Oncofusion

start / graph / gene-cd22

Genestrong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.Q45-immunophenotype

CD22 (SIGLEC2) - the antigen that comes on BEFORE CD19, which makes it both the backup target and the pre-loaded escape route

also called CD22 · SIGLEC2 · Siglec-2 · BL-CAM · B-cell receptor CD22 · B-lymphocyte cell adhesion molecule · P20273

CD22 is another protein on B cells, and it is the obvious fallback when a CD19-directed therapy stops working. The uncomfortable finding is that CD22 appears earlier in B-cell development than CD19 does, so a CD34-positive, CD19-negative, CD22-positive cell already exists in most patients' marrow before any treatment starts - and the more of them a patient has, the more likely CD19-directed therapy is to fail.

The same thing, for a scientist

CD22 (UniProt **P20273**, CD22_HUMAN, **847 aa**, single reviewed human record, synonym SIGLEC2) is the B-lineage sialic-acid-binding lectin targeted by the antibody-drug conjugate inotuzumab ozogamicin and by CD22 CARs. This graph names it in 17 entries, carries `target-cd22-directed-therapy-kmt2a-r`, and has never written the molecule. **The developmental-ordering result is the reason this entry exists.** Bueno 2022 showed **CD22 expression precedes CD19 expression during B-cell development**, and that **CD34+CD19-CD22+ cells are present in diagnostic and relapsed marrow of about 70% of B-ALL patients**. Their frequency **doubles** in patients in complete remission after CD19 CAR T-cells, and the median pre-treatment frequency was **threefold higher in patients who went on to relapse** after CD19-directed immunotherapy. FISH on sorted fractions and xenografting showed these cells carry the diagnostic genetic abnormality and initiate leukaemia in vivo [E1]. **So CD22 is simultaneously the answer to CD19 escape and the marker of the compartment that causes it.** The authors' own conclusion is dual targeting plus routine CD34/CD19/CD22 immunophenotyping at diagnosis and during CD19-directed therapy. **In KMT2A-rearranged B-ALL, CD22 is present but comparatively dim.** In 1,382 B-ALL, KMT2A-rearranged cases had significantly **lower** CD22 mean fluorescence intensity than KMT2A-germline cases [E2]. That matters because **CD22 antigen density predicts response to inotuzumab**: in COG AALL1621, partial CD22 expression and lower CD22 site density were both associated with lower likelihood of response, and CD22 modulation was identified as a resistance mechanism [E3]. **And the paediatric trial evidence excludes the population this graph is about.** AALL1621 enrolled ages 1-21 [E3]; the ITCC-059 very-high-risk first-relapse cohort enrolled children **older than 1 year** and younger than 18, with KMT2A::AFF1 named explicitly as a qualifying high-risk lesion [E4]. **Infants are outside both.**

Go deeper
What CD22-directed therapy has actually achieved in children, with its costs

COG AALL1621 (phase II, single-arm, n = 48 evaluable, ages 1-21, relapsed/refractory CD22-positive B-ALL): CR/CRi 58.3% after cycle 1; of the 28 responders, 27 had flow MRD measured and 18 (66.7%) were below 0.01%. Costs: 7 of 28 responders (25%) had delayed count recovery past day 42, and 6 of 21 patients proceeding to transplant (28.6%) developed grade 3 sinusoidal obstruction syndrome 3.

That SOS number is the single most important thing in this entry for a therapeutic argument. Nearly three in ten transplanted children developed grade 3 hepatic SOS after inotuzumab. KMT2A-rearranged infant disease is a transplant-going disease. The agent's efficacy and its worst toxicity are concentrated in the same patients.

ITCC-059 moved inotuzumab earlier, to very-high-risk first relapse, and defined very-high-risk to include KMT2A::AFF1 by name - so the regulatory and trial infrastructure has explicitly identified this graph's index fusion as a CD22 target population, while excluding patients under one year 4.

