Feedback loops, sure, but since when has gamma been anti-inflammatory?

July 10, 2008 1:04 AM

I've been out of grad school for a few years and I haven't really been keeping up with my immunology and I was handed a paper today that calls IFNg an anti-inflammatory cytokine. What gives?

I understand the concept of feedback loops (see IL-2, IL-27) such that cytokines end up negatively regulating the very effects that they mediate. This paper is flat-out calling gamma an anti-inflammatory, proposing that elevated levels of serum gamma are a potential inhibitor of precursor lesions in a model of rat prostate cancer. What gives? Is it just that I spent my time in infection and immunity and cancer immunology puts gamma in the anti-sometimes-pro column?

posted by oreonax to science & nature (10 comments total)

I don't think it's just a cancer thing. This paper from 2001, places IFNg firmly in the anti-inflammatory camp for asthma and allergy:

"There is reduced production of IFN-g by T-cells of asthmatic patients and this correlates with disease severity. No polymorphism of the IFN-g gene has been associated with asthma. Administration of exogenous IFN-g prevents the airway eosinophilia and
hyper-responsiveness following allergen exposure in mice."
posted by roofus at 2:37 AM on July 10, 2008

Well, there're two lines of argument, I guess. The main one is that as a Type 1 cytokine, IFNg suppresses the production of Type 2 cytokines (IL-4, IL-5 etc.). You'll probably remember the Thype-1/Type-2 paradigm (pdf). Thus, in the typical Type-2 inflammatory situations (i.e. allergy), Type-1 cytokines such as IFNg might play an anti-inflammatory role. Indeed, a quick look at the literature indicates that administration of this cytokine for various allergic disorders has been tried with some success (reference).

The other is that, in various mouse (and rat) strains, and various disease models, IFNg appears to mediate different effects, which can't be explained in terms of a Type-1/Type-2 balance alone. One of the classic studies in the Multiple Sclerosis field was the finding that intraventricular administration of IFNg ameliorated EAE (a (bad) model of MS). This was a result that most certainly did not hold true for humans.
posted by kisch mokusch at 6:21 AM on July 10, 2008

Right, I remember the Type 1/Type 2 inter-regulation but I've never put prostate cancer into the Type 2 category. When lesions are introduced to the equation, I've always associated it with cytolytic activity (they show an increase in lytic units of NK cells) and inflammation (they show decreased gamma, and increased TNFa).

Thanks for the help guys - I think I'm just a bit out of practice. Shows you what being a slacker will do for you I guess.
posted by oreonax at 7:37 AM on July 10, 2008

What's the paper?
posted by kisch mokusch at 2:19 PM on July 10, 2008

Oh! Right, I meant to add that in, but I managed to forget in the flurry of making sure I had bagels to bring to lab meeting/journal club.

It is:
Cyclic adenosine monophosphate differentiated beta endorphin neurons promote immune function and prevent prostate cancer growth. DK Sarkar, et al. PNAS V105, No. 26, pp. 9105-10.
[Link] to full text at PNAS.

Ultimately, there's not much meat on the immunology side of things - from what I understand, DK Sarkar primarily studies beta-endorphins and not so much cancer immunology.
posted by oreonax at 4:46 PM on July 10, 2008

Heh. The data that I didn't link to (because it's unpublished*) is reports that IFNg KO mice on the Balb/c background appear to be protected in particular neuroinflammatory conditions (as opposed to C57BL/6 mice which displayed an increased susceptibility). It's not the first time I've seen counterintuitive results regarding cytokines in the CNS (hence the link in my previous comment). But in the article you link, the line "antiinflammatory cytokine IFNg" is not something that should be written with such carelessness, especially since it refers to "general immunology" (NK cells in the spleen) rather than in the CNS in particular. One can't write such a thing without qualification, so I certainly don't agree with you when you say that you've been slack with this regard. Frankly, I'm surprised the reviewers didn't question them about it.

I can't see the paper here from home (only the abstract), so I'll have to see where they're getting their ideology from when I'm back at Uni (on Monday). But since the authors aren't immunologists (and I'm speculating here) perhaps, since the authors are neuroscientists, and given the weird data regarding IFNg in the brain, they might be forgiven for thinking of IFNg as an anti-inflammatory cytokine.

