% with complete response of 30%.

Rituximab therapy for patients with newly diagnosed, advanced-stage, follicular grade I non-Hodgkin's lymphoma: a phase II trial in the North Central Cancer Treatment Group.

The second paper was a plenary session at ASH in 2010 which was a redo of the original chlorambucil study but was "watch and wait" versus rituxan. This study remains quite "young" so we do not have good long term data. We cannot yet say whether patients live longer etc. What we can conclude however is that giving rituximab certainly delays the need for subsequent chemotherapy.

An Intergroup Randomised Trial of Rituximab Versus a Watch and Wait Strategy In Patients with Stage II, III, IV, Asymptomatic, Non-Bulky Follicular Lymphoma (Grades 1, 2 and 3a). A Preliminary Analysis.

Finally, the Resort study presented at ASH 2011 was significant. This study took "low risk" patients and randomized to four doses of Rituximab versus four doses followed by one dose every three months forever. In the patients who only got four doses, they could get four more doses if their disease came back. Only those patients who had an initial response were followed long term (70%). So the study design is more about the best way to maintain long term benefit from rituximab with either "re-treatment" or "maintenance" in rituximab sensitive patients.

Results of Eastern Cooperative Oncology Group Protocol E4402 (RESORT): A Randomized Phase III Study Comparing Two Different Rituximab Dosing Strategies for Low Tumor Burden Follicular Lymphoma

The key findings was there was very little difference between the two arms in terms of how long a patient was sensitive to rituximab. There was a slight increase in patients needing chemotherapy in the re-treatment arm but not enough to really make thought leaders think that maintenance was the winner. Perhaps the most significant secondary finding though was that there were more patients doing well in either arm of this study than compared to historical controls by a long shot. There is danger in this sort of comparison because that is not what the study was designed to answer.

With these three studies however, you see more and more editorials asking if "watch and wait" is an outdated strategy. I don't think it is just yet, but I admit, my threshold for giving single agent rituximab is a lot lower than it was before and I think that is reflected in national practice patterns as well. If we swing back to the lymphocare study, I think the impact here has been that single agent rituximab use has gone up quite a bit. Instead of the 14% reported previously, I would wager it is somewhere between 25-35% of patients. This strategy has probably stripped out more of the "watch and wait" group as well as some of the other folks who would have previously gotten R-Chemo.

I think that "watch and wait" was probably underutilized in the lymphocare observational study and probably more so now. A lot of patients are less comfortable with extra CT scans due to the radiation concerns.

I think that rituxan monotherapy opens up a lot of interesting research options. It may prove to be a useful platform for new drug combinations in research. I am particularly excited by a study (which sadly I am not a part of) where it is revlimid-rituxan versus investigator choice of R-chemo regimen. I hope to see studies soon in which the design is rituxan versus rituxan / new agent.

In our next follicular lymphoma post, I will talk about what sorts or R-Chemo regimens I give when I need to reach for chemotherapy regimens.

Wednesday, July 25, 2012

13q Part Three - interpreting test results and understanding limitations

13q test results:

When FISH is ordered and a 13q abnormality is observed, there are a variety of ways it can be reported.  It is possible to have a deletion on one chromosome in many of the cells, a deletion on both chromosomes in some of the cells, or any combination of the two.  Typically the report will indicate the number of cells lacking one or both copies of 13q.

Up until recently, we were not totally clear how these differences influenced clinical outcome.  Several earlier papers seemed to indicate that having both copies of 13q missing was worse than having only one copy missing.  Unfortunately, some of these studies were limited by small sample sizes and may not have statistically accurately reflected what these results mean.

Genome-wide analysis of DNA copy number changes and LOH in CLL using high-density SNP arrays.

Biallelic deletion 13q14.3 in patients with chronic lymphocytic leukemia: cytogenetic, FISH and clinical studies.

More recently, several groups seem to have arrived at consensus that the actual percentage of cells lacking the 13q matters more than whether there is only one copy or two.  Different groups come up with different number thresholds, but that is because they need to make a categorical separation of data that is by nature a continuous variable.  It is probably safe to say that the more cells with 13q deletions the worse it is.

There are a number of things however that FISH simply doesn't tells us.  In some cases we know whether the missing data is important, in some cases, we have yet to figure it out.  FISH does not distinguish between the minimal deleted region (MDR), commonly deleted region (CDR) which contain Dleu7, or the less frequent larger 13q deletions that contain the RB protein (type II deletion).  Unfortunately FISH only tells us if there is a deletion or not - it doesn't tell us the size or what genes are included.