Cell type, cell state, localisation, epigenetics
  • Cell type. B lineage, from earlier than CD19 1; lost on plasma-cell differentiation. Not on T cells or myeloid cells, which is why a lineage switch to AML makes CD22 as useless as CD19.
  • Cell state. Modulable under therapeutic pressure: CD22 site density falls and partial expression emerges, and that modulation is the named resistance mechanism in AALL1621 3. This is epitope/density escape, the same family as CD19 exon-2 skipping, and it is a different failure mode from lineage switch.
  • Localisation. Plasma membrane, type I transmembrane, 847 residues with seven Ig-like domains; rapidly internalised, which is what makes it an ADC target 5.
  • Epigenetics. Nothing in this beat establishes CD22 as a direct KMT2A-fusion target. CD22 is a B-lineage identity gene, so in this disease it is a read-out of whether the blast is still a B cell - which is precisely the quantity that lineage switch destroys.
Facts and figures
uniprotP20273
gene symbolCD22
length AA847
cd numberCD22
familySIGLEC (SIGLEC2); CD33/SIGLEC3 is the other family member in this beat
developmental orderCD22 precedes CD19 in B-cell development; CD34+CD19-CD22+ cells present in ~70% of B-ALL marrows, 2x enriched after CD19 CAR remission, 3x higher pre-treatment in those who later relapsed (PMID 35421218)
accession resolutiongene_exact:CD22 AND organism_id:9606 AND reviewed:true, fields include gene_primary and length; one row; primary name read back as CD22; length 847 asserted for verify-accessions.py.
trials
[
  {
    "trial": "COG AALL1621",
    "ages": "1-21",
    "n_evaluable": 48,
    "cr_cri_pct": 58.3,
    "mrd_under_0_01_pct_of_responders": 66.7,
    "grade3_sos_after_hsct": "6/21 (28.6%)",
    "pmid": "35007127",
    "title": "Phase II Trial of Inotuzumab Ozogamicin in Children and Adolescents With Relapsed or Refractory B-Cell Acute Lymphoblastic Leukemia: Children's Oncology Group Protocol AALL1621.",
    "authors": "O'Brien MM, Ji L, Shah NN, et al.",
    "venue": "J Clin Oncol",
    "year": 2022,
    "title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
  },
  {
    "trial": "ITCC-059 cohort 3",
    "ages": ">1 and <18",
    "design": "single-arm phase 2, very high risk FIRST relapse; KMT2A::AFF1 a qualifying lesion",
    "pmid": "42296977",
    "title": "Inotuzumab ozogamicin in paediatric very high risk first B-cell acute lymphoblastic leukaemia relapse (ITCC-059): a multicentre, single-arm, phase 2 trial.",
    "authors": "Peccatori N, Locatelli F, Ammerlaan ACJ, et al.",
    "venue": "Lancet Haematol",
    "year": 2026,
    "title_provenance": "Title/authors/venue backfilled from PubMed esummary on 2026-09-20 because this citation stored a bare PMID with no title, which made it unscoreable by verify-citations.py. BACKFILLING IS NOT VERIFICATION: it confirms the PMID resolves to a real paper, not that this is the right paper for the claim. That still requires reading."
  }
]
depmap kmt2a selectivity
{
  "chronos_mean_kmt2a_rearranged": -0.0022,
  "chronos_mean_other_leukaemia": -0.0479,
  "difference": 0.0457,
  "pan_essentiality_all_lines": -0.0446,
  "selectivity_rank": 14382,
  "rank_denominator": 17728,
  "provenance": "Joined 2026-09-21 from analyses/depmap-kmt2a-selective-scan-full (generated 2026-09-19T10:12:06Z), the project's own genome-wide scan over 11 KMT2A-rearranged and 43 other leukaemia DepMap lines. NOT re-fetched and NOT independent of that scan: every caveat it carries applies here, including that the KMT2A-rearranged arm is a small number of lines drawn from nine individuals, that the rank denominator is 17,728 genes, and that a difference is not a dependency. Added because only 26 of 308 gene entries carried this number, which made any KMT2A-selectivity question answerable for 8% of the corpus.",
  "how_to_read": "A negative difference means MORE required in KMT2A-rearranged lines than in other leukaemia lines. A rank near 17,728 is ANTI-selective. Neither is evidence of a fusion-driven dependency on its own: check pan_essentiality_all_lines before calling anything selective, and note that a paralogue-redundant gene can score near zero for reasons that have nothing to do with this disease - see ana-the-paralogue-redundancy-excuse-is-not-tested-and-my-proxy-decided-the-answer."
}
How this could bite us back
  • Nearly three in ten children transplanted after inotuzumab developed grade 3 hepatic sinusoidal obstruction syndrome in AALL1621. KMT2A-rearranged infant disease is a transplant-going disease, so the agent's worst toxicity is concentrated in exactly the patients most likely to receive it, and no paediatric CD22 trial found here includes infants.
  • CD22 as the answer to CD19 escape is undermined by the same study that recommends it: the CD34+CD19-CD22+ compartment that predicts CD19 failure is a CD22-POSITIVE compartment, so dual targeting removes the reservoir but also applies pressure to the only remaining B antigen. Project EVOLVE collected lineage switches after CD22 targeting as well as CD19.
  • CD22 site density modulates under therapy and low density predicts non-response. A patient can be CD22-positive at enrolment and functionally CD22-negative at retreatment, with no mutation and nothing for a sequencing assay to find.
  • KMT2A-rearranged B-ALL is CD22-dim by MFI relative to KMT2A-germline disease, and the resistance-relevant quantity is site density. Nobody has measured CD22 site density in infant KMT2A-rearranged blasts, so the genotype most explicitly named as a CD22 trial eligibility criterion is the one whose target density is unmeasured.
Still open
  1. What is the CD22 site density on infant KMT2A-rearranged blasts, and does it fall below the AALL1621 non-response threshold? This is a single flow measurement on banked material and would change trial eligibility reasoning.
  2. Is the CD34+CD19-CD22+ escape reservoir present in infant KMT2A-rearranged B-ALL, and does its size predict lineage switch rather than CD19-negative relapse?
  3. Does CD22-directed pressure shorten the time to lineage switch relative to CD19-directed pressure alone? Project EVOLVE has cases after both and no comparison.
  4. Why do paediatric CD22 trials exclude infants while naming KMT2A::AFF1 as a qualifying high-risk lesion? The answer is likely pharmacokinetic and hepatic, and it is not stated in any abstract read here.
Every source behind this entry (5)
  1. E1strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    CD22 expression precedes CD19 expression during B-cell development. CD34+CD19-CD22+ cells are found in diagnostic and relapsed bone marrow samples of approximately 70% of patients with B-ALL, their frequency increases twofold in patients in complete remission after CD19 CAR T-cell therapy, and the median frequency before treatment was threefold higher in patients who relapsed after CD19-directed immunotherapy over a median follow-up of 24 months. FISH on flow-sorted populations and xenograft modelling showed these cells harbour the genetic abnormalities present at diagnosis and initiate leukaemogenesis in vivo. The authors conclude that preleukaemic CD34+CD19-CD22+ progenitors underlie phenotypic escape from CD19-directed immunotherapies and encourage implementation of CD34/CD19/CD22 immunophenotyping for diagnosis and monitoring.