A comment with regards to what "type" of response an anti-cancer response is: You are correct in thinking that cytotoxic (CD8) T cell activity places the response in the Type-1 category (under the current interpretation of the paradigm). But ever since I saw Chris Parish successfully kill cancer with eosinophols (link) - which are generally pretty firmly in the Type-2 category - I refuse to try to categorise anti-cancer immune responses in terms of Th1 vs. Th2. The immune system is too sophisticated to generalise it in such a way.

*not my data, a colleague's, so I won't delve into the particulars.
posted by kisch mokusch at 5:22 AM on July 11, 2008

Fair enough re: cancer and Th1 or 2 categorization.

And I now that I'm taking the time to properly respond, I seem to remember that MS is treated with IFN injections, but I'm pretty sure I read that in a Faye Kellerman novel, so it's not exactly proper documentation.

You've intrigued me with your colleague's unpublished data - if you don't mind me making sure I understand: Balb/c IFNg knockouts have less neuroinflammation than wildtype C57Bl/6 mice under the same conditions?

From what I typed above, and since Balb/c mice skew towards Th2 responses, it sounds like you have a Th1 mediated inflammation model. Is that unexpected?

I hope you don't take this as prying into your colleague's data, I just rarely get to talk shop anymore that I'm curious how much I actually remember from school. I'd be happy to drop this like a hot rock if it's too many questions.

As for the paper I was reviewing for journal club, their beta-endorphins were undetectable in the periphery, and their cytokine data were all ELISAs from serum. There were a few things that made me wonder about the review process, but I'm not a neurobiologist so it might just be me.

A lot of questions came up that had no answer on the face of the paper. On the one hand, the authors didn't make any claims that weren't solidly backed up by their data. However, the rationale behind gathering the data was not entirely elucidated and remains a mystery to all of us as do a few figures.

The bagels were tasty though.
posted by oreonax at 4:18 PM on July 11, 2008

Aargh. I miss-wrote in my answers. Twice!

Firstly, the reference I gave was wrong. That referred to the Type-I INFs (IFNa and IFNb), not the Type II (but still Th1) cytokine IFNg.

A further look at the literature indicates that current treatments with IFNg put the cytokine pretty firmly in the proinflammatory box, especially in cancer (reference 1, reference 2). There are also a few skin disorders for which the treatment of IFNg might be useful (reference), similarly in a pro-inflammatory capacity.

That's not to say that IFNg can't be anti-inflammatory (there was this study in asthma), but the common/default thinking would still be pro-inflammation (especially in humans).

The comments regarding conflicting data in rodents still stands (esp. with regards to the CNS), but I mixed the results up regarding my colleague's data (clearly yesterday was not a good day for me!). The model is one of virally-induced immunopathology of the brain. One of the findings was a decreased pathology observed in IFNg KO mice on the C57BL/6 background compared to WT (as one would expect, since it's the excessive immune response that's causing the pathology). But the opposite was observed in IFNg KO mice on the Balb/c background (i.e the pathology worsened). So, in the Balb/c mice, the lack of IFNgamma was in some way leading to a bigger immune response. However, the mechanisms haven't been dissected out, so I can't really tell you much more. There are a lot of differences between the two strains, so who knows.

In general immunology, the whole Th1-Th2 paradigm gets used more as a mnemonic, I think, rather than a hard and fast rule. We clearly make both Th1 and Th2 responses simultaneously to most immune challenges, and the degree of skewing is influenced by numerous factors, not simply the nature of the pathogen.

One of the treatments for MS is beta-IFN (i.e. Type I IFNb), which really is an anti-inflammatory cytokine.
posted by kisch mokusch at 10:18 PM on July 11, 2008

Very cool - thanks for the details on the gamma knockouts. If it ever occurs to you, I'd love to know when it gets published, it sounds fascinating. You've got me wondering all kinds of furiously about cytokine profiles and the histology of the brains of these guys.

Whee!

Thanks for the chance to get my toes wet again - it's been truly fun.
posted by oreonax at 1:50 PM on July 12, 2008

No problem. Like talking shop myself :-) Will ask about the gamma KO theories next time I see my colleague, and mail you his thoughts. However, I fear that it will be a long time before the Balb/c story is published (I don't think there's anybody working on it right now). Part of the C57BL/6 story, OTOH, is already out.
posted by kisch mokusch at 1:49 AM on July 13, 2008