Mutations have been found in the Rb protein in patients with CLL, but those are not detectable by FISH.  Other conditions where the cell tries to replace the missing material by making an extra copy of the remaining chromosome (aka uniparental disomy) are missed entirely.  Furthermore, the cancer cell can turn off the mRNA's through a process called methylation and even if you could figure all this out, it is not clear how these findings affect the biology.

For now we know that if you have a 13q deletion and NO OTHER FISH abnormalities, you are probably in good shape.  If you also have a 17p or 11q, those trump the 13q.  Hopefully we will learn how to handle the other information we know is there and create tests that let us sort it all out.

Tuesday, July 17, 2012

13q - part two

So now for more on 13q:

In cases of CLL where the 13q region is missing, there is variability with regards to just how much is actually gone.  You can have the MDR (minimally deleted region), CDR (commonly deleted region) which can occur in the type I deletions, or you can have a huge chunk of DNA gone that eliminates a specific gene called Rb (important tumor suppressor) in which case you have a type II deletion.

If we look at the MDR, we already talked about the miRNA's in the prior post.  In humans, the miRNA's are completely embedded in another RNA transcript known as a "sterile transcript."  I would love to know what this "sterile transcript" actually does - but so far, nobody has figured it out.  It is known as DLEU2 (deleted in leukemia 2).  We know from other literature that these "sterile transcripts" do have a regulatory role on certain proteins etc. we just don't know much about this one.

A savvy group of researchers (article linked in prior post) was able to make two different strains of mice.  One which just lacks the miRNA's and another which lacks both the miRNA's as well as DLEU2.  If you do nothing to the mice except watch them over time, the group lacking the miRNA's gets monoclonal B cell lymphocytosis, some CLL, and some NHL.  Interestingly the group that lacks DLEU2 gets more CLL, NHL, and even some diffuse large B cell lymphoma.  In summary, the double deletion (miRNA and DLEU2) is more aggressive than just the miRNA.

They did some further experiments and showed that if you take a mouses B cells with these abnormalities and give them a growth stimulus, they start growing faster, and grow longer than normal B cells.  Some of these "cell cycle" genes that are affected are classically known to be cancer related - including one of my favorite proteins  CHK1  (more to follow when we talk about 11q deletions).

Just a little farther ways out the 13q chromosome lies the DLEU7 gene (different than DLEU2).  Instead of being a "sterile transcript" DLEU7 is responsible for actually for the synthesis of a specific protein.  This protein has a regulatory role in signaling through a surface molecule known as the BAFF (B cell activating factor) system.

BAFF is interesting to me because it is clearly an important cytokine (aka: micro hormone) in CLL.  In fact there are several BAFF inhibitor drugs out there, even one that has been approved in lupus.  That drug is locked up in a "custody battle" between two big pharma companies.  I've tried to get it for research studies but I don't see that happening soon.  There are other anti BAFF drugs out there and I think we will get a readout as to their efficacy in CLL at some point.

One of my research buddies, Jennifer Brown at MGH in Boston has been studying familial CLL and she has identified a family where all the affected family members have a lost both copies of their DLEU7 gene - very interesting.

If you have a really big chunk of DNA missing from chromosome 13q14 (ie. type II deletion), you knock out a copy of the RB gene.  RB (aka retinoblastoma - a gene identified as the key driver of an uncommon eye tumor in kids) is a gene that plays a role analogous to p53 in some ways.  It is a critical determinant as to whether a cell proceeds through cell cycle or gets halted and may even directly interact with p53 (the key molecule altered in 17p deletions).  It appears that having RB deleted adversly affects prognosis.

So in this small area, you have several different molecules that all cooperate to form CLL.  This is a pretty power packed part of the genome as far as B cell biology is concerned.  Furthermore, it helps us understand ways to go after the fundamental biology of CLL instead of just throwing chemo at the problem.

Next, we will talk about how we test for it currently and what the limitations are of our current testing

Sunday, July 15, 2012

13q - what is it and why does it matter? (part 1 of several)

July 15

I am really fascinated by the genomic alterations that have been observed in CLL.  One lesson we have learned over the years is that if you have a pattern - there is normally some key biology to be learned.  Once you learn the biology - then you can begin to make more informed ideas about how to treat such a disease.

