top of page


Updated: Mar 17, 2022

Reminder of the NK cell activation threshold as a predictor of survival

This, we know.

· INKmune generates memory-like NK cells in vivo, referred to as tumor-induced memory-like NK cells (TIML), whereas cytokine-induced NK cells are those that are generated ex vivo by administration of cytokines to NK cells (CIML).

· NK cells carry out an essential function of immune surveillance in the body.

· A patient’s NK cells’ activation level – or ‘cytotoxicity’ – is a predictor of overall survival (2002 and 2019). In patients with NK cytotoxicity above a certain threshold, there is substantially higher (9 in 11) overall survival (OS) over the course of two years. In patients with NK cell activation below that level, there is a substantially higher (7 out of 8) death rate over the same period.

Thus, the higher the level of NK killing, the better the chances of overall and long-term survival of a patient.

The consequences of a memory-like phenotype of NK cells

· Conventional NK cells are thought to have a lifespan of 12-14 days (some literature mentions even shorter lifespan, for conventional NK cells).

· NK cells were thought not to be able to show cytolytic activity due to previous activation, contrary to T-cells. In recent years, this has been proven wrong by the discovery of memory-like NK cells.

· Memory-like NK cells have enhanced function, are capable of persistent survival, expansion, avoidance of anergy in vivo, have a lower threshold of activation in response to cytokine restimulation, enhanced cytotoxicity, and they undergo clonal expansion creating self-renewing immune memory. This subset of natural killer cells is basically a super-natural killer cell, and what you want as drug developer as this subset also makes the crossover to the adaptive immunity.

· An activated profile apparently expresses more CD69+ and CD25+. There are also subphenotypes within memory-like NK cells.

· The presence of memory-like NK cells also seems to trigger considerable expansion of total NK cell count.

Thus, the longer a memory-like NK cell persists and the better its cytotoxic function, the more one will have cancer-killing, and hopefully the longer a remission can be sustained.

Consequences for investors

In oncology, duration of remission is essential to judge efficacy of a drug candidate.

The presence and activation of memory-like NK cells could be seen as an indicator of the likelihood of success of any drug therapy in trials.

The percentage of memory-like NK cells and their avidity can therefore be seen as parameters to identify (i) the likelihood of success of a drug candidate vis-à-vis other drug candidates, and (ii) the probability of success in the immune system’s fight against cancer.

Efforts to create memory-like NK cells

Over the past years, some efforts have been made to create memory-like NK cells first ex vivo, then in vitro. So far, these memory-like NK cells have only been able to be created by the addition of cytokines, such as IL-12, IL-15 and IL-18. These efforts have led to so-called cytokine-induced memory-like NK cells (CIML).

One of the main reasons for the direction of the oncology field towards NK cell therapy was the relative safety of NK cell killing, due to the reduced risk of cytokine release syndrome.

However, adding cytokines to NK cells ex vivo is possible but is not sustainable in the same manner in vivo. Continued cytokine support in vivo may also raise toxicity risk and neurologic issues (see below – 4th tentative comparison), thereby diminishing the main of at least one of the initial advantages of NK cell-based immuno-oncology.

INKmune’s tumor-primed NK cells in the first patient have shown the following in the first patient:

Percentage of activated NK cells

Day 119 was the last day measured. That is at least 17 weeks of persistence, or 4 months, without cytokine support. Persistence probably continued after those 4 months.

Proof of proliferation of memory-like NK cells and killing capacity stems from the below.

In vitro, the INKmune™ activated NK cells were better at killing cancer cells than the patient’s own NK cells prior to treatment, with an 82% increase in lysis of K562 leukemia cells and a 47% increase in lysis of NK-resistant RAJI lymphoma cell tumor cells as early as day eight.

One could also add, at this point, the remarkable benefit INKmune has brought in the patients treated, including two terminally ill patients.

Tumor-primed NK cells (TpNK) also show better avidity than resting NK cells and NK cells with added IL-2 and IL-15 as shown here:

How do TIML compare to CIML?

TIML = tumor-induced memory-like NK cell (INKmune)

CIML = cytokine-induced memory-like NK cell

v First tentative comparison – CIML without additional therapy/cytokines

In 2009, cytokine-induced memory-like NK cells had been generated in vitro. A persistence of about three weeks had been reported, which was considered remarkable.

Three weeks as reported here is less than one fourth (<25%) of four months reported by INMB. Adding 300% duration - and possibly more - is a massive improvement.

v Second tentative comparison – CIML in AML

In 2018, it has been reported that CIML exhibit enhanced responses in AML myeloid leukemia. “After transfer, memory-like NK cells were detectable in the blood and BM of patients for weeks, proliferated extensively, and expanded in vivo.” No more detail was given as to the duration of response in the concerned article. In a related article on the same trial, it was reported that their persistence duration was 2-4 weeks, adding that this short persistence could have been due to HLA incompatibility, and arguing that transfer into an immune-compatible environment would prolong their survival.

This may mean that memory-like NK cell persistence may be problematic or uncertain in case of adoptive transfer of CIML cells which do not originate from the patient itself (autologous).

Some weeks as reported (2-4 weeks) is far from the four months (and possibly more) reported in the INMB trial.

