OAS Wednesday - Semen Boosts the Immune System! Helps HIV

One of the things that bugs me most in pop-sci and woo-woo science is the obsession with "boosting" the immune system. The immune system is in a constant balancing act - tip it too far one way and even normally harmless bacteria become life-threatening. But tilt it too far in the other direction, and you can end up with things like allergies and autoimmunity. And some pathogens have learned to take advantage of your normal immune responses, meaning that "boosting" your immune system can sometimes do more harm than good. A good case-in-point is a recent paper published in PLoS Pathogens:

IL-7 enhances replication of laboratory strains and primary isolates of HIV-1 in human lymphoid and cervico-vaginal tissues ex vivo


Interleukin 7 (IL-7) is a type molecule called a "cytokine." Cytokines are chemical signals that cells of the immune system uses to communicate with each other. Some cytokines activate the immune system, some block the immune system, some act on particular cells to stimulate growth or activate one cell type while inhibiting others.

IL-7 is commonly thought of as a growth factor, specifically for T-cells, one of the major parts of the adaptive immune system. During HIV infection, one of the major problems leading to the disease AIDS is a dramatic decrease in T-cells that leaves people vulnerable to other infections. Because IL-7 increases the development of T-cells, several groups have been interested in using IL-7 as a therapy for HIV infected individuals to bring up their T-cell count. This idea seems rational enough: the problem with AIDS patients is a lack of T-cells. IL-7 increases T-cell counts. Therefore, we should use IL-7 as a therapy for AIDS patients. In fact, the company Cytheris is in a stage 2 clinical trial to do just that.

But if you know anything about HIV, you might have noticed a problem. HIV loooooves to infect T-cells. In fact, that's why AIDS patients have low T-cell counts: HIV has killed them all. This means that increasing T-cell counts may have the perverse effect of giving HIV more targets to infect and more factories in which to replicate.

As far as therapy goes, this probably isn't a concern. If a patient has so few T-cells that they need this therapy, it's clear that HIV is winning. In that case, reviving the patient's immune system is the most pressing concern, and the long term consequences of helping HIV are moot. If someone is bleeding out, you need to stop the bleeding, even if it means using your dirty hands that might give them an infection.

But the propensity of IL-7 to help out HIV is a concern in another realm - transmission. In someone that isn't already infected, having more T-cells around could make the difference between an infectious encounter and a non-infectious encounter. That's where this new research comes in. It seems that HIV positive men may actually prime their partner's immune system to make them more susceptible to infection with HIV.

See, semen is not just sperm in an innocuous fluid. It contains all sorts of other factors, including hormones that may improve mood*, stuff from food that the man has eaten (including allergens), and cytokines that modulate the partner's immune system. Scientists have long known that semen contains cytokines that dampen the immune response, a feature that makes sense from an evolutionary perspective (you don't want your partner's immune system attacking your sperm). But scientists have also noticed that semen contains IL-7, and that levels of IL-7 are higher in men with HIV.

So Andrea Introini et al decided to see what effect if any this might have on HIV transmission rates.

The concentration of interleukin (IL)-7, one of the most prominent cytokines in semen of healthy individuals, is further increased in semen of HIV-1-infected men. Here, we investigated the potential role of IL-7 in HIV-1 vaginal transmission in an ex vivo system of human cervico-vaginal tissue.


Since it would be tough to do experiments on couples in the wild, they instead turned to an experimental model, using explants of vaginal tissue** grown in a petri dish. Basically, they would take donations of tissue from patients that were undergoing hysterectomies, and keep that tissue alive in the lab. In this way, they could expose the tissue to HIV in the presence of experimentally controlled levels of IL-7. There are certainly problems with drawing conclusions from this system, but it's probably as close as you can get to doing controlled experiments in a natural system without actually infecting people. You'll lose some of the interactions with the rest of the immune system, but at least you've got an entire tissue instead of isolated cells.

Using this system, they then measured a number of different outputs, from HIV replication (literally counting the number of viruses produced), to the number of T-cells that are infected with HIV. Towards the end of the paper, they tried to explain a mechanism by which IL-7 could cause the results they see. I'll get to how that was done in the results section.


The most notable result is probably evident from the title: IL-7 increases HIV infection and replication.

From http://goo.gl/f0alA Figure 2a and 2b show increased HIV replication in vaginal tissue that has been exposed to IL-7

In the top graph, they are showing you the concentration of HIV virus particles found in the medium the tissues are bathed in. The dotted line is with tissue grown w/o IL-7, the dashed line is tissue grown in a low concentration of IL-7, and the solid line is tissue grown with a high amount of IL-7. The lower graph is basically the same data, but looking at the final concentration of HIV in each of the IL-7 conditions compared to the non-IL-7 condition. In other words, tissues in the presence of low dose IL-7 tended to have about twice as much HIV replication as tissue grown without IL-7, and the high dose tended to increase replication by 5-6 fold.

But why have both graphs if they're showing essentially the same thing?