13q14 is important for anybody who thinks about blood cancer.  It is the most common genomic alteration in CLL (about 50% of patients at diagnosis) and it is also quite common in NHL, myeloma, and even solid tumors.  In CLL - it is considered a "favorable" alteration.  Of all the FISH abnormalities you can have, if that is the only one, you are likely going to live longer than other patients with CLL.  In fact, if you have an mutated BCR, isolated 13q deletion, and early stage disease - chances are you may never actually need treatment.

So what have we learned about 13q14?

The nomenclature here is important.  You have MDR (minimal deleted region), CDR (commonly deleted region), type I (short, but typically inclusive of MDR/CDR), type II (big, inclusive of RB gene).  These can be confusing at first but should make more sense as we go through it.

FISH (fluorescent in situ hybridization) is a test where you use a probe that has a fluorescent tag.  The probe has a certain DNA segment attached that is very specific to a region of the genome.  You then "melt" the DNA in the specimen (blood lymphocytes, lymph node tissue, marrow), drop the probe in - let them bind, then you count the signals.  Since we have two copies of each chromosome, you should have two signals for each probe.  If you have a region of chromosome 13q deleted, you will only see one signal.  Occasionally, a patient will have both copies of 13q deleted (about 15% of time) and you will get no signals.  You use a control probe to make sure the test actually worked.  You can do the same for chromosome 12, 17p, 11q, or any other probe you want.

As you may know, we have 23 pairs of chromosomes - 46 total (although sometimes men lose their y chromosome without any consequence which is an oddity that always makes me laugh).  Chromosomes are rarely symmetrical, and the long arm is referred to as the "q" arm and the short is the "p" (think petit).  You can then use specific staining which creates bands along the chromosome visible to a microscope.  If you go out (starting from the central part known as the centromere) the long arm of chromosome 13 a total of 14 bands you get to a region known as 13q14 - and here is where there is a lot of action in CLL.

 In the case of chromosome 13q14 deletions, the standard probe binds to a region which is deleted in just about every case of 13q14 abnormalities.  It is the smallest known segment to be absent, so it is called the "minimal deleted region" or MDR.  Right at the "landing spot" of the probe you are in the middle of a segment that encodes an RNA molecule known as DLEU2 (deleted in leukemia #2).  We don't actually know what DLEU2 is supposed to do, but it gets faithfully copied from DNA to RNA.  In the process two smaller segments of RNA get cut up known as miRNA (for micro-RNA).  In fact, miRNA's were discovered in the first place when folks were studying this small stretch of genomic alteration in CLL patients.

These two miRNA's are very important.  They are miRNA 15 and 16.  When the cell makes these miRNA's they go out to other parts of the cell and find mRNA with similar sequence and they bind to each other.  Brief detour - DNA is the master template, and it gets copied to mRNA.  mRNA then leaves the nucleus, goes to the cytoplasm of the cell and gets "transcribed" into a protein - these are the engines that make a cell do whatever it does.

miRNAs 15 and 16 are important because they serve as negative regulators of BCL-2 mRNA and a host of other genes (please allow the simplification for now).  BCL-2 will be the subject of another post, but for now you should know that BCL-2 is a very important protein that keeps CLL cells alive a lot longer than they should.  There are also some very interesting drugs out there that target BCL-2 (ABT-199 AKA GDC-0199, navitoclax, others) Therefore, those patients that lack the negative regulator - get more of the BCL-2 protein than they should.

There are additional complexities about BCL-2 expression levels that an informed critic could say I am ignoring for now, but for the purposes of this post, I would like to leave it in the more simplified version.

Now for the interesting part.  If you take a mouse and get rid of that segment of DNA - they develop CLL (about 25% at 10 months - which in mouse terms - is not that aggressive).  Interestingly, they also develop lymphomas and even monoclonal B cell lymphocytosis.  They also have "stereotyped" B-Cell receptors which is another important feature of human CLL.

In future posts - I want to take a longer look at what else happens at 13q14 - but I had to do some background work here to explain how the test works etc.  Here is a link to one of the key papers 

Thanks for reading.

Saturday, July 14, 2012

CLL: Going nowhere fast or going fast to nowhere? How Growth Kinetics Apply to Clinical Picture

Understanding the disease

Traditional theory of CLL holds that it is a disease of long lived malignant cells that slowly proliferate.  Of course there can be significant variability from patient to patient.  The idea would suggest that typical 13q/mutated patient with a doubling time of 4 years has a very low proliferative rate.