    PMID 35421218DOI
    What this evidence cannot show

    Abstract read in full from raw efetch XML on 2026-09-20; full text not read. **B-ALL generally; the abstract does not report the KMT2A-rearranged fraction**, so the threefold pre-treatment difference cannot be attributed to this graph's index disease. The relapse comparison is retrospective, defined by an outcome that happened after the measurement. Detecting a rare CD34+CD19-CD22+ population reliably is a demanding measurement and the abstract gives no absolute frequencies. 'Preleukaemic' is the authors' term for cells shown to carry the diagnostic abnormality and to engraft; it is not a demonstration that they preceded the leukaemia in time.

    Evidence that argues against this (1)
    • In KMT2A-rearranged disease the dominant escape from CD19-directed therapy is lineage switch to AML rather than retention of a B-progenitor phenotype: 53 of 75 collected lineage switches were B-ALL to AML and 64.3% of the B-ALL-derived cases carried KMT2A rearrangement. Switches were collected after CD22 targeting as well as CD19, so dual B-antigen targeting is not a demonstrated answer in this genotype. Silbert SK et al. 2025
  2. E2strong confidence 4 of 5: Replicated in peer-reviewed work by more than one group.paper

    Among 1,382 newly diagnosed B-ALL cases, KMT2A-rearranged B-ALLs showed significantly lower mean fluorescence intensity of CD34, CD22, TdT and CD123 and significantly higher MFI of CD45, CD38 and CD58 than KMT2A-germline cases, and these differences persisted within the same EGIL immunological subgroup.

    PMID 38468826DOI
    What this evidence cannot show

    Abstract read in full from raw efetch XML on 2026-09-20; full text not read. Single-centre, mixed adult and paediatric. **MFI is not antigen site density**: the AALL1621 resistance finding is stated in site density, and no conversion between the two is possible from these data. So 'KMT2A-r is CD22-dim by MFI' raises the question of whether these patients are worse inotuzumab candidates; it does not answer it.