INKmune should not have that issue of immuno-(in-)compatibility, as it does not inject NK cells, and is an allogeneic therapy.

v Third tentative comparison – tumor irradation + CIML

In mice having been irradiated - using high-energy radiation from x-rays, gamma rays, neutrons, protons, and other sources to kill cancer cells and shrink tumors - IL-12/15/18–preactivated NK cells were still detectable for at least 3 mo after adoptive transfer.

First off, one should compare what is comparable. These results cannot properly be compared to INKmune, as irradiation apparently massively influenced the results (up to six times, with regard to IFN-γ, and as test results were shown in mice. “Radiation therapy (RT) was essential for the antitumor activity of transferred NK cells. […] RT greatly increased numbers and function of transferred NK cells. […] Most importantly, in irradiated hosts, significantly higher levels of IFN-γ, granzyme B, and perforin were detected in transferred IL-12/15/18 NK cells compared with nonirradiated hosts (Fig. 7, b and c). In fact, six times more IFN-γ–expressing transferred IL-12/15/18–preactivated NK cells were observed in irradiated compared with nonirradiated hosts (Fig. 7 c).

Nonetheless, INKmune still outperformed the above reporting (3 months in irradiated mice vs. 4 months in non-irradiated humans). That speaks for INKmune, and also implies that INKmune may have additional effect if combined with irradiation.

v Fourth tentative comparison – CIML with added cytokine support, in vitro

In 2019, it had been reported that, after adoptive transfer of CIML, memory-like NK cells were sustained for more than 2 months, constituting 20-50% of total NK cell count at day 60. These results have been generated with added cytokine support and in vitro.

Again, these results cannot properly be compared. In vitro, one can add any cytokine such as IL-15 to keep NK cells alive/activated as long as one wants, but this does not work in vivo. After drug administration, adding cytokine support may be problematic as it may lead to toxicity and neurologic issues.

Nonetheless, INKmune still outperforms the reported result:

- (i) with regards to duration (4 months for INKmune vs. 2 months as reported in vitro with cytokine support);

- (ii) with regards to total NK cell count (60-70% for INKmune vs. 20-50%), assuming - worst case - that activated NK cell count would go down from 70% to 60% over days 29 and 119 (in a better scenario, NK cell count could still go up after day 29);

- (iii) with the trial having taken place in vivo vs. in vitro;

- (iv) with no added cytokine support for INKmune (and hence no side effects) vs. with added cytokine support (and possible side effects, which can be severe at times).

v Fifth tentative comparison – CIML + added cytokine support in cases of post-transplant relapse

In 2021, a report came out on 5 patients who had relapsed post haploidentical stem cell transplant, and had undergone CIML infusion with added IL-2 cytokine support until day 12.

A considerable expansion of total NK cell count after infusion of CIML was reported, with the development of a memory-like phenotype by day +28. By day +60 post the 12-day IL-2 cytokine support, the NK cells phenotypically resembled the NK cells present prior to infusion despite being present in significantly. Responses were attained in 4 of 5 patients at day +28, and treatment occurred in three cancers: acute myeloid leukemia (AML), myelodysplastic syndrome (MDS), and blastic plasmacytoid dendritic cell neoplasm (BPDCN). Three patients had complete remission (patients 2, 3 and 4). Patient 2 died from septic shock due to fungal pneumonia 8 weeks after infusion of CIML NK cells, but had no evidence of disease relapse. Patient 3 stayed in remission for 4 months and patient 4 for 6 months. One patient had a morphologic leukemia-free state, but had a persistent FLT3-ITD mutation detectable. Patient 5 had disease progression.

The report showed proof of NK cells trafficking to sites of disease after CIML-NK cell infusion (bone marrow in this case).

Following side effects due to cytokine-administration were reported:

- fever (temperature ranging from 38.1 ̊C to 39.3 ̊C) during the 12 days of IL-2 administration;

- grade 2 cytokine release syndrome (one patient);

- pancytopenia (4 patients - in two cases the pancytopenia was prolonged.

For reasons of comparison:

- In the INKmune trial, already as of day 8, activation of +50% of NK cells had been shown.

- Persistence of INKmune is obviously longer (4 months and possibly more with INKmune vs. two months, i.e. +100%), suggesting that a durable response with CIML cannot be sustained without continued cytokine support;

- The results this trial has generated are promising, both in relation to disease outcome and remission rate, as in relation to NK cells trafficking to the site of the tumor (including solid tumors – upcoming trial in ovarian cancer);

- INKmune comes with no toxicity, whereas in the above reporting, fever, cytokine release syndrome and pancyotopenia had bene reported.

Conclusion for investors

It would result from all of the above that INKmune’s results so far outperform earlier trials with memory-like NK cells, and may even do so in case of CIML combined with an additional form of therapy allowing for a standalone therapy.

The fair reference for me is shown in the first and second tentative comparison, where a persistence of respectively three weeks or two to four weeks had been noted. INKmune shows consistence persistence after 4 months, possibly longer. Sure, that’s the result in one patient, but it is consistent with earlier trial results, and so one may see it in further patients too. Also, as dosing’s safe and cheap, readministration when a patient’s NK cell count goes down could be considered.

Nothing excludes that, if that wouldn’t be the case or in certain situations, combination therapy can be used.


Outstanding work here, Carl!

Carl Kestens
Carl Kestens
Feb 28, 2022
Replying to

Thank you very much!

bottom of page