One of my favorite CLL papers shatters this theory:

In vivo measurements document the dynamic cellular kinetics of chronic lymphocytic leukemia B cells

In this paper, researchers had patients drink a daily allocation of deuterium (aka heavy water).  They could then take the malignant cells and measure how much deuterium was incorporated into their DNA and from there calculate a "birth rate" and upon cessation of deuterium a "death rate" of the population of CLL cells.

In summary, they found that CLL is much more like a stationary treadmill than a slow walk.  In other words, there is a significant and balanced birth and death of CLL cells that can be as high as 3% of the total population of cells per day.  Progression occurs when there is an imbalance between a higher birth rate and a lower death rate.  Occasionally you will see patients WBC fall quite a bit.  This is an imbalance where death rate exceeds birth rate.  At 3%, you could envision virtually the entire population of cancer cells "turning over" in a period of several months.

This is important because we know that CLL is a disease in which the original clone begets several sub-clones.  If a sub-clone gains a growth or survival advantage, in this model, it is a lot easier for that clone to become dominant.  This is why it is important to recheck FISH when disease relapses after therapy - if you kill off the easy disease, you may be left with more resistant disease.  This is why you see a higher rate of 11q, 17p at relapse than original presentation.  This is also why a disease that remains "stable" for quite some time can suddenly become more "active."

In my own way of thinking, it is why I am trying to move the field away from traditional fludarabine / bendamustine based regimens in front line treatment.  I worry that we are selecting for resistance in the same way we remember being told to take all our antibiotics when we were kids with a sore throat because we didn't want the bacteria to become resistant.

Sadly, there is a tremendous under appreciation of how to use FISH out there with many docs.  It is not used as frequently as it should be.  It is not a static measurement for a patient - if it can change, and that change influences your therapy, you should be getting it rechecked at major clinical turning points.

Friday, July 13, 2012

Immunoconjugate Drugs

Drug Update

ASH 2012 CD22 ADC update 
ASH 2012 CD79 ADC update

One of the technologies I am most excited about right now are a family of drugs known as "antibody drug conjugates (ADCs)."  I think these compounds may begin to fulfill the dreams we had years ago of effective and non-toxic treatments.

The basic idea is the following.  Take a monoclonal antibody.  You make antibodies to fight off flu, cold viruses, bacteria, etc.  Instead of a naturally occuring antibody, we use technologies that make them target lymphoma or CLL cells.

That is the idea behind rituxan.  The antibody coats the outside of the cancer cell and tells the rest of the immune system to attack the cancer cell.

Depending on the target though, sometimes the antibody gets swallowed by the cancer cell instead of just coating the outside.  Once it gets swallowed, a special part of the cell called the lysozymes helps break the molecule into little pieces.

The true innovation of these drugs is the "linker" segment which allows you to attach super potent chemotherapy to the antibody so that the two are effectively "glued" together until they hit the lysozymes.  Once the ADC binds to the cell and gets swallowed, the super potent chemotherapy is released exclusively into the cancer cell instead of going throughout the entire body.

There have been a handful of these drugs out there for a while.  A well known success/failure is gemtuzumab (aka myelotarg).  This was initially approved by accelerated mechanisms for Acute Myeloid Leukemia, then revoked by the FDA when confirmatory studies failed to support the initial enthusiasm.  Frankly this drug will likely come back but only after we are more confident we know how to use it right.  In some regards, Ontak is sort of the same idea.  One can even put zevalin in this category but here you used radiation instead of chemotherapy and that made it logistically difficult for practices to administer what is a very effective treatment.

The problem with the early molecules was that the linker was not so good and free drug got loose in the body.  In the case of gemtuzumab that caused a lot of problems with the liver and low blood counts.

The folks from Seattle Genetics seem to have solved the linker issue.  Their first approved drug - Adceteris / brentuximab is truly revolutionary in Hodgkin's Disease.  Targeting CD30 (a marker for HD or Anaplastic Large Cell Lymphoma), this well tollerated single agent gets durable remissions in patients whose cancer has been refractory to just about everything else.

We have also found CD30 shows up on a number of non-hodgkin's lymphomas.  This was presented at ASCO this summer.

Brentuximab vedotin for relapsed or refractory non-Hodgkin lymphoma: Preliminary results from a phase II study.

They have licensed the technology to a number of other companies and now a host of drugs are working through the system.  An entire class of drugs is now working their way through the system that may prove to be effective in patients with CLL/NHL.  I wager that these drugs are going to be quite active and offer a new type of treatment for patients at all stages of disease. 

Genentech has a little video series for the interested reader - excuse the commercial reference.