    Evidence that argues against this (1)
    • ITCC-059 enrolled very-high-risk first-relapse children with KMT2A::AFF1 named as a qualifying lesion, which means the trial community judged these patients to be adequately CD22-positive to treat. A lower group-mean MFI is compatible with essentially all individual patients exceeding a clinical positivity threshold. Peccatori N et al. 2026
  3. E3established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    In Children's Oncology Group trial AALL1621, a single-arm phase II study of inotuzumab ozogamicin in patients aged 1-21 years with relapsed or refractory CD22-positive B-ALL, 48 patients were evaluable: 19 achieved complete response and 9 complete response with incomplete count recovery after cycle 1, a CR/CRi rate of 58.3% (two-sided 90% CI 46.5 to 69.3). Of 28 responders, 27 had flow MRD measured and 18 (66.7%) were below 0.01%. Seven of 28 responders (25%) had delayed count recovery past day 42 in cycle 1. Of 21 patients undergoing transplant after inotuzumab, 6 (28.6%) developed grade 3 sinusoidal obstruction syndrome. Partial CD22 expression and lower CD22 site density were associated with lower likelihood of response, and CD22 modulation was identified as a mechanism of resistance.

    PMID 35007127DOI
    What this evidence cannot show

    Abstract read in full from raw efetch XML on 2026-09-20; full text not read. **Single-arm, n = 48, heavily pretreated**, so there is no comparator for either efficacy or SOS rate. **Ages 1-21: infants are excluded**, which is the population this graph is about. CD22 expression was evaluated retrospectively by central flow, so the site-density-response association is an exploratory correlative analysis. The abstract does not report the KMT2A-rearranged fraction.

    Evidence that argues against this (1)
    • CD22 site density falling under pressure is antigen modulation, the same class of escape as CD19 exon-2 skipping - and in KMT2A-rearranged disease the competing escape is lineage switch, which this trial's design (single-arm, short follow-up, B-lineage response criteria) is not built to detect. A B-lineage flow panel scores a switched myeloid blast as remission. Kovach AE et al. 2026
  4. E4established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.paper

    ITCC-059 was a multicentre international phase 1-2 trial of inotuzumab ozogamicin in paediatric relapsed or refractory CD22-positive acute lymphoblastic leukaemia. A 2021 amendment added a third cohort at 20 hospitals across 12 countries enrolling children older than 1 year and younger than 18 years with CD22-positive very high risk FIRST B-ALL relapse. Very high risk relapse was defined as isolated bone marrow or combined relapse occurring less than 18 months after initial diagnosis, or with high-risk cytogenetic or molecular characteristics: KMT2A::AFF1, TCF3::PBX1, TCF3::HLF, hypodiploidy, or TP53 mutation/deletion. Inotuzumab was given at 1.8 mg/m2 in cycle 1 and 1.5 mg/m2 per subsequent cycle to a maximum of 6 cycles. Due to slow recruitment the statistical design was amended to reduce the sample size.

    PMID 42296977DOI
    What this evidence cannot show

    Abstract read from raw efetch XML on 2026-09-20 and was truncated by the record at the safety-analysis sentence, so **the efficacy result is NOT stated in this entry** - only the design and the eligibility criteria, which is what it is cited for. Full text not read. **Single-arm**, and **the statistical design was amended mid-trial to reduce the sample size because of slow recruitment**, with the primary-endpoint analysis restricted to the initial population meeting the amended minimum - an analysis-population decision that should be read in the full paper before any efficacy number from this trial is used. **Children older than 1 year only: infants are excluded.**

    Evidence that argues against this (1)
    • The explicit naming of KMT2A::AFF1 as a very-high-risk qualifier establishes that CD22-directed therapy is being moved earlier in exactly this graph's index genotype, while every paediatric CD22 trial found in this beat excludes patients under one year - the age band in which KMT2A::AFF1 is most common. The evidence base and the target population do not overlap. O'Brien MM et al. 2022
  5. E5established confidence 5 of 5: Multiple independent lines of evidence; treated as settled in the field.database

    CD22 is UniProt P20273 (CD22_HUMAN), B-cell receptor CD22 / B-lymphocyte cell adhesion molecule (BL-CAM) / Sialic acid-binding Ig-like lectin 2, 847 amino acids, a single reviewed human record with primary gene name CD22 and gene synonym SIGLEC2.

    What this evidence cannot show

    Fetched 2026-09-20 with `gene_exact:CD22 AND organism_id:9606 AND reviewed:true`, requesting `gene_primary` and `length` per rules 62 and 65; primary name read back; exactly one reviewed row. **CD22 is SIGLEC2 and CD33 is SIGLEC3 - two targets in the same beat from the same family**, which is the paralogue shape rule 63 warns about; the length 847 against CD33's 364 is the check that separates them.

    Evidence that argues against this (1)
    • The canonical length says nothing about what inotuzumab can bind. AALL1621 found that PARTIAL CD22 expression and lower site density predict non-response, so the therapeutically relevant quantity is molecules per cell on a given patient's blasts, which no accession supplies. O'Brien MM et al. 2022
Connected to