When working with human tissue, you can often get wildly different results in terms of actual numbers, but scientists don't trust data that comes from a single experiment - you need to replicate it several times.  But if the authors had included multiple trials in that top graph, they could have had enormous error bars due to variation in the tissue that was used, so the top graph is showing a "representative" experiment. In the second graph, they normalize each experiment to the non-IL-7 control within that experiment, so they can graph them all together. Had they been using genetically identical mice, this might have raised some eyebrows, but when doing experiments with human tissue, it's totally acceptable.

In order to look at why IL-7 might be doing this, they decided to look at the effects of IL-7 on the T-cells present in the tissues that they were treating. T-cells that are infected with virus are often profoundly unhappy, and unhappy cells often commit cell-suicide. Cells committing suicide is often a good thing, because it prevents problems (like viral infections) from spreading, but IL-7 seems to prevent these infected T-cells killing themselves off, giving HIV more time to replicate and more opportunities to infect other T-cells.

This later result isn't terribly surprising, as IL-7 is known to be a growth factor , and several studies have shown that it can inhibit cell suicide, but it was important for these authors to demonstrate that it is still having this effect in the presence of HIV infection

Caveats / Discussion

I think this paper is an important advance, if for no other reason than highlighting a potential problem with using IL-7 as a treatment for HIV. It's important to understand factors that contribute to increasing HIV transmissiveness, and I was pleased that the authors did not suggest that this might lead to potential treatments or preventions, since blocking IL-7 during sex would be pretty impractical. Still, it's important to qualify what this paper does and doesn't show.

The paper does not show the effect of IL-7 in semen on the cells of an actual vagina. These are chunks of tissue floating in a bath of media, with a fairly high concentration of a cytokine present for long periods of time. The authors cite papers showing that semen from healthy men can contain 1-2 nanograms (1/1,000,000,000 of a gram) per milliliter, and that concentrations can be up to 100 times higher in HIV infected men. The lowest does the authors use in this study is 5 ng/mL, which seems reasonable until you realize that a typical ejaculate only contains a couple of milliliters of fluid, and that this would likely be significantly diluted out in the tissue. Further, I'm not sure anyone knows how long these cytokines are likely to persist before being used up or destroyed in the woman's genital track.

The authors did try to address that in Figure 3, where they showed that a brief treatment of IL-7 was enough to enhance replication of HIV. But it's still a fairly long pre-treatment (~16 hours), and it's hard to know if the concentrations are anywhere near what would be experienced in real life.

Shorter pre-treatment with IL-7 is sufficient to enhance HIV replication (Figure 3) Shorter pre-treatment with IL-7 is sufficient to enhance HIV replication (Figure 3)

Another problem is that this study mostly evaluates the cytokine IL-7 in isolation. As I mentioned above, semen contains all kinds of stuff, including cytokines that dampen the immune system. Unfortunately, the authors note that seminal plasma from donors is toxic to cells in culture, so the best experiment (exposing tissue to seminal plasma from healthy vs HIV infected donors) is not really possible. The authors did experiments in lymphoid tissue (which contains A LOT of immune cells) with diluted seminal plasma, but I'm not convinced that these experiments add much.

I should emphasize that none of these caveats are deal breakers. Experimenting with HIV is hard, and every model we have has significant limitations. This is an interesting paper, filled with good experiments. There's definitely more to do, and more to know, but then again, that's always the case in science. It's also worth noting that the authors of this paper did not stretch their results to make grandiose claims based on their data. I don't know if this is because of the authors' restraint or good peer review, but it's always nice to see in papers with a high potential for exaggeration.

[OAS (open access science) wednesday is an attempt to highlight research published in Open Access Journals like PLoS, eLife and PeerJ. Here are some reasons why I think OAS is important. Also check out the Open Science Federation]


*See this post for why the claims that hormones in semen probably don't affect mood.

**As an aside, the authors actually didn't restrict the study to vaginal tissue. They also looked at tonsillar tissue (representing the throat), and actually led with a figure on lymphoid tissue (part of the immune system). I'm guessing they did the initial experiments in lymphoid tissue because it's filled with immune cells and probably easier to get enough cells to measure. The results were essentially the same in all of the tissues that they looked at though.

More like this

New, weird, out-of-left-field, ideas-- we need them to stop HIV. This one makes sense, in retrospect, but I wouldnt have thought to try what these folks did: Targeting α4β7 integrin reduces mucosal transmission of simian immunodeficiency virus and protects gut-associated lymphoid tissue from…
For quite a while, now, there has been a connection between Endogenous Retroviruses and HIV. For some unknown reason, some of the young ERVs in humans, the ones that can still code for a protein here and there, are reactivated in HIV+ patients. Scientists have found ERV RNA in HIV+ patient plasma,…
A helpful reader (hat tip to easy hiker) sent along a story from New Scientist concerning a new report in The New England Journal of Medicine. The NEJM paper is a case series of six subjects who almost died as a result of a clinical trial of a monoclonal antibody being tested as a drug for…
You all might have heard about 'delta32' or 'delta-CCR5' people in association with HIV infection. People who naturally, by chance, have deletions and mutations in the CCR5 gene of their DNA dont make functional CCR5 proteins. It doesnt appear to be 'a big deal', and people who have this particular…